version 1.3.0

Changes

• Fixing linters, remove circle-ci from project (#102)
• Add support for ai.config command (#98)
• Upgrade to redis-py 4.1 (#94)
• Disabling poetry virtualenv control in ci (#92)

(Replacing 1.2.3 release and upgrading to a new minor version due to API addition)

Version 1.2.3

Changes

• Fixing linters, remove circle-ci from project (#102)
• Add support for ai.config command (#98)
• Upgrade to redis-py 4.1 (#94)
• Disabling poetry virtualenv control in ci (#92)

Version 1.2.2

Changes

• Add support in another two types of tensors: String tensor (#90) and Boolean tensors (#89)
• Allow using device id in modelstore command. That is, the device string can be "cpu/gpu:" where n is an integer, and not just "cpu/gpu" as it was until now. (#91)

Version 1.2.1

Changes

• Version and redis pin (#88)

v1.2.0

First release supporting RedisAI 1.2
Supports RedisAI 1.2.4 and up

Changes

• replacing the version in the toml with the tag (#78)
• Update config.yml - lint ci is 30m (#75)
• release changes (#74)
• Add dagexecute support (#69)
• SCRIPTEXECUTE support (#68)
• Tflite test fix (#73)
• Fix bug in tflite test (#72)
• modern python (#65)
• Add MODELSTORE and MODELEXECUTE (#60)
• venv packages spelling fix (#64)
• pointing to edge docker (#63)

Version 1.0.2

Changes

• Restructuring and formatting @hhsecond (#56)
• TESNORGET as mutable numpy ndarray @alonre24 (#54)
• Remove last special character from DAG before run @hhsecond (#53)
• Fix DAGRUN with no LOAD or PERSIST @lantiga (#46)
• [WIP] Pipeline command @hhsecond (#41)
• Update test.py @gkorland (#40)
• RTD test @hhsecond (#38)

Maintenance

• Added release automation (creates drafts automatically and pushes to pypi on release) @filipecosta90 (#50)

Release 1.0.1

Bug Fixes & Improvements

• PyPI documentation format issue
• Better error message on wrong device and wrong backend

Release 1.0.0

Bug fixes & improvements

• More pythonic APIs
• Numpy objects are the default tensors
• redisai specific DType, Backend and Device objects are gone in favor of native python strings
• Standardization of return data type
• SCRIPTLIST & MODELLIST -> SCRIPTSCAN & MODELSCAN
• Documentation in ReadTheDocs

New features

• Automatic chunking of big models
• DAG

Release 0.5.0

• TFLite support
• Tensor API stabilization
  • tensorget expect numpy arrays by default
  • tensorset returns numpy array by default
• more test cases

Release 0.4.1

Bugfix:

• pip install with python2.7 raises an exception now
• mypy uses type supported by early 3.xs