Core

Connections

class client.Connections(client)

Store and manage connections.

ConfigurationMetaNames = <ibm_watsonx_ai.metanames.ConnectionMetaNames object>

MetaNames for connection creation.

create(meta_props)

Create a connection. Examples of PROPERTIES field input:

1. MySQL

   client.connections.ConfigurationMetaNames.PROPERTIES: {
       "database": "database",
       "password": "password",
       "port": "3306",
       "host": "host url",
       "ssl": "false",
       "username": "username"
   }
   

2. Google BigQuery

   1. Method 1: Using service account json. The generated service account json can be provided as input as-is. Provide actual values in json. The example below is only indicative to show the fields. For information on how to generate the service account json, refer to Google BigQuery documentation.

      client.connections.ConfigurationMetaNames.PROPERTIES: {
          "type": "service_account",
          "project_id": "project_id",
          "private_key_id": "private_key_id",
          "private_key": "private key contents",
          "client_email": "client_email",
          "client_id": "client_id",
          "auth_uri": "https://accounts.google.com/o/oauth2/auth",
          "token_uri": "https://oauth2.googleapis.com/token",
          "auth_provider_x509_cert_url": "https://www.googleapis.com/oauth2/v1/certs",
          "client_x509_cert_url": "client_x509_cert_url"
      }
      

   2. Method 2: Using OAuth Method. For information on how to generate a OAuth token, refer to Google BigQuery documentation.

      client.connections.ConfigurationMetaNames.PROPERTIES: {
          "access_token": "access token generated for big query",
          "refresh_token": "refresh token",
          "project_id": "project_id",
          "client_secret": "This is your gmail account password",
          "client_id": "client_id"
      }
      

3. MS SQL

   client.connections.ConfigurationMetaNames.PROPERTIES: {
       "database": "database",
       "password": "password",
       "port": "1433",
       "host": "host",
       "username": "username"
   }
   

4. Teradata

   client.connections.ConfigurationMetaNames.PROPERTIES: {
       "database": "database",
       "password": "password",
       "port": "1433",
       "host": "host",
       "username": "username"
   }
   

Parameters:

meta_props (dict) –

metadata of the connection configuration. To see available meta names, use:

client.connections.ConfigurationMetaNames.get()

Returns:

metadata of the stored connection

Return type:

dict

Example:

sqlserver_data_source_type_id = client.connections.get_datasource_type_id_by_name('sqlserver')
connections_details = client.connections.create({
    client.connections.ConfigurationMetaNames.NAME: "sqlserver connection",
    client.connections.ConfigurationMetaNames.DESCRIPTION: "connection description",
    client.connections.ConfigurationMetaNames.DATASOURCE_TYPE: sqlserver_data_source_type_id,
    client.connections.ConfigurationMetaNames.PROPERTIES: { "database": "database",
                                                            "password": "password",
                                                            "port": "1433",
                                                            "host": "host",
                                                            "username": "username"}
})

delete(connection_id)

Delete a stored connection.

Parameters:

connection_id (str) – unique ID of the connection to be deleted

Returns:

status (“SUCCESS” or “FAILED”)

Return type:

str

Example:

client.connections.delete(connection_id)

get_datasource_type_details_by_id(datasource_type_id, connection_properties=False)

Get datasource type details for the given datasource type ID.

Parameters:

• datasource_type_id (str) – ID of the datasource type

• connection_properties (bool) – if True, the connection properties are included in the returned details. defaults to False

Returns:

Datasource type details

Return type:

dict

Example:

client.connections.get_datasource_type_details_by_id(datasource_type_id)

get_datasource_type_id_by_name(name)

Get a stored datasource type ID for the given datasource type name.

Parameters:

name (str) – name of datasource type

Returns:

ID of datasource type

Return type:

str

Example:

client.connections.get_datasource_type_id_by_name('cloudobjectstorage')

get_datasource_type_uid_by_name(name)

Get a stored datasource type ID for the given datasource type name.

Deprecated: Use Connections.get_datasource_type_id_by_name(name) instead.

Parameters:

name (str) – name of datasource type

Returns:

ID of datasource type

Return type:

str

Example:

client.connections.get_datasource_type_uid_by_name('cloudobjectstorage')

get_details(connection_id=None)

Get connection details for the given unique connection ID. If no connection_id is passed, details for all connections are returned.

Parameters:

connection_id (str) – unique ID of the connection

Returns:

metadata of the stored connection

Return type:

dict

Example:

connection_details = client.connections.get_details(connection_id)
connection_details = client.connections.get_details()

static get_id(connection_details)

Get ID of a stored connection.

Parameters:

connection_details (dict) – metadata of the stored connection

Returns:

unique ID of the stored connection

Return type:

str

Example:

connection_id = client.connection.get_id(connection_details)

static get_uid(connection_details)

Get the unique ID of a stored connection.

Deprecated: Use Connections.get_id(details) instead.

Parameters:

connection_details (dict) – metadata of the stored connection

Returns:

unique ID of the stored connection

Return type:

str

Example:

connection_uid = client.connection.get_uid(connection_details)

get_uploaded_db_drivers()

Get uploaded db driver jar names and paths. Supported for IBM Cloud Pak® for Data, version 4.6.1 and up.

Output

Important

Returns dictionary containing name and path for connection files.

return type: Dict[Str, Str]

Example:

>>> result = client.connections.get_uploaded_db_drivers()

list()

Return pd.DataFrame table with all stored connections in a table format.

Returns:

pandas.DataFrame with listed connections

Return type:

pandas.DataFrame

Example:

client.connections.list()

list_datasource_types()

Print stored datasource types assets in a table format.

Returns:

pandas.DataFrame with listed datasource types

Return type:

pandas.DataFrame

Example: https://test.cloud.ibm.com/apidocs/watsonx-ai#trainings-list

client.connections.list_datasource_types()

list_uploaded_db_drivers()

Return pd.DataFrame table with uploaded db driver jars in table a format. Supported for IBM Cloud Pak® for Data only.

Returns:

pandas.DataFrame with listed uploaded db drivers

Return type:

pandas.DataFrame

Example:

client.connections.list_uploaded_db_drivers()

sign_db_driver_url(jar_name)

Get a signed db driver jar URL to be used during JDBC generic connection creation. The jar name passed as argument needs to be uploaded into the system first. Supported for IBM Cloud Pak® for Data only, version 4.0.4 and later.

Parameters:

jar_name (str) – name of db driver jar

Returns:

URL of signed db driver

Return type:

str

Example:

jar_uri = client.connections.sign_db_driver_url('db2jcc4.jar')

upload_db_driver(path)

Upload db driver jar. Supported for IBM Cloud Pak® for Data only, version 4.0.4 and up.

Parameters:

path (str) – path to the db driver jar file

Example:

client.connections.upload_db_driver('example/path/db2jcc4.jar')

class metanames.ConnectionMetaNames

Set of MetaNames for Connection.

Available MetaNames:

MetaName	Type	Required	Example value
NAME	str	Y	my_space
DESCRIPTION	str	N	my_description
DATASOURCE_TYPE	str	Y	1e3363a5-7ccf-4fff-8022-4850a8024b68
PROPERTIES	dict	Y	{'database': 'db_name', 'host': 'host_url', 'password': 'password', 'username': 'user'}
FLAGS	list	N	['personal_credentials']

Data assets

class client.Assets(client)

Store and manage data assets.

ConfigurationMetaNames = <ibm_watsonx_ai.metanames.AssetsMetaNames object>

MetaNames for Data Assets creation.

create(name, file_path)

Create a data asset and upload content to it.

Parameters:

• name (str) – name to be given to the data asset

• file_path (str) – path to the content file to be uploaded

Returns:

metadata of the stored data asset

Return type:

dict

Example:

asset_details = client.data_assets.create(name="sample_asset", file_path="/path/to/file")

delete(asset_id=None, **kwargs)

Delete a stored data asset.

Parameters:

asset_id (str) – unique ID of the data asset

Returns:

status (“SUCCESS” or “FAILED”) or dictionary, if deleted asynchronously

Return type:

str or dict

Example:

client.data_assets.delete(asset_id)

download(asset_id=None, filename='', **kwargs)

Download and store the content of a data asset.

Parameters:

• asset_id (str) – unique ID of the data asset to be downloaded

• filename (str) – filename to be used for the downloaded file

Returns:

normalized path to the downloaded asset content

Return type:

str

Example:

client.data_assets.download(asset_id,"sample_asset.csv")

get_content(asset_id=None, **kwargs)

Download the content of a data asset.

Parameters:

asset_id (str) – unique ID of the data asset to be downloaded

Returns:

the asset content

Return type:

bytes

Example:

content = client.data_assets.get_content(asset_id).decode('ascii')

get_details(asset_id=None, get_all=None, limit=None, **kwargs)

Get data asset details. If no asset_id is passed, details for all assets are returned.

Parameters:

• asset_id (str) – unique ID of the asset

• limit (int, optional) – limit number of fetched records

• get_all (bool, optional) – if True, it will get all entries in ‘limited’ chunks

Returns:

metadata of the stored data asset

Return type:

dict

Example:

asset_details = client.data_assets.get_details(asset_id)

static get_href(asset_details)

Get the URL of a stored data asset.

Parameters:

asset_details (dict) – details of the stored data asset

Returns:

href of the stored data asset

Return type:

str

Example:

asset_details = client.data_assets.get_details(asset_id)
asset_href = client.data_assets.get_href(asset_details)

static get_id(asset_details)

Get the unique ID of a stored data asset.

Parameters:

asset_details (dict) – details of the stored data asset

Returns:

unique ID of the stored data asset

Return type:

str

Example:

asset_id = client.data_assets.get_id(asset_details)

list(limit=None)

Lists stored data assets in a table format. If limit is set to none, only the first 50 records are shown.

Parameters:

limit (int) – limit number for fetched records

Return type:

DataFrame

Returns:

listed elements

Example:

client.data_assets.list()

store(meta_props)

Create a data asset and upload content to it.

Parameters:

meta_props (dict) –

metadata of the space configuration. To see available meta names, use:

client.data_assets.ConfigurationMetaNames.get()

Example:

Example of data asset creation for files:

metadata = {
    client.data_assets.ConfigurationMetaNames.NAME: 'my data assets',
    client.data_assets.ConfigurationMetaNames.DESCRIPTION: 'sample description',
    client.data_assets.ConfigurationMetaNames.DATA_CONTENT_NAME: 'sample.csv'
}
asset_details = client.data_assets.store(meta_props=metadata)

Example of data asset creation using a connection:

metadata = {
    client.data_assets.ConfigurationMetaNames.NAME: 'my data assets',
    client.data_assets.ConfigurationMetaNames.DESCRIPTION: 'sample description',
    client.data_assets.ConfigurationMetaNames.CONNECTION_ID: '39eaa1ee-9aa4-4651-b8fe-95d3ddae',
    client.data_assets.ConfigurationMetaNames.DATA_CONTENT_NAME: 't1/sample.csv'
}
asset_details = client.data_assets.store(meta_props=metadata)

Example of data asset creation with a database sources type connection:

metadata = {
    client.data_assets.ConfigurationMetaNames.NAME: 'my data assets',
    client.data_assets.ConfigurationMetaNames.DESCRIPTION: 'sample description',
    client.data_assets.ConfigurationMetaNames.CONNECTION_ID: '23eaf1ee-96a4-4651-b8fe-95d3dadfe',
    client.data_assets.ConfigurationMetaNames.DATA_CONTENT_NAME: 't1'
}
asset_details = client.data_assets.store(meta_props=metadata)

class metanames.AssetsMetaNames

Set of MetaNames for Data Asset Specs.

Available MetaNames:

MetaName	Type	Required	Example value
NAME	str	Y	my_data_asset
DATA_CONTENT_NAME	str	Y	/test/sample.csv
CONNECTION_ID	str	N	39eaa1ee-9aa4-4651-b8fe-95d3ddae
DESCRIPTION	str	N	my_description

Deployments

class client.Deployments(client)

Deploy and score published artifacts (models and functions).

class HardwareRequestSizes(value)

An enum class that represents the different hardware request sizes available.

capitalize()

Return a capitalized version of the string.

More specifically, make the first character have upper case and the rest lower case.

casefold()

Return a version of the string suitable for caseless comparisons.

center(width, fillchar=' ', /)

Return a centered string of length width.

Padding is done using the specified fill character (default is a space).

count(sub[, start[, end]]) → int

Return the number of non-overlapping occurrences of substring sub in string S[start:end]. Optional arguments start and end are interpreted as in slice notation.

encode(encoding='utf-8', errors='strict')

Encode the string using the codec registered for encoding.

encoding

The encoding in which to encode the string.

errors

The error handling scheme to use for encoding errors. The default is ‘strict’ meaning that encoding errors raise a UnicodeEncodeError. Other possible values are ‘ignore’, ‘replace’ and ‘xmlcharrefreplace’ as well as any other name registered with codecs.register_error that can handle UnicodeEncodeErrors.

endswith(suffix[, start[, end]]) → bool

Return True if S ends with the specified suffix, False otherwise. With optional start, test S beginning at that position. With optional end, stop comparing S at that position. suffix can also be a tuple of strings to try.

expandtabs(tabsize=8)

Return a copy where all tab characters are expanded using spaces.

If tabsize is not given, a tab size of 8 characters is assumed.

find(sub[, start[, end]]) → int

Return the lowest index in S where substring sub is found, such that sub is contained within S[start:end]. Optional arguments start and end are interpreted as in slice notation.

Return -1 on failure.

format(*args, **kwargs) → str

Return a formatted version of S, using substitutions from args and kwargs. The substitutions are identified by braces (‘{’ and ‘}’).

format_map(mapping) → str

Return a formatted version of S, using substitutions from mapping. The substitutions are identified by braces (‘{’ and ‘}’).

index(sub[, start[, end]]) → int

Return the lowest index in S where substring sub is found, such that sub is contained within S[start:end]. Optional arguments start and end are interpreted as in slice notation.

Raises ValueError when the substring is not found.

isalnum()

Return True if the string is an alpha-numeric string, False otherwise.

A string is alpha-numeric if all characters in the string are alpha-numeric and there is at least one character in the string.

isalpha()

Return True if the string is an alphabetic string, False otherwise.

A string is alphabetic if all characters in the string are alphabetic and there is at least one character in the string.

isascii()

Return True if all characters in the string are ASCII, False otherwise.

ASCII characters have code points in the range U+0000-U+007F. Empty string is ASCII too.

isdecimal()

Return True if the string is a decimal string, False otherwise.

A string is a decimal string if all characters in the string are decimal and there is at least one character in the string.

isdigit()

Return True if the string is a digit string, False otherwise.

A string is a digit string if all characters in the string are digits and there is at least one character in the string.

isidentifier()

Return True if the string is a valid Python identifier, False otherwise.

Call keyword.iskeyword(s) to test whether string s is a reserved identifier, such as “def” or “class”.

islower()

Return True if the string is a lowercase string, False otherwise.

A string is lowercase if all cased characters in the string are lowercase and there is at least one cased character in the string.

isnumeric()

Return True if the string is a numeric string, False otherwise.

A string is numeric if all characters in the string are numeric and there is at least one character in the string.

isprintable()

Return True if the string is printable, False otherwise.

A string is printable if all of its characters are considered printable in repr() or if it is empty.

isspace()

Return True if the string is a whitespace string, False otherwise.

A string is whitespace if all characters in the string are whitespace and there is at least one character in the string.

istitle()

Return True if the string is a title-cased string, False otherwise.

In a title-cased string, upper- and title-case characters may only follow uncased characters and lowercase characters only cased ones.

isupper()

Return True if the string is an uppercase string, False otherwise.

A string is uppercase if all cased characters in the string are uppercase and there is at least one cased character in the string.

join(iterable, /)

Concatenate any number of strings.

The string whose method is called is inserted in between each given string. The result is returned as a new string.

Example: ‘.’.join([‘ab’, ‘pq’, ‘rs’]) -> ‘ab.pq.rs’

ljust(width, fillchar=' ', /)

Return a left-justified string of length width.

Padding is done using the specified fill character (default is a space).

lower()

Return a copy of the string converted to lowercase.

lstrip(chars=None, /)

Return a copy of the string with leading whitespace removed.

If chars is given and not None, remove characters in chars instead.

static maketrans()

Return a translation table usable for str.translate().

If there is only one argument, it must be a dictionary mapping Unicode ordinals (integers) or characters to Unicode ordinals, strings or None. Character keys will be then converted to ordinals. If there are two arguments, they must be strings of equal length, and in the resulting dictionary, each character in x will be mapped to the character at the same position in y. If there is a third argument, it must be a string, whose characters will be mapped to None in the result.

partition(sep, /)

Partition the string into three parts using the given separator.

This will search for the separator in the string. If the separator is found, returns a 3-tuple containing the part before the separator, the separator itself, and the part after it.

If the separator is not found, returns a 3-tuple containing the original string and two empty strings.

removeprefix(prefix, /)

Return a str with the given prefix string removed if present.

If the string starts with the prefix string, return string[len(prefix):]. Otherwise, return a copy of the original string.

removesuffix(suffix, /)

Return a str with the given suffix string removed if present.

If the string ends with the suffix string and that suffix is not empty, return string[:-len(suffix)]. Otherwise, return a copy of the original string.

replace(old, new, count=-1, /)

Return a copy with all occurrences of substring old replaced by new.

count

Maximum number of occurrences to replace. -1 (the default value) means replace all occurrences.

If the optional argument count is given, only the first count occurrences are replaced.

rfind(sub[, start[, end]]) → int

Return the highest index in S where substring sub is found, such that sub is contained within S[start:end]. Optional arguments start and end are interpreted as in slice notation.

Return -1 on failure.

rindex(sub[, start[, end]]) → int

Return the highest index in S where substring sub is found, such that sub is contained within S[start:end]. Optional arguments start and end are interpreted as in slice notation.

Raises ValueError when the substring is not found.

rjust(width, fillchar=' ', /)

Return a right-justified string of length width.

Padding is done using the specified fill character (default is a space).

rpartition(sep, /)

Partition the string into three parts using the given separator.

This will search for the separator in the string, starting at the end. If the separator is found, returns a 3-tuple containing the part before the separator, the separator itself, and the part after it.

If the separator is not found, returns a 3-tuple containing two empty strings and the original string.

rsplit(sep=None, maxsplit=-1)

Return a list of the substrings in the string, using sep as the separator string.

sep

The separator used to split the string.

When set to None (the default value), will split on any whitespace character (including \n \r \t \f and spaces) and will discard empty strings from the result.

maxsplit

Maximum number of splits (starting from the left). -1 (the default value) means no limit.

Splitting starts at the end of the string and works to the front.

rstrip(chars=None, /)

Return a copy of the string with trailing whitespace removed.

If chars is given and not None, remove characters in chars instead.

split(sep=None, maxsplit=-1)

Return a list of the substrings in the string, using sep as the separator string.

sep

The separator used to split the string.

When set to None (the default value), will split on any whitespace character (including \n \r \t \f and spaces) and will discard empty strings from the result.

maxsplit

Maximum number of splits (starting from the left). -1 (the default value) means no limit.

Note, str.split() is mainly useful for data that has been intentionally delimited. With natural text that includes punctuation, consider using the regular expression module.

splitlines(keepends=False)

Return a list of the lines in the string, breaking at line boundaries.

Line breaks are not included in the resulting list unless keepends is given and true.

startswith(prefix[, start[, end]]) → bool

Return True if S starts with the specified prefix, False otherwise. With optional start, test S beginning at that position. With optional end, stop comparing S at that position. prefix can also be a tuple of strings to try.

strip(chars=None, /)

Return a copy of the string with leading and trailing whitespace removed.

If chars is given and not None, remove characters in chars instead.

swapcase()

Convert uppercase characters to lowercase and lowercase characters to uppercase.

title()

Return a version of the string where each word is titlecased.

More specifically, words start with uppercased characters and all remaining cased characters have lower case.

translate(table, /)

Replace each character in the string using the given translation table.

table

Translation table, which must be a mapping of Unicode ordinals to Unicode ordinals, strings, or None.

The table must implement lookup/indexing via __getitem__, for instance a dictionary or list. If this operation raises LookupError, the character is left untouched. Characters mapped to None are deleted.

upper()

Return a copy of the string converted to uppercase.

zfill(width, /)

Pad a numeric string with zeros on the left, to fill a field of the given width.

The string is never truncated.

create(artifact_id=None, meta_props=None, rev_id=None, **kwargs)

Create a deployment from an artifact. An artifact is a model or function that can be deployed.

Parameters:

• artifact_id (str) – ID of the published artifact (the model or function ID)

• meta_props (dict, optional) –

  meta props. To see the available list of meta names, use:

  client.deployments.ConfigurationMetaNames.get()
  

• rev_id (str, optional) – revision ID of the deployment

Returns:

metadata of the created deployment

Return type:

dict

Example:

meta_props = {
    client.deployments.ConfigurationMetaNames.NAME: "SAMPLE DEPLOYMENT NAME",
    client.deployments.ConfigurationMetaNames.ONLINE: {},
    client.deployments.ConfigurationMetaNames.HARDWARE_SPEC : { "id":  "e7ed1d6c-2e89-42d7-aed5-8sb972c1d2b"},
    client.deployments.ConfigurationMetaNames.SERVING_NAME : 'sample_deployment'
}
deployment_details = client.deployments.create(artifact_id, meta_props)

create_job(deployment_id, meta_props, retention=None, transaction_id=None, _asset_id=None)

Create an asynchronous deployment job.

