Skip to main content

TIM API

r2023.5​

New features​

TIM Forecast

Changes​

Various improvements and bug fixes

r2023.3​

New features​

TIM DB

TIM Detect

Changes​

TIM Forecast

TIM Detect

r2023.2​

New features​

TIM Detect

Bug fixes​

TIM DB

  • Fixed estimation of dataset slice size preventing some ML jobs to run on gappy data

Changes​

TIM Forecast

  • Changed default rebuilding policy for forecasting jobs from "NewSituations" to "All"

r2023.1​

New features​

TIM Detect

TIM DB

TIM Explain

  • Performance improvements to RCA

Changes​

TIM Detect

  • Increased maximum dataset row limit for system driven AD job to 500 000

v5.4.6 [2022-11-03] (LTS 2022.10)​

Changes​

  • Added queue priority to the log of a job
  • Offset limit can not be set in job configuration (not even to 0) if allow offsets is false
  • Removed Month base unit from prediction from, prediction to, rolling window and rebuilding policy job configuration settings. Only Sample base unit is supported for monthly and yearly data.
  • Removed model quality parameter from the TIMForecast Model Zoo and introduced new version v5.6 of the Model Zoo
  • Modified behavior of a what-if analysis job status - if no predictions are affected, the job ends up in FinishedWithWarning status instead of Failed and a warning message is added to the job log
  • Increased the ratio of data used to create offsets and windows to the number of all observations when generating TIM Forecast features

Bug fixes​

  • Improved json schema validation for POST endpoints in case of invalid type of registration body
  • Added the category filter to the calculation of number of observations in the job validation
  • Added error message for a use case with no dataset assigned in TIM Detect
  • Fixed edge case when a TIM Detect job is registered with rows parameter set to Sample 0
  • Fixed out-of-sample rows configuration parameter in TIM Forecast so that it does not influence the preselection of features
  • Fixed counting number of observations for panel data in TIM DB
  • Proper API error message returned when trying to upload timestamp with year lower than 1 to TIM DB

v5.4.5 [2022-10-20]​

Changes​

  • Renamed multivariate approach to system-driven approach TIM Detect

Bug fixes​

  • Added warning message to the log instead of the failed job when results table is empty in TIM Forecast

  • Fixed generation of forecasts for unaligned panel data in TIM Forecast

  • Fixed last target timestamp in the data alignment configuration in TIM Forecast so that it can be greater than the last timestamp in the dataset

v5.4.4 [2022-10-13]​

Changes​

  • Enabled usage of longer offsets than automatically detected threshold if offsetLimit is set in TIM Forecast
  • Modified offset limit detection for short datasets in TIM Forecast

v5.4.3 [2022-09-29]​

New features​

Changes​

Bug fixes​

v5.4.2 [2022-08-26]​

New features​

Changes​

Bug fixes​

v5.4.1 [2022-07-18]​

Bug fixes​

  • Handled redelivered messages from queue manager
  • Handled reconnection to queue manager

v5.4.0 [2022-07-06]​

New features​

Changes​

Bug fixes​

v5.3.1 [2022-05-17]​

New features​

Changes​

  • Improved authentication
  • Telemetry endpoints
    • GET /ops/dataset-calls
      • endpointId filter accepts DatasetManagement endpoints (DM01 to DM16)
      • Removed datasetMicroservice filter
      • Renamed datasetExistsType filter to datasetState
      • Renamed time to madeAt in response
    • GET /ops/job-calls
      • endpointId filter accepts JobManagement endpoints (JM01 to JM47)
      • Removed jobMicroservice filter
      • Renamed jobExistsType filter to jobState
      • Renamed time to madeAt in response

Bug fixes​

v5.3.0 [2022-04-19] (LTS 2022.4)​

New features​

Changes​

  • Removed space (' ') as possibility for automatically detected csvSeparator parameter in CSV upload and update

Bug fixes​

v5.2.1 [2022-03-08]​

New features​

Changes​

  • Updated and improved model results in Swagger
  • Removed GET /datasets/supported-formats
  • Improved descriptiveness of error messages
  • Made type of offsetLimit optional in configuration in registration bodies of POST /forecast-jobs/build-model and POST /forecast-jobs/{id}/rebuild-model
  • Improved timestamp validation during dataset upload
  • Improved timestamp format validation and added supported formats to Swagger schema
  • Removed unused fields from token payload
  • Changed expiresIn of a token to an integer representing seconds

