CHANGELOG.md

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

Changes are grouped as follows

• Added for new features.
• Changed for changes in existing functionality.
• Deprecated for soon-to-be removed features.
• Removed for now removed features.
• Fixed for any bug fixes.
• Security in case of vulnerabilities.

[0.60.33]

Changed

• Make geopandas and shapely dependencies optional

[0.60.32]

Changed

• Geospatial search: update to reflect API change in order_by

[0.60.31]

Changed

• Geospatial new path support and new feature type update format support.

[0.60.30]

Added

• new filter by transformation_external_id for transformations.jobs.list().

[0.60.29]

Changed

• transformation.job.raw_query renamed to query, because of API changes.

[0.60.28]

Removed

• transformation_schedules.retrieve_multiple() no longer acepts is_public parameter, it's asumed the user wants to retrieve the given ids either they are marked as public or not.
• internal changes in the way transformation.preview() sends its parameters to the backend.

[0.60.27]

Added

• Geospatial: support allow_crs_transformation for creating, updating and searching features.

[0.60.26]

Fixed

• added TransformationBlockedInfo time property to fix decoding Transformation objects.

[0.60.25]

Added

• added from_geopandas static method to FeatureList.

[0.60.23]

Added

• added to_geopandas method on FeatureList.

[0.60.22]

Fixed

• str(TransformationJobMetric), str(TransformationJob), str(TransformationNotification) and str(TransformationSchedule) no longer throw an exception.

[0.60.21]

Added

• added destination filter on TransformationNotificationsAPI.list().

[0.60.20]

Added

• added Transformation.running_job, Transformation.last_finished_job, Transformation.schedule and Transformation.blocked fields.

[0.60.19]

Fixed

• str(TransformationPreviewResult) no longer throws an exception.

[0.60.18]

Added

• added support for ignore_unknown_ids parameter in transformations.retrieve_multiple and transformations.delete.

[0.60.17]

Changed

• Dependencies upgrade

[0.60.16]

Added

• added transformationsAPI.preview() method, to preview the result of a query before running a transformation.

[0.60.15]

Added

• optional timeout parameter for TransformationJob.wait() and TransformationJob.wait_async() methods.

[0.60.14]

Added

• GeospatialAPI.stream_features method to stream features.

[0.60.13]

Added

• GeospatialAPI.update_feature_types method to update feature types.

[0.60.12]

Changed

• TransformationDestination.raw() now uses type = 'raw' istead of 'raw_table' and doesn't use deprecated raw_type parameter.

[0.60.11]

Added

• TransformationJobsAPI.list_metrics method to retrieve job metrics.
• transformations.jobs.retrieve_multiple method to get multiple jobs by id in a single API call.

[0.60.10]

Added

• oicdCredentials.audience parameter for transformations.

[0.60.9]

Added

• transformation_id parameter in TransformationJobsAPI.list(), to list only the jobs associated with a given transformation.

[0.60.8]

Added

• TransformationsApi.retrieve_multiple method, to get multiple transformations from the backend in a single request.

[0.60.6]

Added

• TransformationJob.transformation_id field to asossiate a job with it's source transformation.

[0.60.5]

Changed

• transformations.notificationsnow uses transformation_id and transformation_external_id instead of config_id and config_external_id.

[0.60.4]

Added

• transformations.jobs api client, which retrieves the status of transformation runs.
• transformations.run method, which runs transformations.

[0.60.4]

Changed

• Extraction pipeline entity updated to be same with API entity.

[0.60.3]

Added

• transformations.schema api client, which allows the retrieval of the expected schema of sql transformations based on the destination data type.

[0.60.2]

Changed

• FunctionsAPI.call and FunctionSchedulesAPI.create fail with CogniteAPIError if the request to the sessions API is unsuccesful (instead of creating calls/schedules without sessions).

[0.60.1]

Added

• transformations.schedules.retrieve_multiple to retrieve multiple transformation schedules by transformation ids and external ids.