Parameters:

• deployment_id (str) – unique ID of the deployment

• meta_props (dict) – metaprops. To see the available list of metanames, use client.deployments.ScoringMetaNames.get() or client.deployments.DecisionOptimizationmetaNames.get()

• retention (int, optional) – how many job days job meta should be retained, takes integer values >= -1, supported only on Cloud

• transaction_id (str, optional) – transaction ID to be passed with the payload

Returns:

metadata of the created async deployment job

Return type:

dict or str

Note

• The valid payloads for scoring input are either list of values, pandas or numpy dataframes.

Example:

scoring_payload = {client.deployments.ScoringMetaNames.INPUT_DATA: [{'fields': ['GENDER','AGE','MARITAL_STATUS','PROFESSION'],
                                                                         'values': [['M',23,'Single','Student'],
                                                                                    ['M',55,'Single','Executive']]}]}
async_job = client.deployments.create_job(deployment_id, scoring_payload)

delete(deployment_id=None, **kwargs)

Delete a deployment.

Parameters:

deployment_id (str) – unique ID of the deployment

Returns:

status (“SUCCESS” or “FAILED”)

Return type:

str

Example:

client.deployments.delete(deployment_id)

delete_job(job_id=None, hard_delete=False, **kwargs)

Delete a deployment job that is running. This method can also delete metadata details of completed or canceled jobs when hard_delete parameter is set to True.

Parameters:

• job_id (str) – unique ID of the deployment job to be deleted

• hard_delete (bool, optional) –

  specify True or False:

  True - To delete the completed or canceled job.

  False - To cancel the currently running deployment job.

Returns:

status (“SUCCESS” or “FAILED”)

Return type:

str

Example:

client.deployments.delete_job(job_id)

generate(deployment_id, prompt=None, params=None, guardrails=False, guardrails_hap_params=None, guardrails_pii_params=None, concurrency_limit=10, async_mode=False, validate_prompt_variables=True)

Generate a raw response with prompt for given deployment_id.

Parameters:

• deployment_id (str) – unique ID of the deployment

• prompt (str, optional) – prompt needed for text generation. If deployment_id points to the Prompt Template asset, then the prompt argument must be None, defaults to None

• params (dict, optional) – meta props for text generation, use ibm_watsonx_ai.metanames.GenTextParamsMetaNames().show() to view the list of MetaNames

• guardrails (bool, optional) – If True, then potentially hateful, abusive, and/or profane language (HAP) was detected filter is toggle on for both prompt and generated text, defaults to False

• guardrails_hap_params (dict, optional) – meta props for HAP moderations, use ibm_watsonx_ai.metanames.GenTextModerationsMetaNames().show() to view the list of MetaNames

• concurrency_limit (int, optional) – number of requests to be sent in parallel, maximum is 10

• async_mode (bool, optional) – If True, then yield results asynchronously (using generator). In this case both the prompt and the generated text will be concatenated in the final response - under generated_text, defaults to False

• validate_prompt_variables (bool) – If True, prompt variables provided in params are validated with the ones in Prompt Template Asset. This parameter is only applicable in a Prompt Template Asset deployment scenario and should not be changed for different cases, defaults to True

Returns:

scoring result containing generated content

Return type:

dict

generate_text(deployment_id, prompt=None, params=None, raw_response=False, guardrails=False, guardrails_hap_params=None, guardrails_pii_params=None, concurrency_limit=10, validate_prompt_variables=True)

Given the selected deployment (deployment_id), a text prompt as input, and the parameters and concurrency_limit, the selected inference will generate a completion text as generated_text response.

Parameters:

• deployment_id (str) – unique ID of the deployment

• prompt (str, optional) – the prompt string or list of strings. If the list of strings is passed, requests will be managed in parallel with the rate of concurency_limit, defaults to None

• params (dict, optional) – meta props for text generation, use ibm_watsonx_ai.metanames.GenTextParamsMetaNames().show() to view the list of MetaNames

• raw_response (bool, optional) – returns the whole response object

• guardrails (bool, optional) – If True, then potentially hateful, abusive, and/or profane language (HAP) was detected filter is toggle on for both prompt and generated text, defaults to False

• guardrails_hap_params (dict, optional) – meta props for HAP moderations, use ibm_watsonx_ai.metanames.GenTextModerationsMetaNames().show() to view the list of MetaNames

• concurrency_limit (int, optional) – number of requests to be sent in parallel, maximum is 10

• validate_prompt_variables (bool) – If True, prompt variables provided in params are validated with the ones in Prompt Template Asset. This parameter is only applicable in a Prompt Template Asset deployment scenario and should not be changed for different cases, defaults to True

Returns:

generated content

Return type:

str

Note

By default only the first occurance of HAPDetectionWarning is displayed. To enable printing all warnings of this category, use:

import warnings
from ibm_watsonx_ai.foundation_models.utils import HAPDetectionWarning

warnings.filterwarnings("always", category=HAPDetectionWarning)

generate_text_stream(deployment_id, prompt=None, params=None, raw_response=False, guardrails=False, guardrails_hap_params=None, guardrails_pii_params=None, validate_prompt_variables=True)

Given the selected deployment (deployment_id), a text prompt as input and parameters, the selected inference will generate a streamed text as generate_text_stream.

Parameters:

• deployment_id (str) – unique ID of the deployment

• prompt (str, optional) – the prompt string, defaults to None

• params (dict, optional) – meta props for text generation, use ibm_watsonx_ai.metanames.GenTextParamsMetaNames().show() to view the list of MetaNames

• raw_response (bool, optional) – yields the whole response object

• guardrails (bool, optional) – If True, then potentially hateful, abusive, and/or profane language (HAP) was detected filter is toggle on for both prompt and generated text, defaults to False

• guardrails_hap_params (dict, optional) – meta props for HAP moderations, use ibm_watsonx_ai.metanames.GenTextModerationsMetaNames().show() to view the list of MetaNames

• validate_prompt_variables (bool) – If True, prompt variables provided in params are validated with the ones in Prompt Template Asset. This parameter is only applicable in a Prompt Template Asset deployment scenario and should not be changed for different cases, defaults to True

Returns:

generated content

Return type:

str

Note

By default only the first occurance of HAPDetectionWarning is displayed. To enable printing all warnings of this category, use:

import warnings
from ibm_watsonx_ai.foundation_models.utils import HAPDetectionWarning

warnings.filterwarnings("always", category=HAPDetectionWarning)

get_details(deployment_id=None, serving_name=None, limit=None, asynchronous=False, get_all=False, spec_state=None, _silent=False, **kwargs)

Get information about deployment(s). If deployment_id is not passed, all deployment details are returned.

Parameters:

• deployment_id (str, optional) – unique ID of the deployment

• serving_name (str, optional) – serving name that filters deployments

• limit (int, optional) – limit number of fetched records

• asynchronous (bool, optional) – if True, it will work as a generator

• get_all (bool, optional) – if True, it will get all entries in ‘limited’ chunks

• spec_state (SpecStates, optional) – software specification state, can be used only when deployment_id is None

Returns:

metadata of the deployment(s)

Return type:

dict (if deployment_id is not None) or {“resources”: [dict]} (if deployment_id is None)

Example:

deployment_details = client.deployments.get_details(deployment_id)
deployment_details = client.deployments.get_details(deployment_id=deployment_id)
deployments_details = client.deployments.get_details()
deployments_details = client.deployments.get_details(limit=100)
deployments_details = client.deployments.get_details(limit=100, get_all=True)
deployments_details = []
for entry in client.deployments.get_details(limit=100, asynchronous=True, get_all=True):
    deployments_details.extend(entry)

get_download_url(deployment_details)

Get deployment_download_url from the deployment details.

Parameters:

deployment_details (dict) – created deployment details

Returns:

deployment download URL that is used to get file deployment (for example: Core ML)

Return type:

str

Example:

deployment_url = client.deployments.get_download_url(deployment)

static get_href(deployment_details)

Get deployment_href from the deployment details.

Parameters:

deployment_details (dict) – metadata of the deployment

Returns:

deployment href that is used to manage the deployment

Return type:

str

Example:

deployment_href = client.deployments.get_href(deployment)

static get_id(deployment_details)

Get the deployment ID from the deployment details.

Parameters:

deployment_details (dict) – metadata of the deployment

Returns:

deployment ID that is used to manage the deployment

Return type:

str

Example:

deployment_id = client.deployments.get_id(deployment)

get_job_details(job_id=None, include=None, limit=None, **kwargs)

Get information about deployment job(s). If deployment job_id is not passed, all deployment jobs details are returned.

Parameters:

• job_id (str, optional) – unique ID of the job

• include (str, optional) – fields to be retrieved from ‘decision_optimization’ and ‘scoring’ section mentioned as value(s) (comma separated) as output response fields

• limit (int, optional) – limit number of fetched records

Returns:

metadata of deployment job(s)

Return type:

dict (if job_id is not None) or {“resources”: [dict]} (if job_id is None)

Example:

deployment_details = client.deployments.get_job_details()
deployments_details = client.deployments.get_job_details(job_id=job_id)

get_job_href(job_details)

Get the href of a deployment job.

Parameters:

job_details (dict) – metadata of the deployment job

Returns:

href of the deployment job

Return type:

str

Example:

job_details = client.deployments.get_job_details(job_id=job_id)
job_status = client.deployments.get_job_href(job_details)

get_job_id(job_details)

Get the unique ID of a deployment job.

Parameters:

job_details (dict) – metadata of the deployment job

Returns:

unique ID of the deployment job

Return type:

str

Example:

job_details = client.deployments.get_job_details(job_id=job_id)
job_status = client.deployments.get_job_id(job_details)

get_job_status(job_id)

Get the status of a deployment job.

Parameters:

job_id (str) – unique ID of the deployment job

Returns:

status of the deployment job

Return type:

dict

Example:

job_status = client.deployments.get_job_status(job_id)

get_job_uid(job_details)

Get the unique ID of a deployment job.

Deprecated: Use get_job_id(job_details) instead.

Parameters:

job_details (dict) – metadata of the deployment job

Returns:

unique ID of the deployment job

Return type:

str

Example:

job_details = client.deployments.get_job_details(job_uid=job_uid)
job_status = client.deployments.get_job_uid(job_details)

static get_scoring_href(deployment_details)

Get scoring URL from deployment details.

Parameters:

deployment_details (dict) – metadata of the deployment

Returns:

scoring endpoint URL that is used to make scoring requests

Return type:

str

Example:

scoring_href = client.deployments.get_scoring_href(deployment)

static get_serving_href(deployment_details)

Get serving URL from the deployment details.

Parameters:

deployment_details (dict) – metadata of the deployment

Returns:

serving endpoint URL that is used to make scoring requests

Return type:

str

Example:

scoring_href = client.deployments.get_serving_href(deployment)

static get_uid(deployment_details)

Get deployment_uid from the deployment details.

Deprecated: Use get_id(deployment_details) instead.

Parameters:

deployment_details (dict) – metadata of the deployment

Returns:

deployment UID that is used to manage the deployment

Return type:

str

Example:

deployment_uid = client.deployments.get_uid(deployment)

is_serving_name_available(serving_name)

Check if the serving name is available for use.

Parameters:

serving_name (str) – serving name that filters deployments

Returns:

information about whether the serving name is available

Return type:

bool

Example:

is_available = client.deployments.is_serving_name_available('test')

list(limit=None, artifact_type=None)

Returns deployments in a table format. If limit is set to None, only the first 50 records are shown.

Parameters:

• limit (int, optional) – limit number of fetched records

• artifact_type (str, optional) – return only deployments with the specified artifact_type

Returns:

pandas.DataFrame with the listed deployments

Return type:

pandas.DataFrame

Example:

client.deployments.list()

list_jobs(limit=None)

Return the async deployment jobs in a table format. If the limit is set to None, only the first 50 records are shown.

Parameters:

limit (int, optional) – limit number of fetched records

Returns:

pandas.DataFrame with listed deployment jobs

Return type:

pandas.DataFrame

Note

This method list only async deployment jobs created for WML deployment.

Example:

client.deployments.list_jobs()

run_ai_service(deployment_id, ai_service_payload)

Execute an AI service by providing a scoring payload.

Parameters:

• deployment_id (str) – unique ID of the deployment

• ai_service_payload (dict) – AI service payload to be passed to generate the method

Returns:

response of the AI service

Return type:

Any

run_ai_service_stream(deployment_id, ai_service_payload)

Execute an AI service by providing a scoring payload.

Parameters:

• deployment_id (str) – unique ID of the deployment

• ai_service_payload (dict) – AI service payload to be passed to generate the method

Returns:

response of the AI service

Return type:

str

score(deployment_id, meta_props, transaction_id=None)

Make scoring requests against the deployed artifact.

Parameters:

• deployment_id (str) – unique ID of the deployment to be scored

• meta_props (dict) – meta props for scoring, use client.deployments.ScoringMetaNames.show() to view the list of ScoringMetaNames

• transaction_id (str, optional) – transaction ID to be passed with the records during payload logging

Returns:

scoring result that contains prediction and probability

Return type:

dict

Note

• client.deployments.ScoringMetaNames.INPUT_DATA is the only metaname valid for sync scoring.

• The valid payloads for scoring input are either list of values, pandas or numpy dataframes.

Example:

scoring_payload = {client.deployments.ScoringMetaNames.INPUT_DATA:
    [{'fields':
        ['GENDER','AGE','MARITAL_STATUS','PROFESSION'],
        'values': [
            ['M',23,'Single','Student'],
            ['M',55,'Single','Executive']
        ]
    }]
}
predictions = client.deployments.score(deployment_id, scoring_payload)

update(deployment_id=None, changes=None, **kwargs)

Updates existing deployment metadata. If ASSET is patched, then ‘id’ field is mandatory and it starts a deployment with the provided asset id/rev. Deployment ID remains the same.

Parameters:

• deployment_id (str) – unique ID of deployment to be updated

• changes (dict) – elements to be changed, where keys are ConfigurationMetaNames

Returns:

metadata of the updated deployment

Return type:

dict or None

Examples

metadata = {client.deployments.ConfigurationMetaNames.NAME:"updated_Deployment"}
updated_deployment_details = client.deployments.update(deployment_id, changes=metadata)

metadata = {client.deployments.ConfigurationMetaNames.ASSET: {  "id": "ca0cd864-4582-4732-b365-3165598dc945",
                                                                "rev":"2" }}
deployment_details = client.deployments.update(deployment_id, changes=metadata)

class metanames.DeploymentMetaNames

Set of MetaNames for Deployments Specs.

Available MetaNames:

MetaName	Type	Required	Schema	Example value
TAGS	list	N	['string']	['string1', 'string2']
NAME	str	N		my_deployment
DESCRIPTION	str	N		my_deployment
CUSTOM	dict	N		{}
ASSET	dict	N		{'id': '4cedab6d-e8e4-4214-b81a-2ddb122db2ab', 'rev': '1'}
PROMPT_TEMPLATE	dict	N		{'id': '4cedab6d-e8e4-4214-b81a-2ddb122db2ab'}
HARDWARE_SPEC	dict	N		{'id': '3342-1ce536-20dc-4444-aac7-7284cf3befc'}
HARDWARE_REQUEST	dict	N		{'size': 'gpu_s', 'num_nodes': 1}
HYBRID_PIPELINE_HARDWARE_SPECS	list	N		[{'node_runtime_id': 'auto_ai.kb', 'hardware_spec': {'id': '3342-1ce536-20dc-4444-aac7-7284cf3befc', 'num_nodes': '2'}}]
ONLINE	dict	N		{}
BATCH	dict	N		{}
DETACHED	dict	N		{}
R_SHINY	dict	N		{'authentication': 'anyone_with_url'}
VIRTUAL	dict	N		{}
OWNER	str	N		<owner_id>
BASE_MODEL_ID	str	N		google/flan-ul2
BASE_DEPLOYMENT_ID	str	N		76a60161-facb-4968-a475-a6f1447c44bf
PROMPT_VARIABLES	dict	N		{'key': 'value'}

class ibm_watsonx_ai.utils.enums.RShinyAuthenticationValues(value)

Allowable values of R_Shiny authentication.

ANYONE_WITH_URL = 'anyone_with_url'

ANY_VALID_USER = 'any_valid_user'

MEMBERS_OF_DEPLOYMENT_SPACE = 'members_of_deployment_space'

class metanames.ScoringMetaNames

Set of MetaNames for Scoring.

Available MetaNames:

MetaName	Type	Required	Schema	Example value
NAME	str	N		jobs test
INPUT_DATA	list	N	[{'name(optional)': 'string', 'id(optional)': 'string', 'fields(optional)': 'array[string]', 'values': 'array[array[string]]'}]	[{'fields': ['name', 'age', 'occupation'], 'values': [['john', 23, 'student']]}]
INPUT_DATA_REFERENCES	list	N	[{'id(optional)': 'string', 'type(required)': 'string', 'connection(required)': {'href(required)': 'string'}, 'location(required)': {'bucket': 'string', 'path': 'string'}, 'schema(optional)': {'id(required)': 'string', 'fields(required)': [{'name(required)': 'string', 'type(required)': 'string', 'nullable(optional)': 'string'}]}}]
OUTPUT_DATA_REFERENCE	dict	N	{'type(required)': 'string', 'connection(required)': {'href(required)': 'string'}, 'location(required)': {'bucket': 'string', 'path': 'string'}, 'schema(optional)': {'id(required)': 'string', 'fields(required)': [{'name(required)': 'string', 'type(required)': 'string', 'nullable(optional)': 'string'}]}}
EVALUATIONS_SPEC	list	N	[{'id(optional)': 'string', 'input_target(optional)': 'string', 'metrics_names(optional)': 'array[string]'}]	[{'id': 'string', 'input_target': 'string', 'metrics_names': ['auroc', 'accuracy']}]
ENVIRONMENT_VARIABLES	dict	N		{'my_env_var1': 'env_var_value1', 'my_env_var2': 'env_var_value2'}
SCORING_PARAMETERS	dict	N		{'forecast_window': 50}

class metanames.DecisionOptimizationMetaNames

Set of MetaNames for Decision Optimization.

Available MetaNames:

MetaName	Type	Required	Schema	Example value
INPUT_DATA	list	N	[{'name(optional)': 'string', 'id(optional)': 'string', 'fields(optional)': 'array[string]', 'values': 'array[array[string]]'}]	[{'fields': ['name', 'age', 'occupation'], 'values': [['john', 23, 'student']]}]
INPUT_DATA_REFERENCES	list	N	[{'name(optional)': 'string', 'id(optional)': 'string', 'fields(optional)': 'array[string]', 'values': 'array[array[string]]'}]	[{'fields': ['name', 'age', 'occupation'], 'values': [['john', 23, 'student']]}]
OUTPUT_DATA	list	N	[{'name(optional)': 'string'}]
OUTPUT_DATA_REFERENCES	list	N	{'name(optional)': 'string', 'type(required)': 'string', 'connection(required)': {'endpoint_url(required)': 'string', 'access_key_id(required)': 'string', 'secret_access_key(required)': 'string'}, 'location(required)': {'bucket': 'string', 'path': 'string'}, 'schema(optional)': {'id(required)': 'string', 'fields(required)': [{'name(required)': 'string', 'type(required)': 'string', 'nullable(optional)': 'string'}]}}
SOLVE_PARAMETERS	dict	N

class ibm_watsonx_ai.deployments.RuntimeContext(api_client, request_payload_json=None)

Class included to keep the interface compatible with the Deployment’s RuntimeContext used in AIServices implementation.

Parameters:

• api_client (APIClient) – initialized APIClient object with a set project ID or space ID. If passed, credentials and project_id/space_id are not required.

• request_payload_json (dict, optional) – Request payload for testing of generate/ generate_stream call of AI Service.

`` RuntimeContext`` initialized for testing purposes before deployment:

context = RuntimeContext(api_client=client, request_payload_json={"field": "value"})

Examples of RuntimeContext usage within AI Service source code:

def deployable_ai_service(context, **custom):
    task_token = context.generate_token()

    def generate(context) -> dict:
        user_token = context.get_token()
        headers = context.get_headers()
        json_body = context.get_json()
        ...
        return {"body": json_body}

    return generate

generate = deployable_ai_service(context)
generate_output = generate(context)  # returns {"body": {"field": "value"}}

Change the JSON body in RuntimeContext:

context.request_payload_json = {"field2": "value2"}

generate = deployable_ai_service(context)
generate_output = generate(context)  # returns {"body": {"field2": "value2"}}

generate_token()

Return refreshed token.

get_headers()

Return headers with refreshed token

get_json()

Get payload JSON send in body of API request to the generate or generate_stream method in deployed AIService. For testing purposes the payload JSON need to be set in RuntimeContext initialization or later as request_payload_json property.

get_token()

Return user token.

Export/Import

class client.Export(client)

cancel(export_id, space_id=None, project_id=None)

Cancel an export job. space_id or project_id has to be provided.

Note

To delete an export_id job, use delete() API.

Parameters:

• export_id (str) – export job identifier

• space_id (str, optional) – space identifier

• project_id (str, optional) – project identifier

Returns:

status (“SUCCESS” or “FAILED”)

Return type:

str

Example:

client.export_assets.cancel(export_id='6213cf1-252f-424b-b52d-5cdd9814956c',
                            space_id='3421cf1-252f-424b-b52d-5cdd981495fe')

delete(export_id, space_id=None, project_id=None)

Delete the given export_id job. space_id or project_id has to be provided.

Parameters:

• export_id (str) – export job identifier

• space_id (str, optional) – space identifier

• project_id (str, optional) – project identifier

Returns:

status (“SUCCESS” or “FAILED”)

Return type:

str

Example:

client.export_assets.delete(export_id='6213cf1-252f-424b-b52d-5cdd9814956c',
                            space_id= '98a53931-a8c0-4c2f-8319-c793155e4598')

get_details(export_id=None, space_id=None, project_id=None, limit=None, asynchronous=False, get_all=False)

Get metadata of a given export job. If no export_id is specified, all export metadata is returned.