Bug fixes​

v5.2.0 [2022-01-13]​

New features​

  • Implemented the ISO 8601 standard throughout the REST API
    • Added timeZone as an optional parameter for uploading and updating of datasets
    • timeZone information is returned as part of dataset and dataset version metadata
    • All returned timestamps are in UTC format
      • Sliced data (download)
      • Metadata parameters
    • Data aggregation happens in the original dataset timezone

Datasets:

  • Added samplingPeriod parameter to SQL dataset upload
  • SQL dataset upload automatically recognizes the timestamp column

Changes​

Datasets:

  • Disallowed duplicate headers (variable names) in uploaded datasets

Bug fixes​

Forecasting:

  • Fixed handling of a lost AMQP connection

Anomaly Detection:

  • Fixed handling of a lost AMQP connection

Datasets:

  • Improved error message on attempting to upload an empty file as a dataset

Licenses & Users:

  • Improved error message when sending a valid JSON request body that does not match the expected schema

v5.1.20 [2021-12-09]​

New features​

Forecasting:

  • Added parameter offsetLimit to configuration of forecasting jobs of type build-model and rebuild-model

Datasets:

  • Expanded sorting options in GET /datasets to include "+createdAt", "-createdAt", "+updatedAt", "-updatedAt", "+title" and "-title"
  • Expanded GET /datasets/{id}/versions to return complete dataset version entities

Changes​

Bug fixes​

Forecasting:

  • Fixed failing job registration when predictionTo is lower than the sampling period of the data
  • Removed default for dailyCycle parameter in Swagger documentation

Datasets:

  • Improved error message when attempting to upload a dataset with an empty column (with header, but no observations)
  • Caused jobs to fail when column statistics are not correctly stored
  • Fixed uploading of columns with NaN values

Licenses & Users:

  • Made sorting consistent by only accepting predefined values (enumeration instead of string)
  • Enabled users to update their own first and last name

v5.1.19 [2021-11-03]​

Changes​

Datasets:

  • Improved validation of the amount of data (number of rows/observations) that is uploaded

Bug fixes​

  • Improved validation of UUIDs
  • Removed unused schemas from Swagger

Datasets:

  • Improved error messages in log with column names
  • Fixed handling timestampColumn as string (name of column) in POST /datasets and PATCH /datasets/{id}
  • Reintroduced support for timestamp format "yyyy-mm-dd", "mm-dd-yyyy", "mm/dd/yyyy" and "dd/mm/yyyy"

Licenses & Users:

  • Improved error messages on concurrent database operations
  • Disallowed deleting a personal user group
  • Improved error response for POST /user-groups with missing Content-Type header

v5.1.18 [2021-10-19]​

New features​

Forecasting:

  • Added optional allowOffsets parameter in model (re)building configuration that enables the use of influencers that are not not available throughout the whole forecasting horizon by creating their offset transformations

Changes​

  • Implemented ahead-of-time compilation of microservices, increasing the speed of their first run
  • Added support for Azure Service Bus
  • Unified log format

Datasets:

  • Improved implementation of POST /datasets/csv

Bug fixes​

Forecasting:

  • Fixed error message when setting the targetColumn configuration parameter to the timestamp column
  • Fixed usage of predictionFrom and predictionTo configuration parameters in determining job category (light/heavy)
  • Included use of outOfSampleRows in determining job category (light/heavy)
  • Improved ordering of response of GET /forecast-jobs/{id}/status-collect by ordering progress messages with identical timestamps by progress

Anomaly Detection:

  • Improved ordering of response of GET /detection-jobs/{id}/status-collect by ordering progress messages with identical timestamps by progress

Datasets:

  • Fixed handling of samplingPeriod parameter in configuration of POST /datasets/csv
  • Fixed error message when calling PATCH /datasets/{id}/csv with correctly formatted, yet unexisting dataset ID
  • Fixed handling of timestamp formats including 'Z' character (Zulu)
  • Fixed use of timestampFormat parameter in configuration of POST /datasets/sql
  • Fixed validation of workspace ID

Licenses & Users:

  • Enabled updating of main license administrator to a license member who has not yet logged in to the TIM Platform (v5)

v5.1.17 [2021-09-21]​

New features​

Forecasting:

  • Enabled additional filtering during retrieval of lists of forecasting jobs:
    • on minimal date and time of creation (from)
    • on maximal date and time of creation (to)
  • Added DELETE /forecast-jobs endpoint for deleting multiple forecasting jobs
  • Enabled filtering during deletion of multiple forecasting jobs:
    • on experiment
    • on use case
    • on type
    • on status
    • on parent job
    • on minimal date and time of creation (from)
    • on maximal date and time of creation (to)

Anomaly Detection:

  • Enabled additional filtering during retrieval of lists of anomaly detection jobs:
    • on minimal date and time of creation (from)
    • on maximal date and time of creation (to)
  • Added DELETE /detection-jobs endpoint for deleting multiple anomaly detection jobs
  • Enabled filtering during deletion of multiple anomaly detection jobs:
    • on experiment
    • on use case
    • on type
    • on status
    • on parent job
    • on minimal date and time of creation (from)
    • on maximal date and time of creation (to)

Datasets:

  • Enabled additional filtering during retrieval of lists of datasets:
    • on minimal date and time of creation (from)
    • on maximal date and time of creation (to)
    • on status
  • Added DELETE /datasets endpoint for deleting multiple datasets
  • Enabled filtering during deletion of multiple datasets:
    • on minimal date and time of creation (from)
    • on maximal date and time of creation (to)
    • on workspace

Changes​

  • Unified JSON log format
  • Improved parameter descriptions in Swagger
  • Hierarchical API request bodies
  • Hierarchical API responses

Forecasting:

  • For POST /forecast-jobs/{id}/execute, changed HTTP response status code for successful requests from 200 (OK) to 202 (Accepted)

Anomaly Detection:

  • Improved handling of invalid API calls
  • For POST /detection-jobs/{id}/execute, changed HTTP response status code for successful requests from 200 (OK) to 202 (Accepted)

Bug fixes​

Forecasting:

  • Handled scenario throwing exception during holiday column validation
  • In POST /forecast-jobs/{id}/predict, improved schema validation
  • Handled job registration (build, rebuild, predict) with dataset version that has not yet finished uploading
  • Fixed issue of no out-of-sample rows when relativeRange is used
  • For GET /forecast-jobs/{id}/results/table, updated example response
  • For GET /forecast-jobs and GET /forecast-jobs/{id}, updated Swagger response schema to include calculationTime key

Anomaly Detection:

  • Handled job registration (build, rebuild, detect) with dataset version that has not yet finished uploading
  • Handled scenario throwing exception during holiday column validation
  • Improved recognition of endpoints
  • For GET /detection-jobs/{id}/results/table, updated example response

Datasets:

  • Fixed dataset upload remaining stuck in status Registered
  • Fixed foreign key violation on dataset log update
  • In DELETE /datasets/{id} and DELETE /datasets/{id}/versions/{version}, remove use cases using that dataset instead of returning error
  • In POST /datasets/{id}/versions/{version}/slice, fixed handling of imputation type
  • Fixed bug in update of dataset with columns in a different order than in previous versions

Workflow:

  • Improved example responses in Swagger
  • In DELETE /workspaces/{id}, remove datasets inside that workspace instead of returning error

Licenses & Users:

  • Fixed creation of default workspace

v5.1.16 [2021-08-12]​

Changes​

Forecasting:

  • Improved Swagger response for GET /forecast-jobs/{id}/results/rca

Bug fixes​

  • Fixed handling of header X-User-Id
  • Added the following headers to Swagger:
    • tokenAuth
    • X-Log-Reference
    • X-Tim-Client

Forecasting:

  • In POST /forecast-jobs/{id}/rca, made modelIndex required
  • In POST /forecast-jobs/build-model and POST /forecast-jobs/{id}/rebuild model, disabled setting rollingWindow to 0

v5.1.15 [2021-08-10]​

Changes​

  • Formatted console logs as JSONs
  • Hid deprecated warnings from stdout

Bug fixes​

Datasets:

  • Expanded handling of CSV files with a Byte Order Mark (BOM)

v5.1.14 [2021-08-06]​

New features​

Forecasting:

  • Enabled sorting during retrieval of lists of forecasting jobs:
    • ASC and DESC
    • on created_at, executed_at, completed_at and priority
  • Enabled additional filtering during retrieval of lists of forecasting jobs:
    • on status
    • on parent job
  • Added POST /forecast-jobs/{id}/predict endpoint for forecasting with an existing model

Anomaly Detection:

  • Enabled sorting during retrieval of lists of anomaly detection jobs:
    • ASC and DESC
    • on created_at, executed_at, completed_at and priority
  • Enabled additional filtering during retrieval of lists of forecasting jobs:
    • on status
    • on parent job

Datasets:

  • Enabled sorting during retrieval of lists of datasets:
    • ASC and DESC
    • on created_at
  • Enabled additional filtering during retrieval of lists of datasets:
    • on workspace

Workflow:

  • Enabled sorting during retrieval of lists of workspaces:
    • ASC and DESC
    • on created_at, updated_at and title
  • Enabled sorting during retrieval of lists of use cases:
    • ASC and DESC
    • on created_at, updated_at, title, dataset and accuracy impact
  • Enabled sorting during retrieval of lists of experiments:
    • ASC and DESC
    • on created_at, updated_at and title

Licenses & Users:

  • Error codes for error and exception logging

Changes​

  • Unification of Swagger file structure

Forecasting:

  • Removed default value for maxModelComplexity in configuration or build model jobs and rebuild model jobs
  • Attempts at executing an already executed job return status 200 instead of 400
  • In GET /forecast-jobs/{id}/results/rca, renamed the parameter lag to radius

Anomaly Detection:

  • Attempts at executing an already executed job return status 200 instead of 400
  • In GET /detection-jobs/{id}/results/rca, renamed the parameter lag to radius

Bug fixes​

Forecasting:

  • Catch job fails due to too large of a dataset slice upon registration (instead of during execution)
  • Fixed handling of registration of a model building job with an invalid use case ID

Anomaly Detection:

  • Catch job fails due to too large of a dataset slice upon registration (instead of during execution)
  • Fixed handling of registration of a model building job with an invalid use case ID

Datasets:

  • Add size limit for download (POST /datasets/{id}/versions/{version}/slice)

v5.1.13 [2021-07-23]​

New features​

  • Error messages expanded:
    • with codes identifying domain, endpoint and specific error
    • with human-readable message
  • Logging in Apache format for all resources
  • For all list retrieval endpoints:
    • made pagination default to (OFFSET 0, LIMIT 100)
    • introduced sorting options ASC and DESC
    • made sorting default to created_at DESC (most recently created entities returned first)

Forecasting:

  • In GET /forecast-jobs, return jobLoad

Anomaly Detection:

  • In GET /detection-jobs, return jobLoad

Datasets:

  • Added support for timestamp format "yyyy-mm-ddTHH:MM:SSZ" (including Z indicator for Zulu)

Licenses & Users:

  • Made sorting for list of users default to DESC on name (first and last)

Changes​

Forecasting:

  • Improved Swagger responses

Anomaly Detection:

  • Improved Swagger responses

Workflow:

  • Hierarchical API responses

Bug fixes​

Forecasting:

  • Create a new RCA job if execution of the existing RCA job has failed (instead of returning the ID of the existing RCA job upon attempt to register)
  • Introduced X-Tim-Client header (for /execute endpoint)
  • Return error if expectedResultsTableSize is infinite
  • Fixed bug in calculation of expectedResultsTableSize related to rollingWindow parameter for registration of a model building job

Anomaly Detection:

  • Create a new RCA job if execution of the existing RCA job has failed (instead of returning the ID of the existing RCA job upon attempt to register)
  • Fixed functionality to copy jobs

Datasets:

  • Fixed formatting of X-Total-Count response header in list retrieval endpoints

v5.1.12 [2021-07-16]​

New features​

  • Accept X-Log-Reference, X-User-Id and X-Tim-Client request headers
  • Logging of API errors

Forecasting:

  • Only one RCA job created per parent:
    • all registration attempts return the same ID
    • trying to execute an executed RCA job returns a message informing the user that the job has already been executed
  • Introduced X-Tim-Client header (for /execute endpoint)
  • In GET /forecast-jobs, added the possibility to filter on multiple job types in a single request (allowed array in type query parameters)