[0.59.3]

Fixed

• Classes DiagramDetectResults and DiagramConvertResults return items-property correctly when the job is not completed.

[0.59.2]

Added

• transformations.notifications api client, which allows the creation, deletion and retrieval of transformation email notifications.

[0.59.1]

Changed

• In FunctionsAPI.create, the default value of cpu is changed from 0.25 to None, thus deferring the default value to the API, which also is 0.25. The argument is unavailable in Azure.
• In FunctionsAPI.create, the default value of memory is changed from 1.0 to None, thus deferring the default value to the API, which also is 1.0. The argument is unavailable in Azure.

[0.59]

Added

• transformations api client, which allows the creation, deletion, update and retrieval of transformations.
• transformations.schedules api client, which allows the schedule, unschedule and retrieval of recurring runs of a transformation.

[0.57.1]

Added

• session_id property to FunctionSchedule class, which is the id of the session started when creating a schedule with client credentials.

[0.57.0]

Changed

• FunctionSchedulesAPI.list() now accepts the argument function_id. The arguments function_external_id and function_id are mutually exclusive.
• Function.list_schedules() lists schedules attached to the function with both function_id and function_external_id.

[0.56.1]

Fixed

• Bug: FunctionSchedulesAPI.create did not pass function_id to request body.

[0.56.0]

Changed

• Rename Integrations to Extraction Pipelines.

[0.55.0]

Added

• Add functionality to filter integration runs. Some new fields added to integration entity.

[0.54.0]

Changed

• FunctionSchedulesAPI.create now requires function_id to be used when creating a schedule with OIDC through client_credentials.
• FunctionSchedulesAPI.create: The arguments function_id and function_external_id are both optional and mutually exclusive.
• FunctionSchedulesAPI.create: The argument function_external_id has changed position, necessitated by the fact that it is now optional.

Fixed

• Correct error message when specifying both function_id and function_external_id on methods FunctionCallsAPI.list(), FunctionCallsAPI.retrieve(), FunctionCallsAPI.get_response() and FunctionCallsAPI.get_logs().

[0.53.1]

Fixed

• FunctionsAPI.call previously didn't pass the session nonce if data was set to None. This is now fixed.

[0.53.0]

Changed

• FunctionsAPI.call now uses OIDC tokens if the client was instantiated with a token or client credentials.
• FunctionSchedulesAPI.create now supports OIDC tokens through the use of client-credentials, explicitly passed in as an argument.

[0.52.0]

Fixed

• FunctionsAPI.list() and FunctionSchedulesAPI.list() with argument limit equal to None, float(inf) or -1 now returns all resources. Previously, the default limit of the API was used (100).

[0.51.0]

Fixed

• Fixed Templates API

[0.50.0]

Added

• Diagrams API

Removed

• Unstructured API (due to lack of maintenance)
• Templates API (now available in the main SDK). Note that template completion is still available.

[0.49.0]

Added

• Update method for the annotations API

[0.48.2]

Fixed

• FunctionsAPI.list() and FunctionSchedulesAPI.list() were missing cognite_client from a bug introduced in version [0.47.0].

[0.48.1]

Fixed

• Image parameter for plot data extraction

[0.48.0]

Added

• replacements parameter in entity matching fit
• suggest fields endpoint in entity matching

[0.47.0]

Changed

• The FunctionsAPI.list()-method now accepts additional filters: name, owner, file_id, status, external_id_prefix and created_time.
• The FunctionSchedulesAPI.list()-method now accepts additional filters: name, function_external_id, created_time and cron_expression.
• The Function.list_schedules()-method now accepts limit argument.

[0.46.0]

Added

• Add functionality to add/remove elements in integration list fields.

[0.45.2]

Fixed

• The Function.list_schedules()-method now correctly returns all schedules tied to the function.

[0.45.0]

Added

• Possibility to detect entities with multiple names, as this is now a feature in the API.
• Same functionality in document detect.