Parameters:

• export_id (str, optional) – export job identifier

• space_id (str, optional) – space identifier

• project_id (str, optional) – project identifier

• limit (int, optional) – limit number of fetched records

• asynchronous (bool, optional) – if True, it will work as a generator

• get_all (bool, optional) – if True, it will get all entries in ‘limited’ chunks

Returns:

export metadata

Return type:

dict (if export_id is not None) or {“resources”: [dict]} (if export_id is None)

Example:

details = client.export_assets.get_details(export_id, space_id= '98a53931-a8c0-4c2f-8319-c793155e4598')
details = client.export_assets.get_details()
details = client.export_assets.get_details(limit=100)
details = client.export_assets.get_details(limit=100, get_all=True)
details = []
for entry in client.export_assets.get_details(limit=100, asynchronous=True, get_all=True):
    details.extend(entry)

get_exported_content(export_id, space_id=None, project_id=None, file_path=None)

Get the exported content as a zip file.

Parameters:

• export_id (str) – export job identifier

• space_id (str, optional) – space identifier

• project_id (str, optional) – project identifier

• file_path (str, optional) – name of local file to create, this should be absolute path of the file and the file shouldn’t exist

Returns:

path to the downloaded function content

Return type:

str

Example:

client.exports.get_exported_content(export_id,
                                    space_id='98a53931-a8c0-4c2f-8319-c793155e4598',
                                    file_path='/home/user/my_exported_content.zip')

static get_id(export_details)

Get the ID of the export job from export details.

Parameters:

export_details (dict) – metadata of the export job

Returns:

ID of the export job

Return type:

str

Example:

id = client.export_assets.get_id(export_details)

list(space_id=None, project_id=None, limit=None)

Return export jobs in a table format. If limit is set to None, only the first 50 records are shown.

Parameters:

• space_id (str, optional) – space identifier

• project_id (str, optional) – project identifier

• limit (int, optional) – limit number of fetched records

Returns:

pandas.DataFrame with listed connections

Return type:

pandas.DataFrame

Example:

client.export_assets.list()

start(meta_props, space_id=None, project_id=None)

Start the export. You must provide the space_id or the project_id. ALL_ASSETS is by default False. You don’t need to provide it unless it is set to True. You must provide one of the following in the meta_props: ALL_ASSETS, ASSET_TYPES, or ASSET_IDS. Only one of these can be provided.

In the meta_props:

ALL_ASSETS is a boolean. When set to True, it exports all assets in the given space. ASSET_IDS is an array that contains the list of assets IDs to be exported. ASSET_TYPES is used to provide the asset types to be exported. All assets of that asset type will be exported.

Eg: wml_model, wml_model_definition, wml_pipeline, wml_function, wml_experiment, software_specification, hardware_specification, package_extension, script

Parameters:

• meta_props (dict) – metadata, to see available meta names use client.export_assets.ConfigurationMetaNames.get()

• space_id (str, optional) – space identifier

• project_id – project identifier

Returns:

Response json

Return type:

dict

Example:

metadata = {
    client.export_assets.ConfigurationMetaNames.NAME: "export_model",
    client.export_assets.ConfigurationMetaNames.ASSET_IDS: ["13a53931-a8c0-4c2f-8319-c793155e7517",
                                                            "13a53931-a8c0-4c2f-8319-c793155e7518"]}

details = client.export_assets.start(meta_props=metadata, space_id="98a53931-a8c0-4c2f-8319-c793155e4598")

metadata = {
    client.export_assets.ConfigurationMetaNames.NAME: "export_model",
    client.export_assets.ConfigurationMetaNames.ASSET_TYPES: ["wml_model"]}

details = client.export_assets.start(meta_props=metadata, space_id="98a53931-a8c0-4c2f-8319-c793155e4598")

metadata = {
    client.export_assets.ConfigurationMetaNames.NAME: "export_model",
    client.export_assets.ConfigurationMetaNames.ALL_ASSETS: True}

details = client.export_assets.start(meta_props=metadata, space_id="98a53931-a8c0-4c2f-8319-c793155e4598")

class client.Import(client)

cancel(import_id, space_id=None, project_id=None)

Cancel an import job. You must provide the space_id or the project_id.

Note

To delete an import_id job, use delete() api

Parameters:

• import_id (str) – import the job identifier

• space_id (str, optional) – space identifier

• project_id (str, optional) – project identifier

Example:

client.import_assets.cancel(import_id='6213cf1-252f-424b-b52d-5cdd9814956c',
                            space_id='3421cf1-252f-424b-b52d-5cdd981495fe')

delete(import_id, space_id=None, project_id=None)

Deletes the given import_id job. You must provide the space_id or the project_id.

Parameters:

• import_id (str) – import the job identifier

• space_id (str, optional) – space identifier

• project_id (str, optional) – project identifier

Example:

client.import_assets.delete(import_id='6213cf1-252f-424b-b52d-5cdd9814956c',
                            space_id= '98a53931-a8c0-4c2f-8319-c793155e4598')

get_details(import_id=None, space_id=None, project_id=None, limit=None, asynchronous=False, get_all=False)

Get metadata of the given import job. If no import_id is specified, all import metadata is returned.

Parameters:

• import_id (str, optional) – import the job identifier

• space_id (str, optional) – space identifier

• project_id (str, optional) – project identifier

• limit (int, optional) – limit number of fetched records

• asynchronous (bool, optional) – if True, it will work as a generator

• get_all (bool, optional) – if True, it will get all entries in ‘limited’ chunks

Returns:

import(s) metadata

Return type:

dict (if import_id is not None) or {“resources”: [dict]} (if import_id is None)

Example:

details = client.import_assets.get_details(import_id)
details = client.import_assets.get_details()
details = client.import_assets.get_details(limit=100)
details = client.import_assets.get_details(limit=100, get_all=True)
details = []
for entry in client.import_assets.get_details(limit=100, asynchronous=True, get_all=True):
    details.extend(entry)

static get_id(import_details)

Get ID of the import job from import details.

Parameters:

import_details (dict) – metadata of the import job

Returns:

ID of the import job

Return type:

str

Example:

id = client.import_assets.get_id(import_details)

list(space_id=None, project_id=None, limit=None)

Return import jobs in a table format. If limit is set to None, only the first 50 records are shown.

Parameters:

• space_id (str, optional) – space identifier

• project_id (str, optional) – project identifier

• limit (int, optional) – limit number of fetched records

Returns:

pandas.DataFrame with listed assets

Return type:

pandas.DataFrame

Example:

client.import_assets.list()

start(file_path, space_id=None, project_id=None)

Start the import. You must provide the space_id or the project_id.

Parameters:

• file_path (str) – file path to the zip file with exported assets

• space_id (str, optional) – space identifier

• project_id (str, optional) – project identifier

Returns:

response json

Return type:

dict

Example:

details = client.import_assets.start(space_id="98a53931-a8c0-4c2f-8319-c793155e4598",
                                     file_path="/home/user/data_to_be_imported.zip")

Factsheets (IBM Cloud only)

Warning! Not supported for IBM Cloud Pak® for Data.

class client.Factsheets(client)

Link WML Model to Model Entry.

list_model_entries(catalog_id=None)

Return all WKC Model Entry assets for a catalog.

Parameters:

catalog_id (str, optional) – catalog ID where you want to register model. If no catalog_id is provided, WKC Model Entry assets from all catalogs are listed.

Returns:

all WKC Model Entry assets for a catalog

Return type:

dict

Example:

model_entries = client.factsheets.list_model_entries(catalog_id)

register_model_entry(model_id, meta_props, catalog_id=None)

Link WML Model to Model Entry

Parameters:

• model_id (str) – ID of the published model/asset

• meta_props (dict[str, str]) –

  metaprops, to see the available list of meta names use:

  client.factsheets.ConfigurationMetaNames.get()
  

• catalog_id (str, optional) – catalog ID where you want to register model

Returns:

metadata of the registration

Return type:

dict

Example:

meta_props = {
    client.factsheets.ConfigurationMetaNames.ASSET_ID: '83a53931-a8c0-4c2f-8319-c793155e7517'}

registration_details = client.factsheets.register_model_entry(model_id, catalog_id, meta_props)

or

meta_props = {
    client.factsheets.ConfigurationMetaNames.NAME: "New model entry",
    client.factsheets.ConfigurationMetaNames.DESCRIPTION: "New model entry"}

registration_details = client.factsheets.register_model_entry(model_id, meta_props)

unregister_model_entry(asset_id, catalog_id=None)

Unregister WKC Model Entry

Parameters:

• asset_id (str) – ID of the WKC model entry

• catalog_id (str, optional) – catalog ID where the asset is stored, when not provided, default client space or project will be taken

Example:

model_entries = client.factsheets.unregister_model_entry(asset_id='83a53931-a8c0-4c2f-8319-c793155e7517',
                                                         catalog_id='34553931-a8c0-4c2f-8319-c793155e7517')

or

client.set.default_space('98f53931-a8c0-4c2f-8319-c793155e7517')
model_entries = client.factsheets.unregister_model_entry(asset_id='83a53931-a8c0-4c2f-8319-c793155e7517')

class metanames.FactsheetsMetaNames

Set of MetaNames for Factsheets metanames.

Available MetaNames:

MetaName	Type	Required	Example value
ASSET_ID	str	N	13a53931-a8c0-4c2f-8319-c793155e7517
NAME	str	N	New model entry
DESCRIPTION	str	N	New model entry
MODEL_ENTRY_CATALOG_ID	str	Y	13a53931-a8c0-4c2f-8319-c793155e7517

Hardware specifications

class client.HwSpec(client)

Store and manage hardware specs.

ConfigurationMetaNames = <ibm_watsonx_ai.metanames.HwSpecMetaNames object>

MetaNames for Hardware Specification.

delete(hw_spec_id)

Delete a hardware specification.

Parameters:

hw_spec_id (str) – unique ID of the hardware specification to be deleted

Returns:

status (“SUCCESS” or “FAILED”)

Return type:

str

get_details(hw_spec_id=None, **kwargs)

Get hardware specification details.

Parameters:

hw_spec_id (str) – unique ID of the hardware spec

Returns:

metadata of the hardware specifications

Return type:

dict

Example:

hw_spec_details = client.hardware_specifications.get_details(hw_spec_uid)

static get_href(hw_spec_details)

Get the URL of hardware specifications.

Parameters:

hw_spec_details (dict) – details of the hardware specifications

Returns:

href of the hardware specifications

Return type:

str

Example:

hw_spec_details = client.hw_spec.get_details(hw_spec_id)
hw_spec_href = client.hw_spec.get_href(hw_spec_details)

static get_id(hw_spec_details)

Get the ID of a hardware specifications asset.

Parameters:

hw_spec_details (dict) – metadata of the hardware specifications

Returns:

unique ID of the hardware specifications

Return type:

str

Example:

asset_id = client.hardware_specifications.get_id(hw_spec_details)

get_id_by_name(hw_spec_name)

Get the unique ID of a hardware specification for the given name.

Parameters:

hw_spec_name (str) – name of the hardware specification

Returns:

unique ID of the hardware specification

Return type:

str

Example:

asset_id = client.hardware_specifications.get_id_by_name(hw_spec_name)

static get_uid(hw_spec_details)

Get the UID of a hardware specifications asset.

Deprecated: Use get_id(hw_spec_details) instead.

Parameters:

hw_spec_details (dict) – metadata of the hardware specifications

Returns:

unique ID of the hardware specifications

Return type:

str

Example:

asset_uid = client.hardware_specifications.get_uid(hw_spec_details)

get_uid_by_name(hw_spec_name)

Get the unique ID of a hardware specification for the given name.

Deprecated: Use get_id_by_name(hw_spec_name) instead.

Parameters:

hw_spec_name (str) – name of the hardware specification

Returns:

unique ID of the hardware specification

Return type:

str

Example:

asset_uid = client.hardware_specifications.get_uid_by_name(hw_spec_name)

list(name=None, limit=None)

List hardware specifications in a table format.

Parameters:

• name (str, optional) – unique ID of the hardware spec

• limit (int, optional) – limit number of fetched records

Returns:

pandas.DataFrame with listed hardware specifications

Return type:

pandas.DataFrame

Example:

client.hardware_specifications.list()

store(meta_props)

Create a hardware specification.

Parameters:

meta_props (dict) –

metadata of the hardware specification configuration. To see available meta names, use:

client.hardware_specifications.ConfigurationMetaNames.get()

Returns:

metadata of the created hardware specification

Return type:

dict

Example:

meta_props = {
    client.hardware_specifications.ConfigurationMetaNames.NAME: "custom hardware specification",
    client.hardware_specifications.ConfigurationMetaNames.DESCRIPTION: "Custom hardware specification creted with SDK",
    client.hardware_specifications.ConfigurationMetaNames.NODES:{"cpu":{"units":"2"},"mem":{"size":"128Gi"},"gpu":{"num_gpu":1}}
 }

client.hardware_specifications.store(meta_props)

class metanames.HwSpecMetaNames

Set of MetaNames for Hardware Specifications Specs.

Available MetaNames:

MetaName	Type	Required	Example value
NAME	str	Y	Custom Hardware Specification
DESCRIPTION	str	N	my_description
NODES	dict	N	{}
SPARK	dict	N	{}
DATASTAGE	dict	N	{}

Helpers

class ibm_watsonx_ai.helpers.helpers.get_credentials_from_config(env_name, credentials_name, config_path='./config.ini')

Bases:

Load credentials from the config file.

[DEV_LC]

credentials = { }
cos_credentials = { }

Parameters:

• env_name (str) – name of [ENV] defined in the config file

• credentials_name (str) – name of credentials

• config_path (str) – path to the config file

Returns:

loaded credentials

Return type:

dict

Example:

get_credentials_from_config(env_name='DEV_LC', credentials_name='credentials')

Model definitions

class client.ModelDefinition(client)

Store and manage model definitions.

ConfigurationMetaNames = <ibm_watsonx_ai.metanames.ModelDefinitionMetaNames object>

MetaNames for model definition creation.

create_revision(model_definition_id=None, **kwargs)

Create a revision for the given model definition. Revisions are immutable once created. The metadata and attachment of the model definition is taken and a revision is created out of it.

Parameters:

model_definition_id (str) – ID of the model definition

Returns:

revised metadata of the stored model definition

Return type:

dict

Example:

model_definition_revision = client.model_definitions.create_revision(model_definition_id)

delete(model_definition_id=None, **kwargs)

Delete a stored model definition.

Parameters:

model_definition_id (str) – unique ID of the stored model definition

Returns:

status (“SUCCESS” or “FAILED”)

Return type:

str

Example:

client.model_definitions.delete(model_definition_id)

download(model_definition_id, filename=None, rev_id=None, **kwargs)

Download the content of a model definition asset.

Parameters:

• model_definition_id (str) – unique ID of the model definition asset to be downloaded

• filename (str) – filename to be used for the downloaded file

• rev_id (str, optional) – revision ID

Returns:

path to the downloaded asset content

Return type:

str

Example:

client.model_definitions.download(model_definition_id, "model_definition_file")

get_details(model_definition_id=None, limit=None, get_all=None, **kwargs)

Get metadata of a stored model definition. If no model_definition_id is passed, details for all model definitions are returned.

Parameters:

• model_definition_id (str, optional) – unique ID of the model definition

• limit (int, optional) – limit number of fetched records

• get_all (bool, optional) – if True, it will get all entries in ‘limited’ chunks

Returns:

metadata of model definition

Return type:

dict (if model_definition_id is not None)

Example:

get_href(model_definition_details)

Get the href of a stored model definition.

Parameters:

model_definition_details (dict) – details of the stored model definition

Returns:

href of the stored model definition

Return type:

str

Example:

model_definition_id = client.model_definitions.get_href(model_definition_details)

get_id(model_definition_details)

Get the unique ID of a stored model definition asset.

Parameters:

model_definition_details (dict) – metadata of the stored model definition asset

Returns:

unique ID of the stored model definition asset

Return type:

str

Example:

asset_id = client.model_definition.get_id(asset_details)

get_revision_details(model_definition_id=None, rev_id=None, **kwargs)

Get metadata of a model definition.

Parameters:

• model_definition_id (str) – ID of the model definition

• rev_id (str, optional) – ID of the revision. If this parameter is not provided, it returns the latest revision. If there is no latest revision, it returns an error.

Returns:

metadata of the stored model definition

Return type:

dict

Example:

script_details = client.model_definitions.get_revision_details(model_definition_id, rev_id)

get_uid(model_definition_details)

Get the UID of the stored model.

Deprecated: Use get_id(model_definition_details) instead.

Parameters:

model_definition_details (dict) – details of the stored model definition

Returns:

UID of the stored model definition

Return type:

str

Example:

model_definition_uid = client.model_definitions.get_uid(model_definition_details)

list(limit=None)

Return the stored model definition assets in a table format. If limit is set to None, only the first 50 records are shown.

Parameters:

limit (int, optional) – limit number of fetched records

Returns:

pandas.DataFrame with listed model definitions

Return type:

pandas.DataFrame

Example:

client.model_definitions.list()

list_revisions(model_definition_id=None, limit=None, **kwargs)

Return the stored model definition assets in a table format. If limit is set to None, only the first 50 records are shown.

Parameters:

• model_definition_id (str) – unique ID of the model definition

• limit (int, optional) – limit number of fetched records

Returns:

pandas.DataFrame with listed model definitions

Return type:

pandas.DataFrame

Example:

client.model_definitions.list_revisions()

store(model_definition, meta_props)

Create a model definition.

Parameters:

• meta_props (dict) –

  metadata of the model definition configuration. To see available meta names, use:

  client.model_definitions.ConfigurationMetaNames.get()
  

• model_definition (str) – path to the content file to be uploaded

Returns:

metadata of the created model definition

Return type:

dict

Example:

client.model_definitions.store(model_definition, meta_props)

update(model_definition_id, meta_props=None, file_path=None)

Update the model definition with metadata, attachment, or both.

Parameters:

• model_definition_id (str) – ID of the model definition

• meta_props (dict) – metadata of the model definition configuration to be updated

• file_path (str, optional) – path to the content file to be uploaded

Returns:

updated metadata of the model definition

Return type:

dict

Example:

model_definition_details = client.model_definition.update(model_definition_id, meta_props, file_path)

class metanames.ModelDefinitionMetaNames

Set of MetaNames for Model Definition.

Available MetaNames:

MetaName	Type	Required	Schema	Example value
NAME	str	Y		my_model_definition
DESCRIPTION	str	N		my model_definition
PLATFORM	dict	Y	{'name(required)': 'string', 'versions(required)': ['versions']}	{'name': 'python', 'versions': ['3.10']}
VERSION	str	Y		1.0
COMMAND	str	N		python3 convolutional_network.py
CUSTOM	dict	N		{'field1': 'value1'}
SPACE_UID	str	N		3c1ce536-20dc-426e-aac7-7284cf3befc6

Package extensions

class client.PkgExtn(client)

Store and manage software Packages Extension specs.

ConfigurationMetaNames = <ibm_watsonx_ai.metanames.PkgExtnMetaNames object>

MetaNames for Package Extensions creation.

delete(pkg_extn_id)

Delete a package extension.

Parameters:

pkg_extn_id (str) – unique ID of the package extension

Returns:

status (“SUCCESS” or “FAILED”) if deleted synchronously or dictionary with response

Return type:

str or dict

Example:

client.package_extensions.delete(pkg_extn_id)

download(pkg_extn_id, filename)

Download a package extension.

Parameters:

• pkg_extn_id (str) – unique ID of the package extension to be downloaded

• filename (str) – filename to be used for the downloaded file

Returns:

path to the downloaded package extension content

Return type:

str

Example:

client.package_extensions.download(pkg_extn_id,"sample_conda.yml/custom_library.zip")

get_details(pkg_extn_id)

Get package extensions details.

Parameters:

pkg_extn_id (str) – unique ID of the package extension

Returns:

details of the package extension

Return type:

dict

Example:

pkg_extn_details = client.pkg_extn.get_details(pkg_extn_id)

static get_href(pkg_extn_details)

Get the URL of a stored package extension.

Parameters:

pkg_extn_details (dict) – details of the package extension

Returns:

href of the package extension

Return type:

str

Example:

pkg_extn_details = client.package_extensions.get_details(pkg_extn_id)
pkg_extn_href = client.package_extensions.get_href(pkg_extn_details)

static get_id(pkg_extn_details)

Get the unique ID of a package extension.

Parameters:

pkg_extn_details (dict) – details of the package extension

Returns:

unique ID of the package extension

Return type:

str

Example:

asset_id = client.package_extensions.get_id(pkg_extn_details)

get_id_by_name(pkg_extn_name)

Get the ID of a package extension.

Parameters:

pkg_extn_name (str) – name of the package extension

Returns:

unique ID of the package extension

Return type:

str

Example:

asset_id = client.package_extensions.get_id_by_name(pkg_extn_name)

list()

List the package extensions in a table format.

Returns:

pandas.DataFrame with listed package extensions

Return type:

pandas.DataFrame

client.package_extensions.list()

store(meta_props, file_path)

Create a package extension.

Parameters:

• meta_props (dict) –

  metadata of the package extension. To see available meta names, use:

  client.package_extensions.ConfigurationMetaNames.get()
  

• file_path (str) – path to the file to be uploaded as a package extension

Returns:

metadata of the package extension

Return type:

dict

Example:

meta_props = {
    client.package_extensions.ConfigurationMetaNames.NAME: "skl_pipeline_heart_problem_prediction",
    client.package_extensions.ConfigurationMetaNames.DESCRIPTION: "description scikit-learn_0.20",
    client.package_extensions.ConfigurationMetaNames.TYPE: "conda_yml"
}

pkg_extn_details = client.package_extensions.store(meta_props=meta_props, file_path="/path/to/file")

class metanames.PkgExtnMetaNames

Set of MetaNames for Package Extensions Specs.