Anomaly Detection:

  • Only one RCA job created per parent:
    • all registration attempts return the same ID
    • trying to execute an executed RCA job returns a message informing the user that the job has already been executed
  • In GET /detection-jobs, the possibility to filter on multiple job types in a single request (allowed array in type query parameters)

Bug fixes​

  • Fixed issues with logging
  • Fixed header names

Forecasting:

  • Made POST /forecast-jobs/{id}/execute return JSON instead of plain text

Anomaly Detection:

  • Made POST /detection-jobs/{id}/execute return JSON instead of plain text

v5.1.11 [2021-07-13]​

New features​

Licenses & Users:

  • Sorting during retrieval of lists of:
    • license log
    • users
    • user log
    • user groups

Changes​

  • Unified API error responses

Forecasting:

  • Improved Swagger responses for Forecasting Job Management endpoints: removed "###" from implemented keys

Anomaly Detection:

  • Improved Swagger response for GET /detection-jobs/{id}/results/rca

Licenses & Users:

  • Hierarchical API responses

Bug fixes​

  • Fixed API error message creation in API error handler

Forecasting:

  • Removed unimplemented elements (sequences) from Swagger

Anomaly Detection:

  • Introduced X-Tim-Client header (for /execute endpoint)
  • Removed unimplemented elements (sequences) from Swagger

Datasets:

  • Fixed upload of CSV files containing format with hidden characters (quotes)

v5.1.10 [2021-07-12]​

Changes​

Forecasting:

  • Improved Swagger request body for Forecasting Job Management endpoints

Bug fixes​

Forecasting:

  • Disallowed jobs to be registered with an empty columns selection
  • Fixed bug in determination of job load category (light or heavy)

Anomaly Detection:

  • Disallowed jobs to be registered with an empty columns selection

v5.1.9 [2021-07-09]​

New features​

Forecasting:

  • Job load categorization: light and heavy job loads are recognized and send to an appropriate queue

Bug fixes​

Datasets:

  • Updated Swagger response of GET /datasets/{id}/log (renamed timestamp to createdAt for consistency)

v5.1.8 [2021-07-07]​

New features​

Forecasting:

  • During the creation of a forecast job, if no experiment ID is passed and no parent job exists, a default forecasting experiment is created in the relevant use case, in which the job is then created
  • Enabled parallel processing of API requests

Anomaly Detection:

  • During the creation of an anomaly detection job, if no experiment ID is passed and no parent job exists, a default anomaly detection experiment is created in the relevant use case, in which the job is then created
  • Added an "Info" log message about job load category
  • Enabled parallel processing of API requests

Workflow:

  • Introduced X-Tim-Client header
  • Enabled parallel processing of API requests

Changes​

  • Improved deployment pipelines

Datasets:

  • Changed response type of POSTS /datasets/{id}/versions/{version}/slice from JSON object to CSV file
  • Changed edit endpoints from PUT to PATCH
  • Improved Swagger UI for Dataset Management endpoints

Bug fixes​

Forecasting:

  • Forecasting job registration no longer fails if "from" equals "to" in rows ranges (inSampleRows and outOfSampleRows in configuration of model building and model rebuilding jobs, outOfSampleRows in configuration of predicting jobs)

Anomaly Detection:

  • Fixed handling of empty registration body
  • Anomaly detection job registration no longer fails if "from" equals "to" in rows ranges (rows in data configuration of model building, model rebuilding jobs and detection jobs)
  • Fixed failing RCA jobs related to the calculation of expectedResultsTableSize

v5.1.7 [2021-06-28]​

New features​

Anomaly Detection:

  • Job load categorization: light and heavy job loads are recognized and send to an appropriate queue

Datasets:

  • Added POST /datasets/sql/preview endpoint for downloading a preview of a dataset from a connected SQL database

Changes​

Workflow:

  • Changed edit endpoints from PUT to PATCH

v5.1.6 [2021-06-24]​

Changes​

Workflow:

  • Improved Swagger response for POST /use-cases

Bug fixes​

Workflow:

  • Fixed error handling of import and export of use cases

v5.1.5 [2021-06-23]​

New features​

Anomaly Detection:

  • Added expectedResultsTableSize (the expected size of the table returned in the job results (in MB).) to the response of
    • POST /detection-jobs/build-model
    • POST /detection-jobs/{id}/rebuild-model
    • POST /detection-jobs/{id}/detect