[0.44.0]

Changed

• Domains/schemas/universes/domains are now consistently called templates in the SDK, and the completion functions are moved to client.templates.completion
• Annotations filter updated with additional fields, and to reflect annotated resource type being mandatory.

[0.43.0]

Added

• Possibility to add environment variables to functions via the new argument env_vars of the FunctionsAPI.create() method.
• Optional argument limit to Function.list_calls()

[0.42.0] - 21-02-01

Added

• Template functionality.

Removed

• Playground relationships.

[0.41.0] - 2020-01-19

Added

• New entity match rules endpoints are available in client.match_rules

Changed

• EntityMatchingPipeline update accepts its own suggested rules, despite extra fields.

Fixed

• Better exception checks for missing packages in notebook output.
• Fixed missing column error in dataframe in notebook output.

[0.40.1] - 2020-01-19

Changed

• PNID detect no longer blocks, entities passed as a dict are resolved in the .matches helper method.

[0.40] - 2020-01-19

Changed

• Richer output for entity matching pipelines and PNID jobs in notebooks
• More specific job types returned from these endpoints, with helper methods that make processing and visualizing output easier.

Added

• OCR method for PNID parsing

[0.39] - 2021-01-18

Changed

• Integrations and IntegrationRuns data classes updated with new fields.

[0.38] - 2020-01-12

Changed

• The following methods now accepts a limit-parameter, with default of 25, that controls how many elements the methods return: FunctionCallAPI.list, FunctionAPI.list, and FunctionScheduleAPI.list.
• Removed data property from FunctionSchedule class.

[0.37] - 2020-01-11

Changed

• The complete method of the schema completion api was renamed to complete_type

Added

• Method complete_domain to schema completion

[0.36] - 2020-01-05

Changed

• The FunctionCallAPI.retrieve method now utilizes retrieve_multiple in the backend. /calls -> /calls/byids

Added

• Method retrieve to FunctionScheduleAPI-class.

[0.35] - 2020-12-14

Added

• Added IntegrationsAPI and IntegrationRunsAPI

[0.34] - 2020-12-08

Added

• Added relationships_label and use_existing_matches to pipelines

[0.33.0] - 2020-12-02

Added

• Method get_input_data on the FunctionSchedule-class.
• Method get_input_data on the FunctionScheduleAPI-class.

[0.32.0] - 2020-12-03

Fixed

• Fix the client initialization to correctly pass all keyword arguments.

[0.31.2] - 2020-12-03

Fixed

• Fix the filter functionality of annotations

[0.31.1] - 2020-11-30

Added

• Argument function_call_info to the definition of the handle function. This argument is a dictionary with keys function_id and, if the call is scheduled, schedule_id and scheduled_time.
• Property scheduled_time on FunctionCall class.

[0.31.0] - 2020-11-25

Changed

• Core entity matching moved to beta client, experimental now extends this.

[0.30.0] - 2020-11-22

Fixed

• Pnid Object Detection GET endpoint fixed after separation between /findobjects and /findsimilar

[0.29.0] - 2020-11-09

Added

• Arguments confirmed_matches and schedule_interval for EntityMatchingPipeline
• Method retrieve_latest for entity matching pipeline run.
• Various helper methods on EntityMatchingPipeline and EntityMatchingPipelineRun
• EntityMatchingPipeline Update class and method

[0.28.0] - 2020-10-27

Changed

• Renamed match_from and match_to in entity matcher, as well as timestamp->time in contextualization jobs.
• Entity matching matches renamed to true_matches

Added

• Entity matching pipelines rejected_matches and score_threshold options

Fixed

• Entity matching pipelines now correctly takes model_parameters instead of model_id

[0.27.0] - 2020-10-20

Added

• Plot Data Extraction API

[0.26.0] - 2020-10-09

Added

• AnnotationsAPI with create, list, retrieve, retrieve_multiple, delete functionalities

[0.25.0] - 2020-10-08

Changed

• Entity matcher fields renamed, and added a new format for true_matches. complete_missing -> ignore_missing_fields, keys_from_to -> match_fields.