Available MetaNames:

MetaName	Type	Required	Example value
NAME	str	Y	Python 3.10 with pre-installed ML package
DESCRIPTION	str	N	my_description
TYPE	str	Y	conda_yml/custom_library

Parameter Sets

class client.ParameterSets(client)

Store and manage parameter sets.

ConfigurationMetaNames = <ibm_watsonx_ai.metanames.ParameterSetsMetaNames object>

MetaNames for Parameter Sets creation.

create(meta_props)

Create a parameter set.

Parameters:

meta_props (dict) –

metadata of the space configuration. To see available meta names, use:

client.parameter_sets.ConfigurationMetaNames.get()

Returns:

metadata of the stored parameter set

Return type:

dict

Example:

meta_props = {
    client.parameter_sets.ConfigurationMetaNames.NAME: "Example name",
    client.parameter_sets.ConfigurationMetaNames.DESCRIPTION: "Example description",
    client.parameter_sets.ConfigurationMetaNames.PARAMETERS: [
        {
            "name": "string",
            "description": "string",
            "prompt": "string",
            "type": "string",
            "subtype": "string",
            "value": "string",
            "valid_values": [
                "string"
            ]
        }
    ],
    client.parameter_sets.ConfigurationMetaNames.VALUE_SETS: [
        {
            "name": "string",
            "values": [
                {
                    "name": "string",
                    "value": "string"
                }
            ]
        }
    ]
}

parameter_sets_details = client.parameter_sets.create(meta_props)

delete(parameter_set_id)

Delete a parameter set.

Parameters:

parameter_set_id (str) – unique ID of the parameter set

Returns:

status (“SUCCESS” or “FAILED”)

Return type:

str

Example:

client.parameter_sets.delete(parameter_set_id)

get_details(parameter_set_id=None)

Get parameter set details. If no parameter_sets_id is passed, details for all parameter sets are returned.

Parameters:

parameter_set_id (str, optional) – ID of the software specification

Returns:

metadata of the stored parameter set(s)

Return type:

• dict - if parameter_set_id is not None

• {“parameter_sets”: [dict]} - if parameter_set_id is None

Examples

If parameter_set_id is None:

parameter_sets_details = client.parameter_sets.get_details()

If parameter_set_id is given:

parameter_sets_details = client.parameter_sets.get_details(parameter_set_id)

get_id_by_name(parameter_set_name)

Get the unique ID of a parameter set.

Parameters:

parameter_set_name (str) – name of the parameter set

Returns:

unique ID of the parameter set

Return type:

str

Example:

asset_id = client.parameter_sets.get_id_by_name(parameter_set_name)

list(limit=None)

List parameter sets in a table format.

Parameters:

limit (int, optional) – limit number of fetched records

Returns:

pandas.DataFrame with listed parameter sets

Return type:

pandas.DataFrame

Example:

client.parameter_sets.list()

update(parameter_set_id, new_data, file_path)

Update parameter sets.

Parameters:

• parameter_set_id (str) – unique ID of the parameter sets

• new_data (str, list) – new data for parameters

• file_path (str) – path to update

Returns:

metadata of the updated parameter sets

Return type:

dict

Example for description

new_description_data = "New description"
parameter_set_details = client.parameter_sets.update(parameter_set_id, new_description_data, "description")

Example for parameters

new_parameters_data = [
    {
        "name": "string",
        "description": "new_description",
        "prompt": "new_string",
        "type": "new_string",
        "subtype": "new_string",
        "value": "new_string",
        "valid_values": [
            "new_string"
        ]
    }
]
parameter_set_details = client.parameter_sets.update(parameter_set_id, new_parameters_data, "parameters")

Example for value_sets

new_value_sets_data = [
    {
        "name": "string",
        "values": [
            {
                "name": "string",
                "value": "new_string"
            }
        ]
    }
]
parameter_set_details = client.parameter_sets.update_value_sets(parameter_set_id, new_value_sets_data, "value_sets")

class metanames.ParameterSetsMetaNames

Set of MetaNames for Parameter Sets metanames.

Available MetaNames:

MetaName	Type	Required	Example value
NAME	str	Y	sample name
DESCRIPTION	str	N	sample description
PARAMETERS	list	Y	[{'name': 'string', 'description': 'string', 'prompt': 'string', 'type': 'string', 'subtype': 'string', 'value': 'string', 'valid_values': ['string']}]
VALUE_SETS	list	N	[{'name': 'string', 'values': [{'name': 'string', 'value': 'string'}]}]

Repository

class client.Repository(client)

Store and manage models, functions, spaces, pipelines, and experiments using the Watson Machine Learning Repository.

To view ModelMetaNames, use:

client.repository.ModelMetaNames.show()

To view ExperimentMetaNames, use:

client.repository.ExperimentMetaNames.show()

To view FunctionMetaNames, use:

client.repository.FunctionMetaNames.show()

To view PipelineMetaNames, use:

client.repository.PipelineMetaNames.show()

To view AIServiceMetaNames, use:

client.repository.AIServiceMetaNames.show()

class ModelAssetTypes(DO_DOCPLEX_20_1='do-docplex_20.1', DO_OPL_20_1='do-opl_20.1', DO_CPLEX_20_1='do-cplex_20.1', DO_CPO_20_1='do-cpo_20.1', DO_DOCPLEX_22_1='do-docplex_22.1', DO_OPL_22_1='do-opl_22.1', DO_CPLEX_22_1='do-cplex_22.1', DO_CPO_22_1='do-cpo_22.1', WML_HYBRID_0_1='wml-hybrid_0.1', PMML_4_2_1='pmml_4.2.1', PYTORCH_ONNX_1_12='pytorch-onnx_1.12', PYTORCH_ONNX_RT22_2='pytorch-onnx_rt22.2', PYTORCH_ONNX_2_0='pytorch-onnx_2.0', PYTORCH_ONNX_RT23_1='pytorch-onnx_rt23.1', SCIKIT_LEARN_1_1='scikit-learn_1.1', MLLIB_3_3='mllib_3.3', SPSS_MODELER_17_1='spss-modeler_17.1', SPSS_MODELER_18_1='spss-modeler_18.1', SPSS_MODELER_18_2='spss-modeler_18.2', TENSORFLOW_2_9='tensorflow_2.9', TENSORFLOW_RT22_2='tensorflow_rt22.2', TENSORFLOW_2_12='tensorflow_2.12', TENSORFLOW_RT23_1='tensorflow_rt23.1', XGBOOST_1_6='xgboost_1.6', PROMPT_TUNE_1_0='prompt_tune_1.0', CUSTOM_FOUNDATION_MODEL_1_0='custom_foundation_model_1.0')

Data class with supported model asset types.

create_ai_service_revision(ai_service_id, **kwargs)

Create a new AI service revision.

Parameters:

ai_service_id (str) – unique ID of the AI service

Returns:

revised metadata of the stored AI service

Return type:

dict

Example:

client._ai_service.create_revision(ai_service_id)

create_experiment_revision(experiment_id)

Create a new experiment revision.

Parameters:

experiment_id (str) – unique ID of the stored experiment

Returns:

new revision details of the stored experiment

Return type:

dict

Example:

experiment_revision_artifact = client.repository.create_experiment_revision(experiment_id)

create_function_revision(function_id=None, **kwargs)

Create a new function revision.

Parameters:

function_id (str) – unique ID of the function

Returns:

revised metadata of the stored function

Return type:

dict

Example:

client.repository.create_function_revision(pipeline_id)

create_model_revision(model_id=None, **kwargs)

Create a revision for a given model ID.

Parameters:

model_id (str) – ID of the stored model

Returns:

revised metadata of the stored model

Return type:

dict

Example:

model_details = client.repository.create_model_revision(model_id)

create_pipeline_revision(pipeline_id=None, **kwargs)

Create a new pipeline revision.

Parameters:

pipeline_id (str) – unique ID of the pipeline

Returns:

details of the pipeline revision

Return type:

dict

Example:

client.repository.create_pipeline_revision(pipeline_id)

create_revision(artifact_id=None, **kwargs)

Create a revision for passed artifact_id.

Parameters:

artifact_id (str) – unique ID of a stored model, experiment, function, or pipelines

Returns:

artifact new revision metadata

Return type:

dict

Example:

details = client.repository.create_revision(artifact_id)

delete(artifact_id=None, **kwargs)

Delete a model, experiment, pipeline, function, or AI service from the repository.

Parameters:

artifact_id (str) – unique ID of the stored model, experiment, function, pipeline, or AI service

Returns:

status “SUCCESS” if deletion is successful

Return type:

Literal[“SUCCESS”]

Example:

client.repository.delete(artifact_id)

download(artifact_id=None, filename='downloaded_artifact.tar.gz', rev_id=None, format=None, **kwargs)

Download the configuration file for an artifact with the specified ID.

Parameters:

• artifact_id (str) – unique ID of the model or function

• filename (str, optional) – name of the file to which the artifact content will be downloaded

• rev_id (str, optional) – revision ID

• format (str, optional) – format of the content, applicable for models

Returns:

path to the downloaded artifact content

Return type:

str

Examples

client.repository.download(model_id, 'my_model.tar.gz')
client.repository.download(model_id, 'my_model.json') # if original model was saved as json, works only for xgboost 1.3

get_ai_service_details(ai_service_id=None, limit=None, asynchronous=False, get_all=False, spec_state=None, ai_service_name=None, **kwargs)

Get the metadata of AI service(s). If neither AI service ID nor AI service is specified, all AI service metadata is returned. If only AI service name is specified, metadata of AI services with the name is returned (if any).

Parameters:

• ai_service_id (str, optional) – ID of the AI service

• limit (int | None, optional) – limit number of fetched records

• asynchronous (bool, optional) – if True, it will work as a generator, defaults to False

• get_all (bool, optional) – if True, it will get all entries in ‘limited’ chunks, defaults to False

• spec_state (SpecStates | None, optional) – software specification state, can be used only when ai_service_id is None

• ai_service_name (str, optional) – name of the AI service, can be used only when ai_service_id is None

Returns:

metadata of the AI service

Return type:

dict (if ID is not None) or {“resources”: [dict]} (if ID is None)

Note

In the current implementation setting, spec_state=True might break the set limit and return less records than stated in the set limit.

Examples:

ai_service_details = client._ai_service.get_details(ai_service_id)
ai_service_details = client._ai_service.get_details(ai_service_name)
ai_service_details = client._ai_service.get_details()
ai_service_details = client._ai_service.get_details(limit=100)
ai_service_details = client._ai_service.get_details(limit=100, get_all=True)
ai_service_details = []
for entry in client._ai_service.get_details(limit=100, asynchronous=True, get_all=True):
    ai_service_details.extend(entry)

static get_ai_service_id(ai_service_details)

Get the ID of a stored AI service.

Parameters:

ai_service_details (dict) – metadata of the stored AI service

Returns:

ID of the stored AI service

Return type:

str

Example:

ai_service_details = client.repository.get_ai_service_details(ai_service_id)
ai_service_id = client.repository.get_ai_service_id(ai_service_details)

get_ai_service_revision_details(ai_service_id, rev_id, **kwargs)

Get the metadata of a specific revision of a stored AI service.

Parameters:

• ai_service_id (str) – definition of the stored AI service

• rev_id (str) – unique ID of the AI service revision

Returns:

metadata of the stored AI service revision

Return type:

dict

Example:

ai_service_revision_details = client.repository.get_ai_service_revision_details(ai_service_id, rev_id)

get_details(artifact_id=None, spec_state=None, artifact_name=None, **kwargs)

Get metadata of stored artifacts. If artifact_id and artifact_name are not specified, the metadata of all models, experiments, functions, pipelines, and ai services is returned. If only artifact_name is specified, metadata of all artifacts with the name is returned.

Parameters:

• artifact_id (str, optional) – unique ID of the stored model, experiment, function, or pipeline

• spec_state (SpecStates, optional) – software specification state, can be used only when artifact_id is None

• artifact_name (str, optional) – name of the stored model, experiment, function, pipeline, or ai service can be used only when artifact_id is None

Returns:

metadata of the stored artifact(s)

Return type:

• dict (if artifact_id is not None)

• {“models”: dict, “experiments”: dict, “pipeline”: dict, “functions”: dict, “ai_service”: dict} (if artifact_id is None)

Examples

details = client.repository.get_details(artifact_id)
details = client.repository.get_details(artifact_name='Sample_model')
details = client.repository.get_details()

Example of getting all repository assets with deprecated software specifications:

from ibm_watsonx_ai.lifecycle import SpecStates

details = client.repository.get_details(spec_state=SpecStates.DEPRECATED)

get_experiment_details(experiment_id=None, limit=None, asynchronous=False, get_all=False, experiment_name=None, **kwargs)

Get metadata of the experiment(s). If neither experiment ID nor experiment name is specified, all experiment metadata is returned. If only experiment name is specified, metadata of experiments with the name is returned (if any).

Parameters:

• experiment_id (str, optional) – ID of the experiment

• limit (int, optional) – limit number of fetched records

• asynchronous (bool, optional) – if True, it will work as a generator

• get_all (bool, optional) – if True, it will get all entries in ‘limited’ chunks

• experiment_name (str, optional) – name of the experiment, can be used only when experiment_id is None

Returns:

experiment metadata

Return type:

dict (if ID is not None) or {“resources”: [dict]} (if ID is None)

Example:

experiment_details = client.repository.get_experiment_details(experiment_id)
experiment_details = client.repository.get_experiment_details(experiment_name='Sample_experiment')
experiment_details = client.repository.get_experiment_details()
experiment_details = client.repository.get_experiment_details(limit=100)
experiment_details = client.repository.get_experiment_details(limit=100, get_all=True)
experiment_details = []
for entry in client.repository.get_experiment_details(limit=100, asynchronous=True, get_all=True):
    experiment_details.extend(entry)

static get_experiment_href(experiment_details)

Get the href of a stored experiment.

Parameters:

experiment_details (dict) – metadata of the stored experiment

Returns:

href of the stored experiment

Return type:

str

Example:

experiment_details = client.repository.get_experiment_details(experiment_id)
experiment_href = client.repository.get_experiment_href(experiment_details)

static get_experiment_id(experiment_details)

Get the unique ID of a stored experiment.

Parameters:

experiment_details (dict) – metadata of the stored experiment

Returns:

unique ID of the stored experiment

Return type:

str

Example:

experiment_details = client.repository.get_experiment_details(experiment_id)
experiment_id = client.repository.get_experiment_id(experiment_details)

get_experiment_revision_details(experiment_id, rev_id, **kwargs)

Get metadata of a stored experiments revisions.

Parameters:

• experiment_id (str) – ID of the stored experiment

• rev_id (str) – rev_id number of the stored experiment

Returns:

revision metadata of the stored experiment

Return type:

dict

Example:

experiment_details = client.repository.get_experiment_revision_details(experiment_id, rev_id)

get_function_details(function_id=None, limit=None, asynchronous=False, get_all=False, spec_state=None, function_name=None, **kwargs)

Get metadata of function(s). If neither function ID nor function name is specified, the metadata of all functions is returned. If only function name is specified, metadata of functions with the name is returned (if any).

Parameters:

• function_id – ID of the function

• limit (int, optional) – limit number of fetched records

• asynchronous (bool, optional) – if True, it will work as a generator

• get_all (bool, optional) – if True, it will get all entries in ‘limited’ chunks

• spec_state (SpecStates, optional) – software specification state, can be used only when function_id is None

• function_name (str, optional) – name of the function, can be used only when function_id is None

Type:

str, optional

Returns:

metadata of the function

Return type:

dict (if ID is not None) or {“resources”: [dict]} (if ID is None)

Note

In current implementation setting spec_state=True may break set limit, returning less records than stated by set limit.

Examples

function_details = client.repository.get_function_details(function_id)
function_details = client.repository.get_function_details(function_name='Sample_function')
function_details = client.repository.get_function_details()
function_details = client.repository.get_function_details(limit=100)
function_details = client.repository.get_function_details(limit=100, get_all=True)
function_details = []
for entry in client.repository.get_function_details(limit=100, asynchronous=True, get_all=True):
    function_details.extend(entry)

static get_function_href(function_details)

Get the URL of a stored function.

Parameters:

function_details (dict) – details of the stored function

Returns:

href of the stored function

Return type:

str

Example:

function_details = client.repository.get_function_details(function_id)
function_url = client.repository.get_function_href(function_details)

static get_function_id(function_details)

Get ID of stored function.

Parameters:

function_details (dict) – metadata of the stored function

Returns:

ID of stored function

Return type:

str

Example:

function_details = client.repository.get_function_details(function_id)
function_id = client.repository.get_function_id(function_details)

get_function_revision_details(function_id, rev_id, **kwargs)

Get metadata of a specific revision of a stored function.

Parameters:

• function_id (str) – definition of the stored function

• rev_id (str) – unique ID of the function revision

Returns:

stored function revision metadata

Return type:

dict

Example:

function_revision_details = client.repository.get_function_revision_details(function_id, rev_id)

get_id_by_name(artifact_name)

Get the ID of a stored artifact by name.

Parameters:

artifact_name (str) – name of the stored artifact

Returns:

ID of the stored artifact if exactly one with the ‘artifact_name’ exists. Otherwise, raise an error.

Return type:

str

Example:

artifact_id = client.repository.get_id_by_name(artifact_name)

get_model_details(model_id=None, limit=None, asynchronous=False, get_all=False, spec_state=None, model_name=None, **kwargs)

Get metadata of stored models. If neither model ID nor model name is specified, the metadata of all models is returned. If only model name is specified, metadata of models with the name is returned (if any).

Parameters:

• model_id (str, optional) – ID of the stored model, definition, or pipeline

• limit (int, optional) – limit number of fetched records

• asynchronous (bool, optional) – if True, it will work as a generator

• get_all (bool, optional) – if True, it will get all entries in ‘limited’ chunks

• spec_state (SpecStates, optional) – software specification state, can be used only when model_id is None

• model_name (str, optional) – name of the stored model, definition, or pipeline, can be used only when model_id is None

Returns:

metadata of the stored model(s)

Return type:

dict (if ID is not None) or {“resources”: [dict]} (if ID is None)

Note

In current implementation setting spec_state=True may break set limit, returning less records than stated by set limit.

Example:

model_details = client.repository.get_model_details(model_id)
models_details = client.repository.get_model_details(model_name='Sample_model')
models_details = client.repository.get_model_details()
models_details = client.repository.get_model_details(limit=100)
models_details = client.repository.get_model_details(limit=100, get_all=True)
models_details = []
for entry in client.repository.get_model_details(limit=100, asynchronous=True, get_all=True):
    models_details.extend(entry)

static get_model_href(model_details)

Get the URL of a stored model.

Parameters:

model_details (dict) – details of the stored model

Returns:

URL of the stored model

Return type:

str

Example:

model_url = client.repository.get_model_href(model_details)

static get_model_id(model_details)

Get the ID of a stored model.

Parameters:

model_details (dict) – details of the stored model

Returns:

ID of the stored model

Return type:

str

Example:

model_id = client.repository.get_model_id(model_details)

get_model_revision_details(model_id=None, rev_id=None, **kwargs)

Get metadata of a stored model’s specific revision.

Parameters:

• model_id (str) – ID of the stored model, definition, or pipeline

• rev_id (str) – unique ID of the stored model revision

Returns:

metadata of the stored model(s)

Return type:

dict

Example:

model_details = client.repository.get_model_revision_details(model_id, rev_id)

get_pipeline_details(pipeline_id=None, limit=None, asynchronous=False, get_all=False, pipeline_name=None, **kwargs)

Get metadata of stored pipeline(s). If neither pipeline ID nor pipeline name is specified, the metadata of all pipelines is returned. If only pipeline name is specified, metadata of pipelines with the name is returned (if any).

Parameters:

• pipeline_id (str, optional) – ID of the pipeline

• limit (int, optional) – limit number of fetched records

• asynchronous (bool, optional) – if True, it will work as a generator

• get_all (bool, optional) – if True, it will get all entries in ‘limited’ chunks

• pipeline_name (str, optional) – name of the pipeline, can be used only when pipeline_id is None

Returns:

metadata of pipeline(s)

Return type:

dict (if ID is not None) or {“resources”: [dict]} (if ID is None)

Example:

pipeline_details = client.repository.get_pipeline_details(pipeline_id)
pipeline_details = client.repository.get_pipeline_details(pipeline_name='Sample_pipeline')
pipeline_details = client.repository.get_pipeline_details()
pipeline_details = client.repository.get_pipeline_details(limit=100)
pipeline_details = client.repository.get_pipeline_details(limit=100, get_all=True)
pipeline_details = []
for entry in client.repository.get_pipeline_details(limit=100, asynchronous=True, get_all=True):
    pipeline_details.extend(entry)

static get_pipeline_href(pipeline_details)

Get the href from pipeline details.

Parameters:

pipeline_details (dict) – metadata of the stored pipeline

Returns:

href of the pipeline

Return type:

str

Example:

pipeline_details = client.repository.get_pipeline_details(pipeline_id)
pipeline_href = client.repository.get_pipeline_href(pipeline_details)

static get_pipeline_id(pipeline_details)

Get the pipeline ID from pipeline details.

Parameters:

pipeline_details (dict) – metadata of the stored pipeline

Returns:

unique ID of the pipeline

Return type:

str

Example:

pipeline_id = client.repository.get_pipeline_id(pipeline_details)

get_pipeline_revision_details(pipeline_id=None, rev_id=None, **kwargs)

Get metadata of a pipeline revision.