Changes​

Workflow:

  • Removed creation of default experiments upon creation of use case
  • Allowed default forecasting experiment and default anomaly detection experiment of a use case to be set during edit of use case
  • Improved Swagger request body for PUT /use-cases

Bug fixes​

Workflow:

  • Fixed encryption of .tw files containing an exported use case

v5.1.4 [2021-06-22]​

New features​

Forecasting:

  • Added plotlyFriendly parameter in GET /forecast-jobs/{id}/results/table (if set to true (default), gaps in timestamps for all forecastTypes and bins will be filled; if modelIndex is chosen this setting is ignored)

Datasets:

  • Enabled parallel processing of API requests

Changes​

Workflow:

  • Improved Swagger UI for GET /workspaces

Bug fixes​

Forecasting:

  • Allow predictionIntervals to be set in the configuration of a forecasting job
  • Fixed handling of explicitly set target column

Anomaly Detection:

  • Made detectionIntervals in the model building configuration require at least one interval - sending empty detectionIntervals now fails during job registration

Licenses & Users:

  • Fixed inconsistencies in License and User Management endpoints

v5.1.3 [2021-06-14]​

New features​

Forecasting:

  • Added expectedResultsTableSize (the expected size of the table returned in the job results (in MB).) to the response of
    • POST /forecast-jobs/build-model
    • POST /forecast-jobs/{id}/rebuild-model

Bug fixes​

  • Updated Swagger documentation version to v5

Workflow:

  • Fixed issue with job sizes in export of use cases
  • Fixed issue with regenerating IDs of jobs in import of use cases

v5.1.2 [2021-06-07]​

Bug fixes​

  • Fixed dynamic URL filling in Swagger

v5.1.1 [2021-06-06]​

Changes​

  • Dynamic URL filling in Swagger

v5.1.0 [2021-06-04]​

New features​

Anomaly Detection:

  • Ensured result table is sorted chronologically

Datasets:

  • Enabled parallel processing of dataset uploads

Workflow:

  • Added GET /use-cases/{id}/export for exporting of use cases
  • Added POST /use-cases/import for importing of use cases
  • Enabled additional filtering during retrieval of lists of use cases:
    • on user group

Changes​

Datasets:

  • Added custom validation for UUIDs, in:
    • GET /datasets/{id}
    • PUT /datasets/{id}
    • GET /datasets/{id}/versions/{version}
    • GET /datasets/{id}/versions/{version}/status
    • POST /datasets/{id}/versions/{version}/slice

Bug fixes​

Datasets:

  • Fixed handling of CSV files with a Byte Order Mark (BOM)
  • Queue reconnection

Forecasting:

  • Made a list of from-to ranges in configuration require at least one range
  • Updated response of GET /forecast-jobs/{id}/status (renamed timestamp to createdAt for consistency)

Anomaly Detection:

  • Made a list of from-to ranges in configuration require at least one range

Licenses & Users:

  • Made email case-insensitive

v5.0.0 [2021-05-28]​

New features​

Forecasting:

  • Forecast-related requests organized and managed as jobs
  • Forecasting jobs can be:
    • retrieved,
    • relevant logs can be retrieved, their status and the progress of an execution can be retrieved
    • registered (i.e. created, but not yet executed), available job types are:
      • model building,
      • model rebuilding,
      • predicting, and
      • root cause analysis,
    • executed (after being registered),
    • their results (in-sample, out-of-sample and production) can be retrieved,
    • the production forecast can be retrieved separately,
    • root cause analysis results can be retrieved,
    • the models that were built for a job can be retrieved,
    • the accuracies (error measures) can be retrieved,
    • copied to another experiment, and
    • deleted

Anomaly detection:

  • Anomaly detection-related requests organized and managed as jobs
  • Anomaly detection jobs can be:
    • retrieved,
    • relevant logs can be retrieved, their status and the progress of an execution can be retrieved
    • registered (i.e. created, but not yet executed), available job types are:
      • model building,
      • model rebuilding,
      • detecting, and
      • root cause analysis,
    • executed (after being registered),
    • their results can be retrieved,
    • the models that were built for a job can be retrieved,
    • root cause analysis results can be retrieved,
    • copied to another experiment, and
    • deleted