[0.24.0] - 2020-10-01

Changed

• Inherit from the beta client
• Old relationships moved to .relationships_playground, as beta client contains the new version.
• Entity extraction removed as the API no longer exists.

[0.23.1] - 2020-10-01

Added

• Arguments cpu and memory to FunctionsAPI.create(), making it possible to customize function resources.

[0.23.0] - 2020-09-29

Added

• Entity matching pipelines, which allows you to deploy an entity matching model with confirmed matches and rules.

Changed

• Entity matching list calls have a limit parameter which defaults to 100.
• ContextualizationJobs now have timestamp fields consistently as members, and no longer return them in result.

[0.22.3] - 2020-09-15

Changed

• /fit in entity matcher supports defining which field in matchFrom and matchTo to use as the id field by specifying id_field

[0.22.2] - 2020-09-14

Changed

• /detect, /convert, /extractpattern, /ocr supports referring to a file using file_id and/or file_external_id. file_id and file_external_id are returned in the responses.

[0.22.1] - 2020-09-09

Added

• /detect in Pnid Parser can accept entities with type List[Union[str, dict]] and returns entities if the type of input entities has type of List[dict]

[0.22.0] - 2020-09-09

Added

• Added /findobjects in Pnid Object Detection

[0.21.0] - 2020-09-09

Removed

• Removed /parse in Pnid Parser

Added

• Added /convert in Pnid Parser

[0.20.1] - 2020-09-07

Fixed

• Fix in EntityMatcher predict method.

[0.20.0] - 2020-09-01

Added

• Entity matcher predict and refit methods now exist on client.entity_matching in addition to the model object.

[0.19.0] - 2020-08-31

Changed

• Entity matcher methods now all take and return id instead of model_id and accept id or external_id.
• Entity matcher uses the new /entitymatching endpoints

Removed

• fit_ml and predict_ml methods in entity matcher.

[0.18.0] - 2020-08-25

Added

• Add py.typed file to inform that package is typed

[0.17.0] - 2020-08-20

Changed

• Some methods in entity matcher now take id instead of model_id

Added

• Update and retrieve_multiple methods for entity matcher.
• Changed entity matching methods to take external_id where possible.

[0.16.0] - 2020-08-13

Changed

• Change list-method to list all models instead of jobs.
• Changed endpoints used for list and list_jobs to comply with updates to the API.

Added

• Add list_jobs-method that behaves as old list-method, that is returns all jobs.
• Option to filter returned models on based on all input parameters.

[0.15.4] - 2020-08-13

Added

• Add name, description and external-id parameters to fit entity matcher.

[0.15.3] - 2020-08-13

Fixed

• Fix import time bug in functions.

[0.15.2] - 2020-08-13

Added

• Insert sleep cycles for files api queries in functions creation

[0.15.1] - 2020-08-10

Added

• Add model info for entity matcher models (classifier, feature_type, model_to, keys_from_to) to the output of retrieve.

[0.15.0] - 2020-08-10

Changed

• Changed name of the parameter indicating the type of features created from model_type to feature_type, entity matcher.

[0.14.5] - 2020-08-07

Added

• Allow users to specify custom path to handle function by setting function_path.

[0.14.4] - 2020-08-06

Removed

• Removed everything related to resource typing (contextualization).

[0.14.3] - 2020-08-06

Added

• A class for Document parsing, allowing detecting entity in documents
• Detect function also for pnid, similar to the old parse function, which may get deprecated
• min_tokens parameter for entity detecting endpoints. Allowing to set the mimimum number of tokens needed for a detection

[0.14.2] - 2020-08-05

Changed

• Renamed refitml refit entity matcher.

[0.14.1] - 2020-08-03

Added

• Added retry to functions POST endpoints.

[0.14.0] - 2020-08-03

Changed

• Merge fit_ml and fit, and predict_ml and predict endpoints entity matcher - major changes to required parameters and responses.