Parameters:

• pipeline_id (str) – ID of the stored pipeline

• rev_id (str) – revision ID of the stored pipeline

Returns:

revised metadata of the stored pipeline

Return type:

dict

Example:

pipeline_details = client.repository.get_pipeline_revision_details(pipeline_id, rev_id)

Note

rev_id parameter is not applicable in Cloud platform.

list(framework_filter=None)

Get and list stored models, pipelines, functions, experiments, and AI services in a table/DataFrame format. If limit is set to None, only the first 50 records are shown.

Parameters:

framework_filter (str, optional) – get only the frameworks with the desired names

Returns:

DataFrame with listed names and IDs of stored models

Return type:

pandas.DataFrame

Example:

client.repository.list()
client.repository.list(framework_filter='prompt_tune')

list_ai_service_revisions(ai_service_id, limit=None)

Print all revisions for a given AI service ID in a table format.

Parameters:

• ai_service_id (str) – unique ID of the stored AI service

• limit (int, optional) – limit number of fetched records

Returns:

pandas.DataFrame with listed revisions

Return type:

pandas.DataFrame

Example:

client.repository.list_ai_service_revisions(ai_service_id)

list_ai_services(limit=None)

Return stored AI services in a table format. If limit is set to None, only the first 50 records are shown.

Parameters:

limit (int, optional) – limit number of fetched records

Returns:

pandas.DataFrame with listed AI services

Return type:

pandas.DataFrame

Example:

client.repository.list_ai_services()

list_experiments(limit=None)

List stored experiments in a table format. If limit is set to None, only the first 50 records are shown.

Parameters:

limit (int, optional) – limit number of fetched records

Returns:

pandas.DataFrame with listed experiments

Return type:

pandas.DataFrame

Example:

client.repository.list_experiments()

list_experiments_revisions(experiment_id=None, limit=None, **kwargs)

Print all revisions for a given experiment ID in a table format.

Parameters:

• experiment_id (str) – unique ID of the stored experiment

• limit (int, optional) – limit number of fetched records

Returns:

pandas.DataFrame with listed revisions

Return type:

pandas.DataFrame

Example:

client.repository.list_experiments_revisions(experiment_id)

list_functions(limit=None)

Return stored functions in a table format. If limit is set to None, only the first 50 records are shown.

Parameters:

limit (int, optional) – limit number of fetched records

Returns:

pandas.DataFrame with listed functions

Return type:

pandas.DataFrame

Example:

client.repository.list_functions()

list_functions_revisions(function_id=None, limit=None, **kwargs)

Print all revisions for a given function ID in a table format.

Parameters:

• function_id (str) – unique ID of the stored function

• limit (int, optional) – limit number of fetched records

Returns:

pandas.DataFrame with listed revisions

Return type:

pandas.DataFrame

Example:

client.repository.list_functions_revisions(function_id)

list_models(limit=None, asynchronous=False, get_all=False)

List stored models in a table format. If limit is set to None, only the first 50 records are shown.

Parameters:

• limit (int, optional) – limit number of fetched records

• asynchronous (bool, optional) – if True, it will work as a generator

• get_all (bool, optional) – if True, it will get all entries in ‘limited’ chunks

Returns:

pandas.DataFrame with listed models or generator if asynchronous is set to True

Return type:

pandas.DataFrame | Generator

Example:

client.repository.list_models()
client.repository.list_models(limit=100)
client.repository.list_models(limit=100, get_all=True)
[entry for entry in client.repository.list_models(limit=100, asynchronous=True, get_all=True)]

list_models_revisions(model_id=None, limit=None, **kwargs)

Print all revisions for the given model ID in a table format.

Parameters:

• model_id (str) – unique ID of the stored model

• limit (int, optional) – limit number of fetched records

Returns:

pandas.DataFrame with listed revisions

Return type:

pandas.DataFrame

Example:

client.repository.list_models_revisions(model_id)

list_pipelines(limit=None)

List stored pipelines in a table format. If limit is set to None, only the first 50 records are shown.

Parameters:

limit (int, optional) – limit number of fetched records

Returns:

pandas.DataFrame with listed pipelines

Return type:

pandas.DataFrame

Example:

client.repository.list_pipelines()

list_pipelines_revisions(pipeline_id=None, limit=None, **kwargs)

List all revision for a given pipeline ID in a table format.

Parameters:

• pipeline_id (str) – unique ID of the stored pipeline

• limit (int, optional) – limit number of fetched records

Returns:

pandas.DataFrame with listed revisions

Return type:

pandas.DataFrame

Example:

client.repository.list_pipelines_revisions(pipeline_id)

load(artifact_id=None, **kwargs)

Load a model from the repository to object in a local environment.

Note

The use of the load() method is restricted and not permitted for AutoAI models.

Parameters:

artifact_id (str) – ID of the stored model

Returns:

trained model

Return type:

object

Example

model = client.repository.load(model_id)

promote_model(model_id, source_project_id, target_space_id)

Promote a model from a project to space. Supported only for IBM Cloud Pak® for Data.

Deprecated: Use client.spaces.promote(asset_id, source_project_id, target_space_id) instead.

store_ai_service(ai_service, meta_props)

Create an AI service asset.

You can use one of the following as an ai_service:

• filepath to gz file

• generator function that takes no argument or arguments that all have primitive python default values, and returns a generate function.

Parameters:

• ai_service (str | Callable) – path to a file with an archived AI service function’s content or a generator function (as described above)

• meta_props (dict) – metadata for storing an AI service asset. To see available meta names use client.repository.AIServiceMetaNames.show()

Returns:

metadata of the stored AI service

Return type:

dict

Examples:

The most simple use of an AI service is:

documentation_request = {
    "application/json": {
        "$schema": "http://json-schema.org/draft-07/schema#",
        "type": "object",
        "properties": {
            "query": {"type": "string"},
            "parameters": {
                "properties": {
                    "max_new_tokens": {"type": "integer"},
                    "top_p": {"type": "number"},
                },
                "required": ["max_new_tokens", "top_p"],
            },
        },
        "required": ["query"],
    }
}

documentation_response = {
    "application/json": {
        "$schema": "http://json-schema.org/draft-07/schema#",
        "type": "object",
        "properties": {
            "query": {"type": "string"},
            "result": {"type": "string"}
        },
        "required": ["query", "result"],
    }
}

meta_props = {
    client.repository.AIServiceMetaNames.NAME: "AI service example",
    client.repository.AIServiceMetaNames.DESCRIPTION: "This is AI service function",
    client.repository.AIServiceMetaNames.SOFTWARE_SPEC_ID: "53dc4cf1-252f-424b-b52d-5cdd9814987f",
    client.repository.AIServiceMetaNames.REQUEST_DOCUMENTATION: request_documentation,
    client.repository.AIServiceMetaNames.RESPONSE_DOCUMENTATION: response_documentation
    }

def deployable_ai_service(context, params={"k1":"v1"}, **kwargs):

    # imports
    from ibm_watsonx_ai import Credentials
    from ibm_watsonx_ai.foundation_models import ModelInference

    task_token = context.generate_token()

    outer_context = context
    url = "https://us-south.ml.cloud.ibm.com"
    project_id = "53dc4cf1-252f-424b-b52d-5cdd9814987f"

    def generate(context):
        task_token = outer_context.generate_token()
        payload = context.get_json()

        model = ModelInference(
            model_id="google/flan-t5-xl",
            credentials=Credentials(
                            url=url,
                            token=task_token
                            ),
            project_id=project_id)

        response = model.generate_text(payload['query'])
        response_body = {'query': payload['query'],
                         'result': response}

        return {'body': response_body}

    return generate

stored_ai_service_details = client.repository.store_ai_service(deployable_ai_service, meta_props)

store_experiment(meta_props)

Create an experiment.

Parameters:

meta_props (dict) –

metadata of the experiment configuration. To see available meta names, use:

client.repository.ExperimentMetaNames.get()

Returns:

metadata of the stored experiment

Return type:

dict

Example:

metadata = {
    client.repository.ExperimentMetaNames.NAME: 'my_experiment',
    client.repository.ExperimentMetaNames.EVALUATION_METRICS: ['accuracy'],
    client.repository.ExperimentMetaNames.TRAINING_REFERENCES: [
        {'pipeline': {'href': pipeline_href_1}},
        {'pipeline': {'href':pipeline_href_2}}
    ]
}
experiment_details = client.repository.store_experiment(meta_props=metadata)
experiment_href = client.repository.get_experiment_href(experiment_details)

store_function(function, meta_props)

Create a function.

As a ‘function’ may be used one of the following:

• filepath to gz file

• ‘score’ function reference, where the function is the function which will be deployed

• generator function, which takes no argument or arguments which all have primitive python default values and as result return ‘score’ function

Parameters:

• function (str or function) – path to file with archived function content or function (as described above)

• meta_props (str or dict) – meta data or name of the function, to see available meta names use client.repository.FunctionMetaNames.show()

Returns:

stored function metadata

Return type:

dict

Examples

The most simple use is (using score function):

meta_props = {
    client.repository.FunctionMetaNames.NAME: "function",
    client.repository.FunctionMetaNames.DESCRIPTION: "This is ai function",
    client.repository.FunctionMetaNames.SOFTWARE_SPEC_UID: "53dc4cf1-252f-424b-b52d-5cdd9814987f"}

def score(payload):
    values = [[row[0]*row[1]] for row in payload['values']]
    return {'fields': ['multiplication'], 'values': values}

stored_function_details = client.repository.store_function(score, meta_props)

Other, more interesting example is using generator function. In this situation it is possible to pass some variables:

creds = {...}

def gen_function(credentials=creds, x=2):
    def f(payload):
        values = [[row[0]*row[1]*x] for row in payload['values']]
        return {'fields': ['multiplication'], 'values': values}
    return f

stored_function_details = client.repository.store_function(gen_function, meta_props)

store_model(model=None, meta_props=None, training_data=None, training_target=None, pipeline=None, feature_names=None, label_column_names=None, subtrainingId=None, round_number=None, experiment_metadata=None, training_id=None)

Create a model.

Here you can explore how to save external models in correct format.

Parameters:

• model (str (for filename or path) or object (corresponding to model type)) –

  Can be one of following:

  • The train model object:

    • scikit-learn

    • xgboost

    • spark (PipelineModel)

  • path to saved model in format:

    • tensorflow / keras (.tar.gz)

    • pmml (.xml)

    • scikit-learn (.tar.gz)

    • spss (.str)

    • spark (.tar.gz)

    • xgboost (.tar.gz)

  • directory containing model file(s):

    • scikit-learn

    • xgboost

    • tensorflow

  • unique ID of the trained model

• meta_props (dict, optional) –

  metadata of the models configuration. To see available meta names, use:

  client.repository.ModelMetaNames.get()
  

• training_data (spark dataframe, pandas dataframe, numpy.ndarray or array, optional) – Spark DataFrame supported for spark models. Pandas dataframe, numpy.ndarray or array supported for scikit-learn models

• training_target (array, optional) – array with labels required for scikit-learn models

• pipeline (object, optional) – pipeline required for spark mllib models

• feature_names (numpy.ndarray or list, optional) – feature names for the training data in case of Scikit-Learn/XGBoost models, this is applicable only in the case where the training data is not of type - pandas.DataFrame

• label_column_names (numpy.ndarray or list, optional) – label column names of the trained Scikit-Learn/XGBoost models

• round_number (int, optional) – round number of a Federated Learning experiment that has been configured to save intermediate models, this applies when model is a training id

• experiment_metadata (dict, optional) – metadata retrieved from the experiment that created the model

• training_id (str, optional) – Run id of AutoAI or TuneExperiment experiment.

Returns:

metadata of the created model

Return type:

dict

Note

• For a keras model, model content is expected to contain a .h5 file and an archived version of it.

• feature_names is an optional argument containing the feature names for the training data in case of Scikit-Learn/XGBoost models. Valid types are numpy.ndarray and list. This is applicable only in the case where the training data is not of type - pandas.DataFrame.

• If the training_data is of type pandas.DataFrame and feature_names are provided, feature_names are ignored.

• For INPUT_DATA_SCHEMA meta prop use list even when passing single input data schema. You can provide multiple schemas as dictionaries inside a list.

Examples

stored_model_details = client.repository.store_model(model, name)

In more complicated cases you should create proper metadata, similar to this one:

sw_spec_id = client.software_specifications.get_id_by_name('scikit-learn_0.23-py3.7')

metadata = {
    client.repository.ModelMetaNames.NAME: 'customer satisfaction prediction model',
    client.repository.ModelMetaNames.SOFTWARE_SPEC_ID: sw_spec_id,
    client.repository.ModelMetaNames.TYPE: 'scikit-learn_0.23'
}

In case when you want to provide input data schema of the model, you can provide it as part of meta:

sw_spec_id = client.software_specifications.get_id_by_name('spss-modeler_18.1')

metadata = {
    client.repository.ModelMetaNames.NAME: 'customer satisfaction prediction model',
    client.repository.ModelMetaNames.SOFTWARE_SPEC_ID: sw_spec_id,
    client.repository.ModelMetaNames.TYPE: 'spss-modeler_18.1',
    client.repository.ModelMetaNames.INPUT_DATA_SCHEMA: [{'id': 'test',
                                                          'type': 'list',
                                                          'fields': [{'name': 'age', 'type': 'float'},
                                                                     {'name': 'sex', 'type': 'float'},
                                                                     {'name': 'fbs', 'type': 'float'},
                                                                     {'name': 'restbp', 'type': 'float'}]
                                                          },
                                                          {'id': 'test2',
                                                           'type': 'list',
                                                           'fields': [{'name': 'age', 'type': 'float'},
                                                                      {'name': 'sex', 'type': 'float'},
                                                                      {'name': 'fbs', 'type': 'float'},
                                                                      {'name': 'restbp', 'type': 'float'}]
    }]
}

store_model() method used with a local tar.gz file that contains a model:

stored_model_details = client.repository.store_model(path_to_tar_gz, meta_props=metadata, training_data=None)

store_model() method used with a local directory that contains model files:

stored_model_details = client.repository.store_model(path_to_model_directory, meta_props=metadata, training_data=None)

store_model() method used with the ID of a trained model:

stored_model_details = client.repository.store_model(trained_model_id, meta_props=metadata, training_data=None)

store_model() method used with a pipeline that was generated by an AutoAI experiment:

metadata = {
    client.repository.ModelMetaNames.NAME: 'AutoAI prediction model stored from object'
}
stored_model_details = client.repository.store_model(pipeline_model, meta_props=metadata, experiment_metadata=experiment_metadata)

metadata = {
    client.repository.ModelMetaNames.NAME: 'AutoAI prediction Pipeline_1 model'
}
stored_model_details = client.repository.store_model(model="Pipeline_1", meta_props=metadata, training_id = training_id)

Example of storing a prompt tuned model:

stored_model_details = client.repository.store_model(training_id = prompt_tuning_run_id)

Example of storing a custom foundation model:

sw_spec_id = client.software_specifications.get_id_by_name('watsonx-cfm-caikit-1.0')

metadata = {
    client.repository.ModelMetaNames.NAME: 'custom FM asset',
    client.repository.ModelMetaNames.SOFTWARE_SPEC_ID: sw_spec_id,
    client.repository.ModelMetaNames.TYPE: client.repository.ModelAssetTypes.CUSTOM_FOUNDATION_MODEL_1_0
}
stored_model_details = client.repository.store_model(model='mistralai/Mistral-7B-Instruct-v0.2', meta_props=metadata)

store_pipeline(meta_props)

Create a pipeline.

Parameters:

meta_props (dict) –

metadata of the pipeline configuration. To see available meta names, use:

client.repository.PipelineMetaNames.get()

Returns:

stored pipeline metadata

Return type:

dict

Example:

metadata = {
    client.repository.PipelineMetaNames.NAME: 'my_training_definition',
    client.repository.PipelineMetaNames.DOCUMENT: {"doc_type":"pipeline",
                                                       "version": "2.0",
                                                       "primary_pipeline": "dlaas_only",
                                                       "pipelines": [{"id": "dlaas_only",
                                                                      "runtime_ref": "hybrid",
                                                                      "nodes": [{"id": "training",
                                                                                 "type": "model_node",
                                                                                 "op": "dl_train",
                                                                                 "runtime_ref": "DL",
                                                                                 "inputs": [],
                                                                                 "outputs": [],
                                                                                 "parameters": {"name": "tf-mnist",
                                                                                                "description": "Simple MNIST model implemented in TF",
                                                                                                "command": "python3 convolutional_network.py --trainImagesFile ${DATA_DIR}/train-images-idx3-ubyte.gz --trainLabelsFile ${DATA_DIR}/train-labels-idx1-ubyte.gz --testImagesFile ${DATA_DIR}/t10k-images-idx3-ubyte.gz --testLabelsFile ${DATA_DIR}/t10k-labels-idx1-ubyte.gz --learningRate 0.001 --trainingIters 6000",
                                                                                                "compute": {"name": "k80","nodes": 1},
                                                                                                "training_lib_href": "/v4/libraries/64758251-bt01-4aa5-a7ay-72639e2ff4d2/content"
                                                                                 },
                                                                                 "target_bucket": "wml-dev-results"
                                                                      }]
                                                       }]
    }
}
pipeline_details = client.repository.store_pipeline(training_definition_filepath, meta_props=metadata)

update_ai_service(ai_service_id, changes, update_ai_service=None)

Updates existing AI service asset metadata.

Parameters:

• ai_service_id (str) – ID of AI service to be updated

• changes (dict) – elements that will be changed, where keys are AIServiceMetaNames

• update_ai_service – path to the file with an archived AI service function’s content or function that will be changed for a specific ai_service_id

Example:

metadata = {
    client.repository.AIServiceMetaNames.NAME: "updated_ai_service"
}

ai_service_details = client.repository.update_ai_service(ai_service_id, changes=metadata)

update_experiment(experiment_id=None, changes=None, **kwargs)

Updates existing experiment metadata.

Parameters:

• experiment_id (str) – ID of the experiment with the definition to be updated

• changes (dict) – elements to be changed, where keys are ExperimentMetaNames

Returns:

metadata of the updated experiment

Return type:

dict

Example:

metadata = {
    client.repository.ExperimentMetaNames.NAME: "updated_exp"
}
exp_details = client.repository.update_experiment(experiment_id, changes=metadata)

update_function(function_id, changes=None, update_function=None, **kwargs)

Updates existing function metadata.

Parameters:

• function_id (str) – ID of function which define what should be updated

• changes (dict) – elements which should be changed, where keys are FunctionMetaNames

• update_function (str or function, optional) – path to file with archived function content or function which should be changed for specific function_id, this parameter is valid only for CP4D 3.0.0

Example:

metadata = {
    client.repository.FunctionMetaNames.NAME: "updated_function"
}

function_details = client.repository.update_function(function_id, changes=metadata)

update_model(model_id=None, updated_meta_props=None, update_model=None, **kwargs)

Update an existing model.

Parameters:

• model_id (str) – ID of model to be updated

• updated_meta_props (dict, optional) – new set of updated_meta_props to be updated

• update_model (object or model, optional) – archived model content file or path to directory that contains the archived model file that needs to be changed for the specific model_id

Returns:

updated metadata of the model

Return type:

dict

Example:

model_details = client.repository.update_model(model_id, update_model=updated_content)

update_pipeline(pipeline_id=None, changes=None, rev_id=None, **kwargs)

Update metadata of an existing pipeline.

Parameters:

• pipeline_id (str) – unique ID of the pipeline to be updated

• changes (dict) – elements to be changed, where keys are PipelineMetaNames

• rev_id (str) – revision ID of the pipeline

Returns:

metadata of the updated pipeline

Return type:

dict

Example:

metadata = {
    client.repository.PipelineMetaNames.NAME: "updated_pipeline"
}
pipeline_details = client.repository.update_pipeline(pipeline_id, changes=metadata)

class metanames.ModelMetaNames

Set of MetaNames for models.