Datasets:

  • Datasets can be:
    • retrieved,
    • relevant logs can be retrieved,
    • uploaded as a CSV file,
    • uploaded via connection to an SQL database (supported options are PostgreSQL, MySQL, MariaDB and SQL_Server),
    • updated with a CSV file - this creates a new dataset version,
    • updated via SQL connection to an SQL database (supported options are PostgreSQL, MySQL, MariaDB and SQL_Server) - this creates a new dataset version,
    • updated in terms of metadata, and
    • deleted
  • Dataset versions can be:
    • retrieved,
    • their status and the progress of an upload can be retrieved
    • created, which is done by uploading or updating a dataset,
    • downloaded, and
    • deleted

Workflow:

  • Structured analytics work in a tree-like hierarchy:
    • starting from (forecasting or anomaly detection) jobs that contain molecular requests such as those of model building and model application
    • jobs are grouped in experiments, allowing users to iterate over jobs to optimize the configuration of the particular challenge at hand
    • experiments revolve around use cases, each of which is linked to a single dataset
    • use cases are organized in workspaces
  • Each of the elements in this hierarchy can be
    • retrieved,
    • created,
    • edited, and
    • deleted

Licenses & Users:

  • Support for authentication, logging in and logging out
  • License management; licenses can be:
    • retrieved
    • relevant logs and information about storage and limits can be retrieved, and
    • edited by administrator users
  • Users can be:
    • retrieved,
    • relevant logs can be retrieved,
    • created by administrator users, and
    • edited and deactivated
  • User groups provide the basis for collaboration and sharing of work
  • User groups can be
    • retrieved,
    • created,
    • edited, and
    • deleted

v4.5 [2020-10-23]​

New features​

  • added in-sample results for RTInstantML​
  • added backtesting results for RTInstantML for fast sampling rates​
  • added model view and accuracy through different forecasting horizons​
  • reworked extended importances to match (non-)daily cycle
  • improved execution time (sampling rate agnostic performance) in anomaly detection
  • added isTimeSpecific flag in the response JSON indicating whether Model Zoo is time specific
  • reworked sampling period computation for irregular data with using median

Non-breaking changes​

  • removed constraint of backtesting length in RTInstantML
  • holiday column name if do not exist in csv data return warning
  • several bugfixes for irregular data​

Breaking changes​

This release does not contain any breaking changes.

v4.4 [2020-08-31]​

New features​

Non-breaking changes​

Breaking changes​

This release does not contain any breaking changes but does not support the old Anomaly model from previous versions.

v4.3 [2020-08-06]​

New features​

Non-breaking changes​

Breaking changes​

This release does not contain any breaking changes.

v4.2 [2020-04-15]​

New features​

Non-breaking changes​

  • dictionary Simple Moving Average is now applied to all predictors (not just target)
  • added option to use CSV schema for data in JSON requests
  • reworked treemap strings
  • added an option to return normal behavior signal (forecast) in anomaly detection

Breaking changes​

This release does not contain any breaking changes.

v4.1 [2020-01-30]​

New features​

  • authentication via access tokens
  • customizable interpolation of missing values
  • new dictionaries: Identity and Simple Moving Average
  • added option to turn off offsets
  • support for data with regular gaps

Non-breaking changes​

  • data doesn't have to be aligned
  • optional HTTP headers X-Client-Name and X-Client-Version
  • changed warnings for extreme values coming to prediction to IQR standard instead of relative difference
  • turned off Piecewise Linearity dictionary for too short datasets
  • turned off Weekdays and Weekrest dictionaries for datasets with too few Mondays
  • fixed bug with predictors that are updated not every day of week
  • not using low quality predictions in anomaly detection

Breaking changes​

This release does not contain any breaking changes.

v4.0 [2019-10-15]​

New features​

  • support for general sampling rates

Non-breaking changes​

  • using JSON schemas for request validation (affects error messages)
  • API definition greatly simplified and converted to OpenAPI 3.0 (previously Swagger 2.0)

Breaking changes​

  • new licensing system, License-Key HTTP header is required
  • Api-Key HTTP header is required
  • removed "OneTime" usage (deprecated by RTInstantML)
  • removed meteo support
  • refactored anomaly detection features