[0.13.0] - 2020-07-23

Changed

• Base version of the SDK changed to 2.x

[0.12.1] - 2020-07-07

Added

• The entity matching methods now have additional options for classifier and feature types.
• PNID detect patterns endpoint added.
• Schema completion endpoint added.

[0.12.0] - 2020-07-07

Changed

• Synthetic time series and labels removed, as they were moved to non-experimental status in the main SDK.

[0.11.2] - 2020-06-19

Added

• The method Function.update(), which updates the function object.

[0.11.1] - 2020-06-10

Fixed

• Client no longer gives an error on using a token instead of an API key.

[0.11.0] - 2020-06-08

Changed

• Removed FunctionCallResponse class. Functions FunctionCallsAPI.get_response() and FunctionCall.get_response() now returns the actual function response (without being wrapped in an object with call_id, function_id and response).

[0.10.0] - 2020-06-04

Changed

• POST endpoints such as search and list will be retried in experimental endpoints, as they are in v1.

[0.9.1] - 2020-06-04

Changed

• Updated documentation for functions.

[0.9.0] - 2020-05-29

Changed

• Function response is no longer a property of the FunctionCall class. Instead, the response can be retrieved by the methods FunctionCallsAPI.get_response() or FunctionCall.get_response().
• The documentation for function schedules is put under the expand/collapse header Function Schedules.

Added

• Filtering of function calls given call attributes and added an attribute for schedule id to the FunctionCall data class.

[0.8.3] - 2020-05-28

Fixed

• The method Function.call() now takes the argument wait (defaults to True) instead of asynchronous. This change was supposed to be a part of release 0.8.0.

[0.8.2] - 2020-05-27

Fixed

• Allow true_matches to be None in fit_ml.

[0.8.1] - 2020-05-26

Changed

• Function calls now returns functionId, so the getting logs of a call has a simplified internal structure.

[0.8.0] - 2020-05-26

Changed

• In FunctionsAPI.call(), the asynchronous argument has been removed, reflecting the Functions API which now only supports asynchronous calls. A new argument wait has been introduced. When wait=True (default), FunctionsAPI.call() will block until the call is finished.

[0.7.3] - 2020-05-19

Fixed

• Dosctring for FunctionCallsAPI.list() erroneously listed external_id as optional argument. This has been corrected to function_external_id.

[0.7.2] - 2020-05-15

Fixed

• Support for no entities passed on predict_ml.

[0.7.1] - 2020-05-15

Added

• Support for ML Entity matcher from/to columns and retrain.

[0.7.0] - 2020-05-14

Added

• Labels endpoint support
• Assets labelling support

[0.6.0] - 2020-05-08

Changed

• SDK updated for changes in ML Entity matcher endpoints.

Fixed

• FunctionSchedulesAPI.create() works when argument name contains the character /.

[0.5.8] - 2020-04-30

• Assets playground API uses playground for retrieve

[0.5.7] - 2020-04-29

Added

• Added checks that verify that the function handler in an uploaded function is correctly constructed.

Fixed

• FunctionSchedulesAPI.create() works without providing the data argument.

[0.5.6] - 2020-04-28

Added

• Data classes for contextualization models and job now include time stamps for created_time, start_time, status_time.
• Unstructured search endpoints added in client.files.unstructured.

[0.5.5] - 2020-04-21

Changed

• Data class Function now has the additional attribute error.

[0.5.4] - 2020-04-20

Added

• fit_ml function on entity matcher and predict_ml on model object.

[0.5.3] - 2020-04-16

Added

• Sphinx documentation for Cognite Functions

[0.5.2] - 2020-04-16

Added

• A FunctionSchedule class and corresponding api attached to the Functions api to interact with function schedules. Function schedules can now also be listed using the Function object through the list_schedules method.

[0.5.1] - 2020-04-15

Added

• Argument function_handle to FunctionsAPI.create(), which can be used to create a function directly from code.

[0.5.0] - 2020-04-08

Added

• FunctionsAPI to interact with the Cognite Functions API in CDF.

Changed

• Refactor of contextualization endpoints to not use asyncio and block later.