Available MetaNames:

MetaName	Type	Required	Schema	Example value
NAME	str	Y		my_model
DESCRIPTION	str	N		my_description
INPUT_DATA_SCHEMA	list	N	{'id(required)': 'string', 'fields(required)': [{'name(required)': 'string', 'type(required)': 'string', 'nullable(optional)': 'string'}]}	{'id': '1', 'type': 'struct', 'fields': [{'name': 'x', 'type': 'double', 'nullable': False, 'metadata': {}}, {'name': 'y', 'type': 'double', 'nullable': False, 'metadata': {}}]}
TRAINING_DATA_REFERENCES	list	N	[{'name(optional)': 'string', 'type(required)': 'string', 'connection(required)': {'endpoint_url(required)': 'string', 'access_key_id(required)': 'string', 'secret_access_key(required)': 'string'}, 'location(required)': {'bucket': 'string', 'path': 'string'}, 'schema(optional)': {'id(required)': 'string', 'fields(required)': [{'name(required)': 'string', 'type(required)': 'string', 'nullable(optional)': 'string'}]}}]	[]
TEST_DATA_REFERENCES	list	N	[{'name(optional)': 'string', 'type(required)': 'string', 'connection(required)': {'endpoint_url(required)': 'string', 'access_key_id(required)': 'string', 'secret_access_key(required)': 'string'}, 'location(required)': {'bucket': 'string', 'path': 'string'}, 'schema(optional)': {'id(required)': 'string', 'fields(required)': [{'name(required)': 'string', 'type(required)': 'string', 'nullable(optional)': 'string'}]}}]	[]
OUTPUT_DATA_SCHEMA	dict	N	{'id(required)': 'string', 'fields(required)': [{'name(required)': 'string', 'type(required)': 'string', 'nullable(optional)': 'string'}]}	{'id': '1', 'type': 'struct', 'fields': [{'name': 'x', 'type': 'double', 'nullable': False, 'metadata': {}}, {'name': 'y', 'type': 'double', 'nullable': False, 'metadata': {}}]}
LABEL_FIELD	str	N		PRODUCT_LINE
TRANSFORMED_LABEL_FIELD	str	N		PRODUCT_LINE_IX
TAGS	list	N	['string', 'string']	['string', 'string']
SIZE	dict	N	{'in_memory(optional)': 'string', 'content(optional)': 'string'}	{'in_memory': 0, 'content': 0}
PIPELINE_ID	str	N		53628d69-ced9-4f43-a8cd-9954344039a8
RUNTIME_ID	str	N		53628d69-ced9-4f43-a8cd-9954344039a8
TYPE	str	Y		mllib_2.1
CUSTOM	dict	N		{}
DOMAIN	str	N		Watson Machine Learning
HYPER_PARAMETERS	dict	N
METRICS	list	N
IMPORT	dict	N	{'name(optional)': 'string', 'type(required)': 'string', 'connection(required)': {'endpoint_url(required)': 'string', 'access_key_id(required)': 'string', 'secret_access_key(required)': 'string'}, 'location(required)': {'bucket': 'string', 'path': 'string'}}	{'connection': {'endpoint_url': 'https://s3-api.us-geo.objectstorage.softlayer.net', 'access_key_id': '***', 'secret_access_key': '***'}, 'location': {'bucket': 'train-data', 'path': 'training_path'}, 'type': 's3'}
TRAINING_LIB_ID	str	N		53628d69-ced9-4f43-a8cd-9954344039a8
MODEL_DEFINITION_ID	str	N		53628d6_cdee13-35d3-s8989343
SOFTWARE_SPEC_ID	str	N		53628d69-ced9-4f43-a8cd-9954344039a8
TF_MODEL_PARAMS	dict	N		{'save_format': 'None', 'signatures': 'struct', 'options': 'None', 'custom_objects': 'string'}
FAIRNESS_INFO	dict	N		{'favorable_labels': ['X']}
MODEL_LOCATION	dict	N		{'connection_id': '53628d69-ced9-4f43-a8cd-9954344039a8', 'bucket': 'cos_sample_bucket', 'file_path': 'path/to/model/on/cos'}
FRAMEWORK	str	N		custom_foundation_model
VERSION	str	N		1.0

Note: project (MetaNames.PROJECT_ID) and space (MetaNames.SPACE_ID) meta names are not supported and considered as invalid. Instead use client.set.default_space(<SPACE_ID>) to set the space or client.set.default_project(<PROJECT_ID>).

class metanames.ExperimentMetaNames

Set of MetaNames for experiments.

Available MetaNames:

MetaName	Type	Required	Schema	Example value
NAME	str	Y		Hand-written Digit Recognition
DESCRIPTION	str	N		Hand-written Digit Recognition training
TAGS	list	N	[{'value(required)': 'string', 'description(optional)': 'string'}]	[{'value': 'dsx-project.<project-guid>', 'description': 'DSX project guid'}]
EVALUATION_METHOD	str	N		multiclass
EVALUATION_METRICS	list	N	[{'name(required)': 'string', 'maximize(optional)': 'boolean'}]	[{'name': 'accuracy', 'maximize': False}]
TRAINING_REFERENCES	list	Y	[{'pipeline(optional)': {'href(required)': 'string', 'data_bindings(optional)': [{'data_reference(required)': 'string', 'node_id(required)': 'string'}], 'nodes_parameters(optional)': [{'node_id(required)': 'string', 'parameters(required)': 'dict'}]}, 'training_lib(optional)': {'href(required)': 'string', 'compute(optional)': {'name(required)': 'string', 'nodes(optional)': 'number'}, 'runtime(optional)': {'href(required)': 'string'}, 'command(optional)': 'string', 'parameters(optional)': 'dict'}}]	[{'pipeline': {'href': '/v4/pipelines/6d758251-bb01-4aa5-a7a3-72339e2ff4d8'}}]
SPACE_UID	str	N		3c1ce536-20dc-426e-aac7-7284cf3befc6
LABEL_COLUMN	str	N		label
CUSTOM	dict	N		{'field1': 'value1'}

class metanames.FunctionMetaNames

Set of MetaNames for AI functions.

Available MetaNames:

MetaName	Type	Required	Schema	Example value
NAME	str	Y		ai_function
DESCRIPTION	str	N		This is ai function
SOFTWARE_SPEC_ID	str	N		53628d69-ced9-4f43-a8cd-9954344039a8
SOFTWARE_SPEC_UID	str	N		53628d69-ced9-4f43-a8cd-9954344039a8
INPUT_DATA_SCHEMAS	list	N	[{'id(required)': 'string', 'fields(required)': [{'name(required)': 'string', 'type(required)': 'string', 'nullable(optional)': 'string'}]}]	[{'id': '1', 'type': 'struct', 'fields': [{'name': 'x', 'type': 'double', 'nullable': False, 'metadata': {}}, {'name': 'y', 'type': 'double', 'nullable': False, 'metadata': {}}]}]
OUTPUT_DATA_SCHEMAS	list	N	[{'id(required)': 'string', 'fields(required)': [{'name(required)': 'string', 'type(required)': 'string', 'nullable(optional)': 'string'}]}]	[{'id': '1', 'type': 'struct', 'fields': [{'name': 'multiplication', 'type': 'double', 'nullable': False, 'metadata': {}}]}]
TAGS	list	N	['string']	['tags1', 'tags2']
TYPE	str	N		python
CUSTOM	dict	N		{}
SAMPLE_SCORING_INPUT	dict	N	{'id(optional)': 'string', 'fields(optional)': 'array', 'values(optional)': 'array'}	{'input_data': [{'fields': ['name', 'age', 'occupation'], 'values': [['john', 23, 'student'], ['paul', 33, 'engineer']]}]}

class metanames.PipelineMetanames

Set of MetaNames for pipelines.

Available MetaNames:

MetaName	Type	Required	Schema	Example value
NAME	str	Y		Hand-written Digit Recognitionu
DESCRIPTION	str	N		Hand-written Digit Recognition training
SPACE_ID	str	N		3c1ce536-20dc-426e-aac7-7284cf3befc6
SPACE_UID	str	N		3c1ce536-20dc-426e-aac7-7284cf3befc6
TAGS	list	N	[{'value(required)': 'string', 'description(optional)': 'string'}]	[{'value': 'dsx-project.<project-guid>', 'description': 'DSX project guid'}]
DOCUMENT	dict	N	{'doc_type(required)': 'string', 'version(required)': 'string', 'primary_pipeline(required)': 'string', 'pipelines(required)': [{'id(required)': 'string', 'runtime_ref(required)': 'string', 'nodes(required)': [{'id': 'string', 'type': 'string', 'inputs': 'list', 'outputs': 'list', 'parameters': {'training_lib_href': 'string'}}]}]}	{'doc_type': 'pipeline', 'version': '2.0', 'primary_pipeline': 'dlaas_only', 'pipelines': [{'id': 'dlaas_only', 'runtime_ref': 'hybrid', 'nodes': [{'id': 'training', 'type': 'model_node', 'op': 'dl_train', 'runtime_ref': 'DL', 'inputs': [], 'outputs': [], 'parameters': {'name': 'tf-mnist', 'description': 'Simple MNIST model implemented in TF', 'command': 'python3 convolutional_network.py --trainImagesFile ${DATA_DIR}/train-images-idx3-ubyte.gz --trainLabelsFile ${DATA_DIR}/train-labels-idx1-ubyte.gz --testImagesFile ${DATA_DIR}/t10k-images-idx3-ubyte.gz --testLabelsFile ${DATA_DIR}/t10k-labels-idx1-ubyte.gz --learningRate 0.001 --trainingIters 6000', 'compute': {'name': 'k80', 'nodes': 1}, 'training_lib_href': '/v4/libraries/64758251-bt01-4aa5-a7ay-72639e2ff4d2/content'}, 'target_bucket': 'wml-dev-results'}]}]}
CUSTOM	dict	N		{'field1': 'value1'}
IMPORT	dict	N	{'name(optional)': 'string', 'type(required)': 'string', 'connection(required)': {'endpoint_url(required)': 'string', 'access_key_id(required)': 'string', 'secret_access_key(required)': 'string'}, 'location(required)': {'bucket': 'string', 'path': 'string'}}	{'connection': {'endpoint_url': 'https://s3-api.us-geo.objectstorage.softlayer.net', 'access_key_id': '***', 'secret_access_key': '***'}, 'location': {'bucket': 'train-data', 'path': 'training_path'}, 'type': 's3'}
RUNTIMES	list	N		[{'id': 'id', 'name': 'tensorflow', 'version': '1.13-py3'}]
COMMAND	str	N		convolutional_network.py --trainImagesFile train-images-idx3-ubyte.gz --trainLabelsFile train-labels-idx1-ubyte.gz --testImagesFile t10k-images-idx3-ubyte.gz --testLabelsFile t10k-labels-idx1-ubyte.gz --learningRate 0.001 --trainingIters 6000
COMPUTE	dict	N		{'name': 'k80', 'nodes': 1}

class metanames.AIServiceMetaNames

Set of MetaNames for AI services.

Available MetaNames:

MetaName	Type	Required	Schema	Example value
NAME	str	Y		ai_service
DESCRIPTION	str	N		This is AI service
SOFTWARE_SPEC_ID	str	N		53628d69-ced9-4f43-a8cd-9954344039a8
REQUEST_DOCUMENTATION	dict	N		{'application/json': {'$schema': 'http://json-schema.org/draft-07/schema#', 'type': 'object', 'properties': {'query': {'type': 'string'}, 'parameters': {'properties': {'max_new_tokens': {'type': 'integer'}, 'top_p': {'type': 'number'}}, 'required': ['max_new_tokens', 'top_p']}}, 'required': ['query']}}
RESPONSE_DOCUMENTATION	dict	N		{'application/json': {'$schema': 'http://json-schema.org/draft-07/schema#', 'type': 'object', 'properties': {'query': {'type': 'string'}, 'result': {'type': 'string'}}, 'required': ['query', 'result']}}
TAGS	list	N	['string']	['tags1', 'tags2']
CODE_TYPE	str	N		python
CUSTOM	dict	N		{'key1': 'value1'}

Script

class client.Script(client)

Store and manage script assets.

ConfigurationMetaNames = <ibm_watsonx_ai.metanames.ScriptMetaNames object>

MetaNames for script assets creation.

create_revision(script_id=None, **kwargs)

Create a revision for the given script. Revisions are immutable once created. The metadata and attachment at script_id is taken and a revision is created out of it.

Parameters:

script_id (str) – ID of the script

Returns:

revised metadata of the stored script

Return type:

dict

Example:

script_revision = client.script.create_revision(script_id)

delete(asset_id=None, **kwargs)

Delete a stored script asset.

Parameters:

asset_id (str) – ID of the script asset

Returns:

status (“SUCCESS” or “FAILED”) if deleted synchronously or dictionary with response

Return type:

str | dict

Example:

client.script.delete(asset_id)

download(asset_id=None, filename=None, rev_id=None, **kwargs)

Download the content of a script asset.

Parameters:

• asset_id (str) – unique ID of the script asset to be downloaded

• filename (str) – filename to be used for the downloaded file

• rev_id (str, optional) – revision ID

Returns:

path to the downloaded asset content

Return type:

str

Example:

client.script.download(asset_id, "script_file")

get_details(script_id=None, limit=None, get_all=None, **kwargs)

Get script asset details. If no script_id is passed, details for all script assets are returned.

Parameters:

• script_id (str, optional) – unique ID of the script

• limit (int, optional) – limit number of fetched records

• get_all (bool, optional) – if True, it will get all entries in ‘limited’ chunks

Returns:

metadata of the stored script asset

Return type:

• dict - if script_id is not None

• {“resources”: [dict]} - if script_id is None

Example:

script_details = client.script.get_details(script_id)

static get_href(asset_details)

Get the URL of a stored script asset.

Parameters:

asset_details (dict) – details of the stored script asset

Returns:

href of the stored script asset

Return type:

str

Example:

asset_details = client.script.get_details(asset_id)
asset_href = client.script.get_href(asset_details)

static get_id(asset_details)

Get the unique ID of a stored script asset.

Parameters:

asset_details (dict) – metadata of the stored script asset

Returns:

unique ID of the stored script asset

Return type:

str

Example:

asset_id = client.script.get_id(asset_details)

get_revision_details(script_id=None, rev_id=None, **kwargs)

Get metadata of the script revision.

Parameters:

• script_id (str) – ID of the script

• rev_id (str, optional) – ID of the revision. If this parameter is not provided, it returns the latest revision. If there is no latest revision, it returns an error.

Returns:

metadata of the stored script(s)

Return type:

list

Example:

script_details = client.script.get_revision_details(script_id, rev_id)

list(limit=None)

List stored scripts in a table format. If limit is set to None, only the first 50 records are shown.

Parameters:

limit (int, optional) – limit number of fetched records

Returns:

pandas.DataFrame with listed scripts

Return type:

pandas.DataFrame

Example:

client.script.list()

list_revisions(script_id=None, limit=None, **kwargs)

Print all revisions for the given script ID in a table format.

Parameters:

• script_id (str) – ID of the stored script

• limit (int, optional) – limit number of fetched records

Returns:

pandas.DataFrame with listed revisions

Return type:

pandas.DataFrame

Example:

client.script.list_revisions(script_id)

store(meta_props, file_path)

Create a script asset and upload content to it.

Parameters:

• meta_props (dict) – name to be given to the script asset

• file_path (str) – path to the content file to be uploaded

Returns:

metadata of the stored script asset

Return type:

dict

Example:

metadata = {
    client.script.ConfigurationMetaNames.NAME: 'my first script',
    client.script.ConfigurationMetaNames.DESCRIPTION: 'description of the script',
    client.script.ConfigurationMetaNames.SOFTWARE_SPEC_ID: '0cdb0f1e-5376-4f4d-92dd-da3b69aa9bda'
}

asset_details = client.script.store(meta_props=metadata, file_path="/path/to/file")

update(script_id=None, meta_props=None, file_path=None, **kwargs)

Update a script with metadata, attachment, or both.

Parameters:

• script_id (str) – ID of the script

• meta_props (dict, optional) – changes for the script matadata

• file_path (str, optional) – file path to the new attachment

Returns:

updated metadata of the script

Return type:

dict

Example:

script_details = client.script.update(script_id, meta, content_path)

class metanames.ScriptMetaNames

Set of MetaNames for Script Specifications.

Available MetaNames:

MetaName	Type	Required	Example value
NAME	str	Y	Python script
DESCRIPTION	str	N	my_description
SOFTWARE_SPEC_ID	str	Y	53628d69-ced9-4f43-a8cd-9954344039a8

Service instance

class client.ServiceInstance(client)

Connect, get details, and check usage of a Watson Machine Learning service instance.

get_api_key()

Get the API key of a Watson Machine Learning service.

Returns:

API key

Return type:

str

Example:

instance_details = client.service_instance.get_api_key()

get_details()

Get information about the Watson Machine Learning instance.

Returns:

metadata of the service instance

Return type:

dict

Example:

instance_details = client.service_instance.get_details()

get_instance_id()

Get the instance ID of a Watson Machine Learning service.

Returns:

ID of the instance

Return type:

str

Example:

instance_details = client.service_instance.get_instance_id()

get_password()

Get the password for the Watson Machine Learning service. Applicable only for IBM Cloud Pak® for Data.

Returns:

password

Return type:

str

Example:

instance_details = client.service_instance.get_password()

get_url()

Get the instance URL of a Watson Machine Learning service.

Returns:

URL of the instance

Return type:

str

Example:

instance_details = client.service_instance.get_url()

get_username()

Get the username for the Watson Machine Learning service. Applicable only for IBM Cloud Pak® for Data.

Returns:

username

Return type:

str

Example:

instance_details = client.service_instance.get_username()

Set

class client.Set(client)

Set a space_id or a project_id to be used in the subsequent actions.

default_project(project_id)

Set a project ID.

Parameters:

project_id (str) – ID of the project to be used

Returns:

status (“SUCCESS” if succeeded)

Return type:

str

Example:

client.set.default_project(project_id)

default_space(space_id=None, **kwargs)

Set a space ID.

Parameters:

space_id (str) – ID of the space to be used

Returns:

status (“SUCCESS” if succeeded)

Return type:

str

Example:

client.set.default_space(space_id)

Shiny (IBM Cloud Pak® for Data only)

Warning! Not supported for IBM Cloud.

class client.Shiny(client)

Store and manage shiny assets.

ConfigurationMetaNames = <ibm_watsonx_ai.metanames.ShinyMetaNames object>

MetaNames for Shiny Assets creation.

create_revision(shiny_id=None, **kwargs)

Create a revision for the given shiny asset. Revisions are immutable once created. The metadata and attachment at script_id is taken and a revision is created out of it.

Parameters:

shiny_id (str) – ID of the shiny asset

Returns:

revised metadata of the stored shiny asset

Return type:

dict

Example:

shiny_revision = client.shiny.create_revision(shiny_id)

delete(shiny_id=None, **kwargs)

Delete a stored shiny asset.

Parameters:

shiny_id (str) – unique ID of the shiny asset

Returns:

status (“SUCCESS” or “FAILED”) if deleted synchronously or dictionary with response

Return type:

str | dict

Example:

client.shiny.delete(shiny_id)

download(shiny_id=None, filename=None, rev_id=None, **kwargs)

Download the content of a shiny asset.

Parameters:

• shiny_id (str) – unique ID of the shiny asset to be downloaded

• filename (str) – filename to be used for the downloaded file

• rev_id (str, optional) – ID of the revision

Returns:

path to the downloaded shiny asset content

Return type:

str

Example:

client.shiny.download(shiny_id, "shiny_asset.zip")

get_details(shiny_id=None, limit=None, get_all=None, **kwargs)

Get shiny asset details. If no shiny_id is passed, details for all shiny assets are returned.

Parameters:

• shiny_id (str, optional) – unique ID of the shiny asset

• limit (int, optional) – limit number of fetched records

• get_all (bool, optional) – if True, it will get all entries in ‘limited’ chunks

Returns:

metadata of the stored shiny asset

Return type:

• dict - if shiny_id is not None

• {“resources”: [dict]} - if shiny_id is None

Example:

shiny_details = client.shiny.get_details(shiny_id)

static get_href(shiny_details)

Get the URL of a stored shiny asset.

Parameters:

shiny_details (dict) – details of the stored shiny asset

Returns:

href of the stored shiny asset

Return type:

str

Example:

shiny_details = client.shiny.get_details(shiny_id)
shiny_href = client.shiny.get_href(shiny_details)

static get_id(shiny_details)

Get the unique ID of a stored shiny asset.

Parameters:

shiny_details (dict) – metadata of the stored shiny asset

Returns:

unique ID of the stored shiny asset

Return type:

str

Example:

shiny_id = client.shiny.get_id(shiny_details)

get_revision_details(shiny_id=None, rev_id=None, **kwargs)

Get metadata of the shiny_id revision.

Parameters:

• shiny_id (str) – ID of the shiny asset

• rev_id (str, optional) – ID of the revision. If this parameter is not provided, it returns the latest revision. If there is no latest revision, it returns an error.

Returns:

stored shiny(s) metadata

Return type:

list

Example:

shiny_details = client.shiny.get_revision_details(shiny_id, rev_id)

static get_uid(shiny_details)

Get the Unique ID of a stored shiny asset.

Deprecated: Use get_id(shiny_details) instead.

Parameters:

shiny_details (dict) – metadata of the stored shiny asset

Returns:

unique ID of the stored shiny asset

Return type:

str

Example:

shiny_id = client.shiny.get_uid(shiny_details)

list(limit=None)

List stored shiny assets in a table format. If limit is set to None, only the first 50 records are shown.

Parameters:

limit (int, optional) – limit number of fetched records

Returns:

pandas.DataFrame with listed shiny assets

Return type:

pandas.DataFrame

Example:

client.shiny.list()

list_revisions(shiny_id=None, limit=None, **kwargs)

List all revisions for the given shiny asset ID in a table format.

Parameters:

• shiny_id (str) – ID of the stored shiny asset

• limit (int, optional) – limit number of fetched records

Returns:

pandas.DataFrame with listed shiny revisions

Return type:

pandas.DataFrame

Example:

client.shiny.list_revisions(shiny_id)

store(meta_props, file_path)

Create a shiny asset and upload content to it.

Parameters:

• meta_props (dict) – metadata of the shiny asset

• file_path (str) – path to the content file to be uploaded

Returns:

metadata of the stored shiny asset

Return type:

dict

Example:

meta_props = {
    client.shiny.ConfigurationMetaNames.NAME: "shiny app name"
}

shiny_details = client.shiny.store(meta_props, file_path="/path/to/file")

update(shiny_id=None, meta_props=None, file_path=None, **kwargs)

Update a shiny asset with metadata, attachment, or both.

Parameters:

• shiny_id (str) – ID of the shiny asset

• meta_props (dict, optional) – changes to the metadata of the shiny asset

• file_path (str, optional) – file path to the new attachment

Returns:

updated metadata of the shiny asset

Return type:

dict

Example:

script_details = client.script.update(shiny_id, meta, content_path)

Software specifications

class client.SwSpec(client)

Store and manage software specs.

ConfigurationMetaNames = <ibm_watsonx_ai.metanames.SwSpecMetaNames object>

MetaNames for Software Specification creation.

add_package_extension(sw_spec_id=None, pkg_extn_id=None, **kwargs)

Add a package extension to a software specification’s existing metadata.

Parameters:

• sw_spec_id (str) – unique ID of the software specification to be updated

• pkg_extn_id (str) – unique ID of the package extension to be added to the software specification

Returns:

status

Return type:

str

Example:

client.software_specifications.add_package_extension(sw_spec_id, pkg_extn_id)

delete(sw_spec_id=None, **kwargs)

Delete a software specification.

Parameters:

sw_spec_id (str) – unique ID of the software specification

Returns:

status (“SUCCESS” or “FAILED”)

Return type:

str

Example:

client.software_specifications.delete(sw_spec_id)

delete_package_extension(sw_spec_id=None, pkg_extn_id=None, **kwargs)

Delete a package extension from a software specification’s existing metadata.

Parameters:

• sw_spec_id (str) – unique ID of the software specification to be updated

• pkg_extn_id (str) – unique ID of the package extension to be deleted from the software specification

Returns:

status

Return type:

str

Example:

client.software_specifications.delete_package_extension(sw_spec_uid, pkg_extn_id)

get_details(sw_spec_id=None, state_info=False, **kwargs)

Get software specification details. If no sw_spec_id is passed, details for all software specifications are returned.

Parameters:

• sw_spec_id (bool) – ID of the software specification

• state_info – works only when sw_spec_id is None, instead of returning details of software specs, it returns the state of the software specs information (supported, unsupported, deprecated), containing suggested replacement in case of unsupported or deprecated software specs

Returns:

metadata of the stored software specification(s)

Return type:

• dict - if sw_spec_uid is not None

• {“resources”: [dict]} - if sw_spec_uid is None

Examples

sw_spec_details = client.software_specifications.get_details(sw_spec_uid)
sw_spec_details = client.software_specifications.get_details()
sw_spec_state_details = client.software_specifications.get_details(state_info=True)

static get_href(sw_spec_details)

Get the URL of a software specification.

Parameters:

sw_spec_details (dict) – details of the software specification

Returns:

href of the software specification

Return type:

str

Example:

sw_spec_details = client.software_specifications.get_details(sw_spec_id)
sw_spec_href = client.software_specifications.get_href(sw_spec_details)

static get_id(sw_spec_details)

Get the unique ID of a software specification.

Parameters:

sw_spec_details (dict) – metadata of the software specification

Returns:

unique ID of the software specification

Return type:

str

Example:

asset_id = client.software_specifications.get_id(sw_spec_details)

get_id_by_name(sw_spec_name)

Get the unique ID of a software specification.

Parameters:

sw_spec_name (str) – name of the software specification

Returns:

unique ID of the software specification

Return type:

str

Example:

asset_uid = client.software_specifications.get_id_by_name(sw_spec_name)

static get_uid(sw_spec_details)

Get the unique ID of a software specification.

Deprecated: Use get_id(sw_spec_details) instead.

Parameters:

sw_spec_details (dict) – metadata of the software specification

Returns:

unique ID of the software specification

Return type:

str

Example:

asset_uid = client.software_specifications.get_uid(sw_spec_details)

get_uid_by_name(sw_spec_name)

Get the unique ID of a software specification.

Deprecated: Use get_id_by_name(self, sw_spec_name) instead.

Parameters:

sw_spec_name (str) – name of the software specification

Returns:

unique ID of the software specification

Return type:

str

Example:

asset_uid = client.software_specifications.get_uid_by_name(sw_spec_name)

list(limit=None)

List software specifications in a table format.

Parameters:

limit (int, optional) – limit number of fetched records

Returns:

pandas.DataFrame with listed software specifications

Return type:

pandas.DataFrame

Example:

client.software_specifications.list()

store(meta_props)

Create a software specification.

Parameters:

meta_props (dict) –

metadata of the space configuration. To see available meta names, use:

client.software_specifications.ConfigurationMetaNames.get()

Returns:

metadata of the stored space

Return type:

dict

Example:

meta_props = {
    client.software_specifications.ConfigurationMetaNames.NAME: "skl_pipeline_heart_problem_prediction",
    client.software_specifications.ConfigurationMetaNames.DESCRIPTION: "description scikit-learn_0.20",
    client.software_specifications.ConfigurationMetaNames.PACKAGE_EXTENSIONS_UID: [],
    client.software_specifications.ConfigurationMetaNames.SOFTWARE_CONFIGURATIONS: {},
    client.software_specifications.ConfigurationMetaNames.BASE_SOFTWARE_SPECIFICATION_ID: "guid"
}

sw_spec_details = client.software_specifications.store(meta_props)

class metanames.SwSpecMetaNames

Set of MetaNames for Software Specifications Specs.

Available MetaNames:

MetaName	Type	Required	Schema	Example value
NAME	str	Y		Python 3.10 with pre-installed ML package
DESCRIPTION	str	N		my_description
PACKAGE_EXTENSIONS	list	N		[{'guid': 'value'}]
SOFTWARE_CONFIGURATION	dict	N	{'platform(required)': 'string'}	{'platform': {'name': 'python', 'version': '3.10'}}
BASE_SOFTWARE_SPECIFICATION	dict	Y		{'guid': 'BASE_SOFTWARE_SPECIFICATION_ID'}

Spaces

class client.Spaces(client)

Store and manage spaces.

ConfigurationMetaNames = <ibm_watsonx_ai.metanames.SpacesMetaNames object>

MetaNames for spaces creation.

MemberMetaNames = <ibm_watsonx_ai.metanames.SpacesMemberMetaNames object>

MetaNames for space members creation.

create_member(space_id, meta_props)

Create a member within a space.

Parameters:

• space_id (str) – ID of the space with the definition to be updated

• meta_props (dict) –

  metadata of the member configuration. To see available meta names, use:

  client.spaces.MemberMetaNames.get()
  

Returns:

metadata of the stored member

Return type:

dict

Note

• role can be any one of the following: “viewer”, “editor”, “admin”

• type can be any one of the following: “user”, “service”

• id can be one of the following: service-ID or IAM-userID

Examples

metadata = {
    client.spaces.MemberMetaNames.MEMBERS: [{"id":"IBMid-100000DK0B",
                                             "type": "user",
                                             "role": "admin" }]
}
members_details = client.spaces.create_member(space_id=space_id, meta_props=metadata)

metadata = {
    client.spaces.MemberMetaNames.MEMBERS: [{"id":"iam-ServiceId-5a216e59-6592-43b9-8669-625d341aca71",
                                             "type": "service",
                                             "role": "admin" }]
}
members_details = client.spaces.create_member(space_id=space_id, meta_props=metadata)

delete(space_id)

Delete a stored space.

Parameters:

space_id (str) – ID of the space

Returns:

status “SUCCESS” if deletion is successful

Return type:

Literal[“SUCCESS”]

Example:

client.spaces.delete(space_id)

delete_member(space_id, member_id)

Delete a member associated with a space.

Parameters:

• space_id (str) – ID of the space

• member_id (str) – ID of the member

Returns:

status (“SUCCESS” or “FAILED”)

Return type:

str

Example:

client.spaces.delete_member(space_id,member_id)

get_details(space_id=None, limit=None, asynchronous=False, get_all=False, space_name=None)

Get metadata of stored space(s).

Parameters:

• space_id (str, optional) – ID of the space

• limit (int, optional) – applicable when space_id is not provided, otherwise limit will be ignored

• asynchronous (bool, optional) – if True, it will work as a generator

• get_all (bool, optional) – if True, it will get all entries in ‘limited’ chunks

• space_name (str, optional) – name of the stored space, can be used only when space_id is None

Returns:

metadata of stored space(s)

Return type:

• dict - if space_id is not None

• {“resources”: [dict]} - if space_id is None

Example:

space_details = client.spaces.get_details(space_id)
space_details = client.spaces.get_details(space_name)
space_details = client.spaces.get_details(limit=100)
space_details = client.spaces.get_details(limit=100, get_all=True)
space_details = []
for entry in client.spaces.get_details(limit=100, asynchronous=True, get_all=True):
    space_details.extend(entry)

static get_id(space_details)

Get the space_id from the space details.

Parameters:

space_details (dict) – metadata of the stored space

Returns:

ID of the stored space

Return type:

str

Example:

space_details = client.spaces.store(meta_props)
space_id = client.spaces.get_id(space_details)

get_id_by_name(space_name)

Get the ID of a stored space by name.

Parameters:

space_name (str) – name of the stored space

Returns:

ID of the stored space

Return type:

str

Example:

space_id = client.spaces.get_id_by_name(space_name)

get_member_details(space_id, member_id)

Get metadata of a member associated with a space.

Parameters:

• space_id (str) – ID of that space with the definition to be updated

• member_id (str) – ID of the member

Returns:

metadata of the space member

Return type:

dict

Example:

member_details = client.spaces.get_member_details(space_id,member_id)

static get_uid(space_details)

Get the unique ID of the space.

Deprecated: Use get_id(space_details) instead.

param space_details:

metadata of the space

type space_details:

dict

return:

unique ID of the space

rtype:

str

Example:

space_details = client.spaces.store(meta_props)
space_uid = client.spaces.get_uid(space_details)

list(limit=None, member=None, roles=None, space_type=None)

List stored spaces in a table format. If limit is set to None, only the first 50 records are shown.

Parameters:

• limit (int, optional) – limit number of fetched records

• member (str, optional) – filters the result list, only includes spaces where the user with a matching user ID is a member

• roles (str, optional) – limit number of fetched records

• space_type (str, optional) – filter spaces by their type, available types are ‘wx’, ‘cpd’, and ‘wca’

Returns:

pandas.DataFrame with listed spaces

Return type:

pandas.DataFrame

Example:

client.spaces.list()

list_members(space_id, limit=None, identity_type=None, role=None, state=None)

Print the stored members of a space in a table format. If limit is set to None, only the first 50 records are shown.

Parameters:

• space_id (str) – ID of the space

• limit (int, optional) – limit number of fetched records

• identity_type (str, optional) – filter the members by type

• role (str, optional) – filter the members by role

• state (str, optional) – filter the members by state

Returns:

pandas.DataFrame with listed members

Return type:

pandas.DataFrame

Example:

client.spaces.list_members(space_id)

promote(asset_id, source_project_id, target_space_id, rev_id=None)

Promote an asset from a project to a space.

Parameters:

• asset_id (str) – ID of the stored asset

• source_project_id (str) – source project, from which the asset is promoted

• target_space_id (str) – target space, where the asset is promoted

• rev_id (str, optional) – revision ID of the promoted asset

Returns:

ID of the promoted asset

Return type:

str

Examples

promoted_asset_id = client.spaces.promote(asset_id, source_project_id=project_id, target_space_id=space_id)
promoted_model_id = client.spaces.promote(model_id, source_project_id=project_id, target_space_id=space_id)
promoted_function_id = client.spaces.promote(function_id, source_project_id=project_id, target_space_id=space_id)
promoted_data_asset_id = client.spaces.promote(data_asset_id, source_project_id=project_id, target_space_id=space_id)
promoted_connection_asset_id = client.spaces.promote(connection_id, source_project_id=project_id, target_space_id=space_id)

store(meta_props, background_mode=True)

Create a space. The instance associated with the space via COMPUTE will be used for billing purposes on the cloud. Note that STORAGE and COMPUTE are applicable only for cloud.

Parameters:

• meta_props (dict) –

  metadata of the space configuration. To see available meta names, use:

  client.spaces.ConfigurationMetaNames.get()
  

• background_mode (bool, optional) – indicator if store() method will run in background (async) or (sync)

Returns:

metadata of the stored space

Return type:

dict

Example:

metadata = {
    client.spaces.ConfigurationMetaNames.NAME: "my_space",
    client.spaces.ConfigurationMetaNames.DESCRIPTION: "spaces",
    client.spaces.ConfigurationMetaNames.STORAGE: {"resource_crn": "provide crn of the COS storage"},
    client.spaces.ConfigurationMetaNames.COMPUTE: {"name": "test_instance",
                                                   "crn": "provide crn of the instance"},
    client.spaces.ConfigurationMetaNames.STAGE: {"production": True,
                                                 "name": "stage_name"},
    client.spaces.ConfigurationMetaNames.TAGS: ["sample_tag_1", "sample_tag_2"],
    client.spaces.ConfigurationMetaNames.TYPE: "cpd",
}
spaces_details = client.spaces.store(meta_props=metadata)

update(space_id, changes)

Update existing space metadata. ‘STORAGE’ cannot be updated. STORAGE and COMPUTE are applicable only for cloud.

Parameters:

• space_id (str) – ID of the space with the definition to be updated

• changes (dict) – elements to be changed, where keys are ConfigurationMetaNames

Returns:

metadata of the updated space

Return type:

dict

Example:

metadata = {
    client.spaces.ConfigurationMetaNames.NAME:"updated_space",
    client.spaces.ConfigurationMetaNames.COMPUTE: {"name": "test_instance",
                                                   "crn": "v1:staging:public:pm-20-dev:us-south:a/09796a1b4cddfcc9f7fe17824a68a0f8:f1026e4b-77cf-4703-843d-c9984eac7272::"
    }
}
space_details = client.spaces.update(space_id, changes=metadata)

update_member(space_id, member_id, changes)

Update the metadata of an existing member.

Parameters:

• space_id (str) – ID of the space

• member_id (str) – ID of the member to be updated

• changes (dict) – elements to be changed, where keys are ConfigurationMetaNames

Returns:

metadata of the updated member

Return type:

dict

Example:

metadata = {
    client.spaces.MemberMetaNames.MEMBER: {"role": "editor"}
}
member_details = client.spaces.update_member(space_id, member_id, changes=metadata)

class metanames.SpacesMetaNames

Set of MetaNames for Platform Spaces Specs.

Available MetaNames:

MetaName	Type	Required	Example value
NAME	str	Y	my_space
DESCRIPTION	str	N	my_description
STORAGE	dict	N	{'type': 'bmcos_object_storage', 'resource_crn': '', 'delegated(optional)': 'false'}
COMPUTE	dict	N	{'name': 'name', 'crn': 'crn of the instance'}
STAGE	dict	N	{'production': True, 'name': 'name of the stage'}
TAGS	list	N	['sample_tag']
TYPE	str	N	cpd

class metanames.SpacesMemberMetaNames

Set of MetaNames for Platform Spaces Member Specs.

Available MetaNames:

MetaName	Type	Required	Schema	Example value
MEMBERS	list	N	[{'id(required)': 'string', 'role(required)': 'string', 'type(required)': 'string', 'state(optional)': 'string'}]	[{'id': 'iam-id1', 'role': 'editor', 'type': 'user', 'state': 'active'}, {'id': 'iam-id2', 'role': 'viewer', 'type': 'user', 'state': 'active'}]
MEMBER	dict	N		{'id': 'iam-id1', 'role': 'editor', 'type': 'user', 'state': 'active'}

Training

class client.Training(client)

Train new models.

cancel(training_id=None, hard_delete=False, **kwargs)

Cancel a training that is currently running. This method can delete metadata details of a completed or canceled training run when hard_delete parameter is set to True.

Parameters:

• training_id (str) – ID of the training

• hard_delete (bool, optional) –

  specify True or False:

  • True - to delete the completed or canceled training run

  • False - to cancel the currently running training run

Returns:

status “SUCCESS” if cancelation is successful

Return type:

Literal[“SUCCESS”]

Example:

client.training.cancel(training_id)

get_details(training_id=None, limit=None, asynchronous=False, get_all=False, training_type=None, state=None, tag_value=None, training_definition_id=None, _internal=False, **kwargs)

Get metadata of training(s). If training_id is not specified, the metadata of all model spaces are returned.

Parameters:

• training_id (str, optional) – unique ID of the training

• limit (int, optional) – limit number of fetched records

• asynchronous (bool, optional) – if True, it will work as a generator

• get_all (bool, optional) – if True, it will get all entries in ‘limited’ chunks

• training_type (str, optional) – filter the fetched list of trainings based on the training type [“pipeline” or “experiment”]

• state (str, optional) – filter the fetched list of training based on their state: [queued, running, completed, failed]

• tag_value (str, optional) – filter the fetched list of training based on their tag value

• training_definition_id (str, optional) – filter the fetched trainings that are using the given training definition

Returns:

metadata of training(s)

Return type:

• dict - if training_id is not None

• {“resources”: [dict]} - if training_id is None

Examples

training_run_details = client.training.get_details(training_id)
training_runs_details = client.training.get_details()
training_runs_details = client.training.get_details(limit=100)
training_runs_details = client.training.get_details(limit=100, get_all=True)
training_runs_details = []
for entry in client.training.get_details(limit=100, asynchronous=True, get_all=True):
    training_runs_details.extend(entry)

static get_href(training_details)

Get the training href from the training details.

Parameters:

training_details (dict) – metadata of the created training

Returns:

training href

Return type:

str

Example:

training_details = client.training.get_details(training_id)
run_url = client.training.get_href(training_details)

static get_id(training_details)

Get the training ID from the training details.

Parameters:

training_details (dict) – metadata of the created training

Returns:

unique ID of the training

Return type:

str

Example:

training_details = client.training.get_details(training_id)
training_id = client.training.get_id(training_details)

get_metrics(training_id=None, **kwargs)

Get metrics of a training run.

Parameters:

training_id (str) – ID of the training

Returns:

metrics of the training run

Return type:

list of dict

Example:

training_status = client.training.get_metrics(training_id)

get_status(training_id=None, **kwargs)

Get the status of a created training.

Parameters:

training_id (str) – ID of the training

Returns:

training_status

Return type:

dict

Example:

training_status = client.training.get_status(training_id)

list(limit=None, asynchronous=False, get_all=False)

List stored trainings in a table format. If limit is set to None, only the first 50 records are shown.

Parameters:

• limit (int, optional) – limit number of fetched records

• asynchronous (bool, optional) – if True, it will work as a generator

• get_all (bool, optional) – if True, it will get all entries in ‘limited’ chunks

Returns:

pandas.DataFrame with listed experiments

Return type:

pandas.DataFrame

Examples

client.training.list()
training_runs_df = client.training.list(limit=100)
training_runs_df = client.training.list(limit=100, get_all=True)
training_runs_df = []
for entry in client.training.list(limit=100, asynchronous=True, get_all=True):
    training_runs_df.extend(entry)

list_intermediate_models(training_id=None, **kwargs)

Print the intermediate_models in a table format.

Parameters:

training_id (str) – ID of the training

Note

This method is not supported for IBM Cloud Pak® for Data.

Example:

client.training.list_intermediate_models()

monitor_logs(training_id=None, **kwargs)

Print the logs of a training created.

Parameters:

training_id (str) – training ID

Note

This method is not supported for IBM Cloud Pak® for Data.

Example:

client.training.monitor_logs(training_id)

monitor_metrics(training_id=None, **kwargs)

Print the metrics of a created training.

Parameters:

training_id (str) – ID of the training

Note

This method is not supported for IBM Cloud Pak® for Data.

Example:

client.training.monitor_metrics(training_id)

run(meta_props, asynchronous=True, **kwargs)

Create a new Machine Learning training.

Parameters:

• meta_props (dict) –

  metadata of the training configuration. To see available meta names, use:

  client.training.ConfigurationMetaNames.show()
  

• asynchronous (bool, optional) –

  • True - training job is submitted and progress can be checked later

  • False - method will wait till job completion and print training stats

Returns:

metadata of the training created

Return type:

dict

Note

You can provide one of the following values for training:

• client.training.ConfigurationMetaNames.EXPERIMENT

• client.training.ConfigurationMetaNames.PIPELINE

• client.training.ConfigurationMetaNames.MODEL_DEFINITION

Examples

Example of meta_props for creating a training run in IBM Cloud Pak® for Data version 3.0.1 or above:

metadata = {
    client.training.ConfigurationMetaNames.NAME: 'Hand-written Digit Recognition',
    client.training.ConfigurationMetaNames.DESCRIPTION: 'Hand-written Digit Recognition Training',
    client.training.ConfigurationMetaNames.PIPELINE: {
        "id": "4cedab6d-e8e4-4214-b81a-2ddb122db2ab",
        "rev": "12",
        "model_type": "string",
        "data_bindings": [
            {
                "data_reference_name": "string",
                "node_id": "string"
            }
        ],
        "nodes_parameters": [
            {
                "node_id": "string",
                "parameters": {}
            }
        ],
        "hardware_spec": {
            "id": "4cedab6d-e8e4-4214-b81a-2ddb122db2ab",
            "rev": "12",
            "name": "string",
            "num_nodes": "2"
        }
    },
    client.training.ConfigurationMetaNames.TRAINING_DATA_REFERENCES: [{
        'type': 's3',
        'connection': {},
        'location': {'href': 'v2/assets/asset1233456'},
        'schema': { 'id': 't1', 'name': 'Tasks', 'fields': [ { 'name': 'duration', 'type': 'number' } ]}
    }],
    client.training.ConfigurationMetaNames.TRAINING_RESULTS_REFERENCE: {
        'id' : 'string',
        'connection': {
            'endpoint_url': 'https://s3-api.us-geo.objectstorage.service.networklayer.com',
            'access_key_id': '***',
            'secret_access_key': '***'
        },
        'location': {
            'bucket': 'wml-dev-results',
            'path' : "path"
        }
        'type': 's3'
    }
}

Example of a Federated Learning training job:

aggregator_metadata = {
    client.training.ConfigurationMetaNames.NAME: 'Federated_Learning_Tensorflow_MNIST',
    client.training.ConfigurationMetaNames.DESCRIPTION: 'MNIST digit recognition with Federated Learning using Tensorflow',
    client.training.ConfigurationMetaNames.TRAINING_DATA_REFERENCES: [],
    client.training.ConfigurationMetaNames.TRAINING_RESULTS_REFERENCE: {
        'type': results_type,
        'name': 'outputData',
        'connection': {},
        'location': { 'path': '/projects/' + PROJECT_ID + '/assets/trainings/'}
    },
    client.training.ConfigurationMetaNames.FEDERATED_LEARNING: {
        'model': {
            'type': 'tensorflow',
            'spec': {
            'id': untrained_model_id
        },
        'model_file': untrained_model_name
    },
    'fusion_type': 'iter_avg',
    'metrics': 'accuracy',
    'epochs': 3,
    'rounds': 10,
    'remote_training' : {
        'quorum': 1.0,
        'max_timeout': 3600,
        'remote_training_systems': [ { 'id': prime_rts_id }, { 'id': nonprime_rts_id} ]
    },
    'hardware_spec': {
        'name': 'S'
    },
    'software_spec': {
        'name': 'runtime-22.1-py3.9'
    }
}

aggregator = client.training.run(aggregator_metadata, asynchronous=True)
aggregator_id = client.training.get_id(aggregator)

class metanames.TrainingConfigurationMetaNames

Set of MetaNames for trainings.

Available MetaNames:

MetaName	Type	Required	Schema	Example value
TRAINING_DATA_REFERENCES	list	Y	[{'name(optional)': 'string', 'type(required)': 'string', 'connection(required)': {'endpoint_url(required)': 'string', 'access_key_id(required)': 'string', 'secret_access_key(required)': 'string'}, 'location(required)': {'bucket': 'string', 'path': 'string'}, 'schema(optional)': {'id(required)': 'string', 'fields(required)': [{'name(required)': 'string', 'type(required)': 'string', 'nullable(optional)': 'string'}]}}]	[{'connection': {'endpoint_url': 'https://s3-api.us-geo.objectstorage.softlayer.net', 'access_key_id': '***', 'secret_access_key': '***'}, 'location': {'bucket': 'train-data', 'path': 'training_path'}, 'type': 's3', 'schema': {'id': '1', 'fields': [{'name': 'x', 'type': 'double', 'nullable': 'False'}]}}]
TRAINING_RESULTS_REFERENCE	dict	Y	{'name(optional)': 'string', 'type(required)': 'string', 'connection(required)': {'endpoint_url(required)': 'string', 'access_key_id(required)': 'string', 'secret_access_key(required)': 'string'}, 'location(required)': {'bucket': 'string', 'path': 'string'}}	{'connection': {'endpoint_url': 'https://s3-api.us-geo.objectstorage.softlayer.net', 'access_key_id': '***', 'secret_access_key': '***'}, 'location': {'bucket': 'test-results', 'path': 'training_path'}, 'type': 's3'}
TEST_DATA_REFERENCES	list	N	[{'name(optional)': 'string', 'type(required)': 'string', 'connection(required)': {'endpoint_url(required)': 'string', 'access_key_id(required)': 'string', 'secret_access_key(required)': 'string'}, 'location(required)': {'bucket': 'string', 'path': 'string'}, 'schema(optional)': {'id(required)': 'string', 'fields(required)': [{'name(required)': 'string', 'type(required)': 'string', 'nullable(optional)': 'string'}]}}]	[{'connection': {'endpoint_url': 'https://s3-api.us-geo.objectstorage.softlayer.net', 'access_key_id': '***', 'secret_access_key': '***'}, 'location': {'bucket': 'train-data', 'path': 'training_path'}, 'type': 's3', 'schema': {'id': '1', 'fields': [{'name': 'x', 'type': 'double', 'nullable': 'False'}]}}]
TEST_OUTPUT_DATA	dict	N	{'name(optional)': 'string', 'type(required)': 'string', 'connection(required)': {'endpoint_url(required)': 'string', 'access_key_id(required)': 'string', 'secret_access_key(required)': 'string'}, 'location(required)': {'bucket': 'string', 'path': 'string'}, 'schema(optional)': {'id(required)': 'string', 'fields(required)': [{'name(required)': 'string', 'type(required)': 'string', 'nullable(optional)': 'string'}]}}	[{'connection': {'endpoint_url': 'https://s3-api.us-geo.objectstorage.softlayer.net', 'access_key_id': '***', 'secret_access_key': '***'}, 'location': {'bucket': 'train-data', 'path': 'training_path'}, 'type': 's3', 'schema': {'id': '1', 'fields': [{'name': 'x', 'type': 'double', 'nullable': 'False'}]}}]
TAGS	list	N	['string']	['string']
PIPELINE	dict	N		{'id': '3c1ce536-20dc-426e-aac7-7284cf3befc6', 'rev': '1', 'modeltype': 'tensorflow_1.1.3-py3', 'data_bindings': [{'data_reference_name': 'string', 'node_id': 'string'}], 'node_parameters': [{'node_id': 'string', 'parameters': {}}], 'hardware_spec': {'id': '4cedab6d-e8e4-4214-b81a-2ddb122db2ab', 'rev': '12', 'name': 'string', 'num_nodes': '2'}, 'hybrid_pipeline_hardware_specs': [{'node_runtime_id': 'string', 'hardware_spec': {'id': '4cedab6d-e8e4-4214-b81a-2ddb122db2ab', 'rev': '12', 'name': 'string', 'num_nodes': '2'}}]}
EXPERIMENT	dict	N		{'id': '3c1ce536-20dc-426e-aac7-7284cf3befc6', 'rev': 1, 'description': 'test experiment'}
PROMPT_TUNING	dict	N		{'task_id': 'generation', 'base_model': {'model_id': 'google/flan-t5-xl'}}
FINE_TUNING	dict	N		{'task_id': 'generation', 'base_model': {'model_id': 'bigscience/bloom-560m'}}
AUTO_UPDATE_MODEL	bool	N		False
FEDERATED_LEARNING	dict	N		3c1ce536-20dc-426e-aac7-7284cf3befc6
SPACE_UID	str	N		3c1ce536-20dc-426e-aac7-7284cf3befc6
MODEL_DEFINITION	dict	N		{'id': '4cedab6d-e8e4-4214-b81a-2ddb122db2ab', 'rev': '12', 'model_type': 'string', 'hardware_spec': {'id': '4cedab6d-e8e4-4214-b81a-2ddb122db2ab', 'rev': '12', 'name': 'string', 'num_nodes': '2'}, 'software_spec': {'id': '4cedab6d-e8e4-4214-b81a-2ddb122db2ab', 'rev': '12', 'name': '...'}, 'command': 'string', 'parameters': {}}
DESCRIPTION	str	Y		tensorflow model training
NAME	str	Y		sample training

Enums

class ibm_watsonx_ai.utils.autoai.enums.ClassificationAlgorithms(value)

Bases: Enum

Classification algorithms that AutoAI can use for IBM Cloud.

DT = 'DecisionTreeClassifier'

EX_TREES = 'ExtraTreesClassifier'

GB = 'GradientBoostingClassifier'

LGBM = 'LGBMClassifier'

LR = 'LogisticRegression'

RF = 'RandomForestClassifier'

SnapBM = 'SnapBoostingMachineClassifier'

SnapDT = 'SnapDecisionTreeClassifier'

SnapLR = 'SnapLogisticRegression'

SnapRF = 'SnapRandomForestClassifier'

SnapSVM = 'SnapSVMClassifier'

XGB = 'XGBClassifier'

class ibm_watsonx_ai.utils.autoai.enums.ClassificationAlgorithmsCP4D(value)

Bases: Enum

Classification algorithms that AutoAI can use for IBM Cloud Pak® for Data(CP4D). The SnapML estimators (SnapDT, SnapRF, SnapSVM, SnapLR) are supported on IBM Cloud Pak® for Data version 4.0.2 and later.

DT = 'DecisionTreeClassifierEstimator'

EX_TREES = 'ExtraTreesClassifierEstimator'

GB = 'GradientBoostingClassifierEstimator'

LGBM = 'LGBMClassifierEstimator'

LR = 'LogisticRegressionEstimator'

RF = 'RandomForestClassifierEstimator'

SnapBM = 'SnapBoostingMachineClassifier'

SnapDT = 'SnapDecisionTreeClassifier'

SnapLR = 'SnapLogisticRegression'

SnapRF = 'SnapRandomForestClassifier'

SnapSVM = 'SnapSVMClassifier'

XGB = 'XGBClassifierEstimator'

class ibm_watsonx_ai.utils.autoai.enums.DataConnectionTypes

Bases: object

Supported types of DataConnection.

CA = 'connection_asset'

CN = 'container'

DS = 'data_asset'

FS = 'fs'

S3 = 's3'

class ibm_watsonx_ai.utils.autoai.enums.Directions

Bases: object

Possible metrics directions

ASCENDING = 'ascending'

DESCENDING = 'descending'

class ibm_watsonx_ai.utils.autoai.enums.DocumentsSamplingTypes

Bases: object

Types of training data sampling.

BENCHMARK_DRIVEN = 'benchmark_driven'

RANDOM = 'random'

class ibm_watsonx_ai.utils.autoai.enums.ForecastingAlgorithms(value)

Bases: Enum

Forecasting algorithms that AutoAI can use for IBM watsonx.ai software with IBM Cloud Pak® for Data.

ARIMA = 'ARIMA'

BATS = 'BATS'

ENSEMBLER = 'Ensembler'

HW = 'HoltWinters'

LR = 'LinearRegression'

RF = 'RandomForest'

SVM = 'SVM'

class ibm_watsonx_ai.utils.autoai.enums.ForecastingAlgorithmsCP4D(value)

Bases: Enum

Forecasting algorithms that AutoAI can use for IBM Cloud.

ARIMA = 'ARIMA'

BATS = 'BATS'

ENSEMBLER = 'Ensembler'

HW = 'HoltWinters'

LR = 'LinearRegression'

RF = 'RandomForest'

SVM = 'SVM'

class ibm_watsonx_ai.utils.autoai.enums.ForecastingPipelineTypes(value)

Bases: Enum

Forecasting pipeline types that AutoAI can use for IBM Cloud Pak® for Data(CP4D).

ARIMA = 'ARIMA'

ARIMAX = 'ARIMAX'

ARIMAX_DMLR = 'ARIMAX_DMLR'

ARIMAX_PALR = 'ARIMAX_PALR'

ARIMAX_RAR = 'ARIMAX_RAR'

ARIMAX_RSAR = 'ARIMAX_RSAR'

Bats = 'Bats'

DifferenceFlattenEnsembler = 'DifferenceFlattenEnsembler'

ExogenousDifferenceFlattenEnsembler = 'ExogenousDifferenceFlattenEnsembler'

ExogenousFlattenEnsembler = 'ExogenousFlattenEnsembler'

ExogenousLocalizedFlattenEnsembler = 'ExogenousLocalizedFlattenEnsembler'

ExogenousMT2RForecaster = 'ExogenousMT2RForecaster'

ExogenousRandomForestRegressor = 'ExogenousRandomForestRegressor'

ExogenousSVM = 'ExogenousSVM'

FlattenEnsembler = 'FlattenEnsembler'

HoltWinterAdditive = 'HoltWinterAdditive'

HoltWinterMultiplicative = 'HoltWinterMultiplicative'

LocalizedFlattenEnsembler = 'LocalizedFlattenEnsembler'

MT2RForecaster = 'MT2RForecaster'

RandomForestRegressor = 'RandomForestRegressor'

SVM = 'SVM'

static get_exogenous()

Get a list of pipelines that use supporting features (exogenous pipelines).

Returns:

list of pipelines using supporting features

Return type:

list[ForecastingPipelineTypes]

static get_non_exogenous()

Get a list of pipelines that are not using supporting features (non-exogenous pipelines).

Returns:

list of pipelines that do not use supporting features

Return type:

list[ForecastingPipelineTypes]

class ibm_watsonx_ai.utils.autoai.enums.ImputationStrategy(value)

Bases: Enum

Missing values imputation strategies.

BEST_OF_DEFAULT_IMPUTERS = 'best_of_default_imputers'

CUBIC = 'cubic'

FLATTEN_ITERATIVE = 'flatten_iterative'

LINEAR = 'linear'

MEAN = 'mean'

MEDIAN = 'median'

MOST_FREQUENT = 'most_frequent'

NEXT = 'next'

NO_IMPUTATION = 'no_imputation'

PREVIOUS = 'previous'

VALUE = 'value'

class ibm_watsonx_ai.utils.autoai.enums.Metrics

Bases: object

Supported types of classification and regression metrics in AutoAI.

ACCURACY_AND_DISPARATE_IMPACT_SCORE = 'accuracy_and_disparate_impact'

ACCURACY_SCORE = 'accuracy'

AVERAGE_PRECISION_SCORE = 'average_precision'

EXPLAINED_VARIANCE_SCORE = 'explained_variance'

F1_SCORE = 'f1'

F1_SCORE_MACRO = 'f1_macro'

F1_SCORE_MICRO = 'f1_micro'

F1_SCORE_WEIGHTED = 'f1_weighted'

LOG_LOSS = 'neg_log_loss'

MEAN_ABSOLUTE_ERROR = 'neg_mean_absolute_error'

MEAN_SQUARED_ERROR = 'neg_mean_squared_error'

MEAN_SQUARED_LOG_ERROR = 'neg_mean_squared_log_error'

MEDIAN_ABSOLUTE_ERROR = 'neg_median_absolute_error'

PRECISION_SCORE = 'precision'

PRECISION_SCORE_MACRO = 'precision_macro'

PRECISION_SCORE_MICRO = 'precision_micro'

PRECISION_SCORE_WEIGHTED = 'precision_weighted'

R2_AND_DISPARATE_IMPACT_SCORE = 'r2_and_disparate_impact'

R2_SCORE = 'r2'

RECALL_SCORE = 'recall'

RECALL_SCORE_MACRO = 'recall_macro'

RECALL_SCORE_MICRO = 'recall_micro'

RECALL_SCORE_WEIGHTED = 'recall_weighted'

ROC_AUC_SCORE = 'roc_auc'

ROOT_MEAN_SQUARED_ERROR = 'neg_root_mean_squared_error'

ROOT_MEAN_SQUARED_LOG_ERROR = 'neg_root_mean_squared_log_error'

class ibm_watsonx_ai.utils.autoai.enums.MetricsToDirections(value)

Bases: Enum

Map of metrics directions.

ACCURACY = 'ascending'

AVERAGE_PRECISION = 'ascending'

EXPLAINED_VARIANCE = 'ascending'

F1 = 'ascending'

F1_MACRO = 'ascending'

F1_MICRO = 'ascending'

F1_WEIGHTED = 'ascending'

NEG_LOG_LOSS = 'descending'

NEG_MEAN_ABSOLUTE_ERROR = 'descending'

NEG_MEAN_SQUARED_ERROR = 'descending'

NEG_MEAN_SQUARED_LOG_ERROR = 'descending'

NEG_MEDIAN_ABSOLUTE_ERROR = 'descending'

NEG_ROOT_MEAN_SQUARED_ERROR = 'descending'

NEG_ROOT_MEAN_SQUARED_LOG_ERROR = 'descending'

NORMALIZED_GINI_COEFFICIENT = 'ascending'

PRECISION = 'ascending'

PRECISION_MACRO = 'ascending'

PRECISION_MICRO = 'ascending'

PRECISION_WEIGHTED = 'ascending'

R2 = 'ascending'

RECALL = 'ascending'

RECALL_MACRO = 'ascending'

RECALL_MICRO = 'ascending'

RECALL_WEIGHTED = 'ascending'

ROC_AUC = 'ascending'

class ibm_watsonx_ai.utils.autoai.enums.PipelineTypes

Bases: object

Supported types of Pipelines.

LALE = 'lale'

SKLEARN = 'sklearn'

class ibm_watsonx_ai.utils.autoai.enums.PositiveLabelClass

Bases: object

Metrics that need positive label definition for binary classification.

AVERAGE_PRECISION_SCORE = 'average_precision'

F1_SCORE = 'f1'

F1_SCORE_MACRO = 'f1_macro'

F1_SCORE_MICRO = 'f1_micro'

F1_SCORE_WEIGHTED = 'f1_weighted'

PRECISION_SCORE = 'precision'

PRECISION_SCORE_MACRO = 'precision_macro'

PRECISION_SCORE_MICRO = 'precision_micro'

PRECISION_SCORE_WEIGHTED = 'precision_weighted'

RECALL_SCORE = 'recall'

RECALL_SCORE_MACRO = 'recall_macro'

RECALL_SCORE_MICRO = 'recall_micro'

RECALL_SCORE_WEIGHTED = 'recall_weighted'

class ibm_watsonx_ai.utils.autoai.enums.PredictionType

Bases: object

Supported types of learning.

BINARY = 'binary'

CLASSIFICATION = 'classification'

FORECASTING = 'forecasting'

MULTICLASS = 'multiclass'

REGRESSION = 'regression'

TIMESERIES_ANOMALY_PREDICTION = 'timeseries_anomaly_prediction'

class ibm_watsonx_ai.utils.autoai.enums.RAGMetrics

Bases: object

Supported types of AutoAI RAG metrics

ANSWER_CORRECTNESS = 'answer_correctness'

CONTEXT_CORRECTNESS = 'context_correctness'

FAITHFULNESS = 'faithfulness'

class ibm_watsonx_ai.utils.autoai.enums.RegressionAlgorithms(value)

Bases: Enum

Regression algorithms that AutoAI can use for IBM Cloud.

DT = 'DecisionTreeRegressor'

EX_TREES = 'ExtraTreesRegressor'

GB = 'GradientBoostingRegressor'

LGBM = 'LGBMRegressor'

LR = 'LinearRegression'

RF = 'RandomForestRegressor'

RIDGE = 'Ridge'

SnapBM = 'SnapBoostingMachineRegressor'

SnapDT = 'SnapDecisionTreeRegressor'

SnapRF = 'SnapRandomForestRegressor'

XGB = 'XGBRegressor'

class ibm_watsonx_ai.utils.autoai.enums.RegressionAlgorithmsCP4D(value)

Bases: Enum

Regression algorithms that AutoAI can use for IBM Cloud Pak® for Data(CP4D). The SnapML estimators (SnapDT, SnapRF, SnapBM) are supported on IBM Cloud Pak® for Data version 4.0.2 and later.

DT = 'DecisionTreeRegressorEstimator'

EX_TREES = 'ExtraTreesRegressorEstimator'

GB = 'GradientBoostingRegressorEstimator'

LGBM = 'LGBMRegressorEstimator'

LR = 'LinearRegressionEstimator'

RF = 'RandomForestRegressorEstimator'

RIDGE = 'RidgeEstimator'

SnapBM = 'SnapBoostingMachineRegressor'

SnapDT = 'SnapDecisionTreeRegressor'

SnapRF = 'SnapRandomForestRegressor'

XGB = 'XGBRegressorEstimator'

class ibm_watsonx_ai.utils.autoai.enums.RunStateTypes

Bases: object

Supported types of AutoAI fit/run.

COMPLETED = 'completed'

FAILED = 'failed'

class ibm_watsonx_ai.utils.autoai.enums.SamplingTypes

Bases: object

Types of training data sampling.

FIRST_VALUES = 'first_n_records'

LAST_VALUES = 'truncate'

RANDOM = 'random'

STRATIFIED = 'stratified'

class ibm_watsonx_ai.utils.autoai.enums.TShirtSize

Bases: object

Possible sizes of the AutoAI POD. Depending on the POD size, AutoAI can support different data set sizes.

• S - small (2vCPUs and 8GB of RAM)

• M - Medium (4vCPUs and 16GB of RAM)

• L - Large (8vCPUs and 32GB of RAM))

• XL - Extra Large (16vCPUs and 64GB of RAM)

L = 'l'

M = 'm'

S = 's'

XL = 'xl'

class ibm_watsonx_ai.utils.autoai.enums.TimeseriesAnomalyPredictionAlgorithms(value)

Bases: Enum

Timeseries Anomaly Prediction algorithms that AutoAI can use for IBM Cloud.

Forecasting = 'Forecasting'

Relationship = 'Relationship'

Window = 'Window'

class ibm_watsonx_ai.utils.autoai.enums.TimeseriesAnomalyPredictionPipelineTypes(value)

Bases: Enum

Timeseries Anomaly Prediction pipeline types that AutoAI can use for IBM Cloud.

PointwiseBoundedBATS = 'PointwiseBoundedBATS'

PointwiseBoundedBATSForceUpdate = 'PointwiseBoundedBATSForceUpdate'

PointwiseBoundedHoltWintersAdditive = 'PointwiseBoundedHoltWintersAdditive'

WindowLOF = 'WindowLOF'

WindowNN = 'WindowNN'

WindowPCA = 'WindowPCA'

class ibm_watsonx_ai.utils.autoai.enums.Transformers

Bases: object

Supported types of congito transformers names in AutoAI.

ABS = 'abs'

CBRT = 'cbrt'

COS = 'cos'

CUBE = 'cube'

DIFF = 'diff'

DIVIDE = 'divide'

FEATUREAGGLOMERATION = 'featureagglomeration'

ISOFORESTANOMALY = 'isoforestanomaly'

LOG = 'log'

MAX = 'max'

MINMAXSCALER = 'minmaxscaler'

NXOR = 'nxor'

PCA = 'pca'

PRODUCT = 'product'

ROUND = 'round'

SIGMOID = 'sigmoid'

SIN = 'sin'

SQRT = 'sqrt'

SQUARE = 'square'

STDSCALER = 'stdscaler'

SUM = 'sum'

TAN = 'tan'

class ibm_watsonx_ai.utils.autoai.enums.VisualizationTypes

Bases: object

Types of visualization options.

INPLACE = 'inplace'

PDF = 'pdf'