v0.4.2

This is a minor release, with updates to the documentation, and various minor improvements:

• Parameterize dtype for TFHubExtractor, add ImageResizingFilter + ImageRescalingFilter (#459)
• Add Librosa fundamental frequency (#461)
• Updated documentation (#445, #448, #447, #446), moved to ReadTheDocs (#482), and added video to README (#454)
• Add more information to docstring for MFCCEnergyExtractor (#463)
• Import ABC from collections.abc for Python 3.10 compatibility (#467)

The following extractor was deprecated:

• Remove MicrosoftAPIFaceEmotionExtractor (#471)

Bug fixes and housekeeping:

• FIX: Google APIs (remove discoveryServiceUrl) (#478)
• FIX: Tests (#457, #477)
• Move to Clarifai-GRPC to fix Clarifai extractors (#474)
• FIX: Change excel engine (#455)
• Build docker image with github actions (#452)
• Update links from Naturalistic Data course (#458)

v0.4.1

This is a minor release that adds one new extractor, and a resampling utility, as well as a large number of minor bugfixes.

New features:

• Add TensorFlow hub extractor (#433)
• Add ExtractedFeature resample utility (#430)

Bug fixes and other house keeping

• Replace BERT with lighter extract in metric extractor tests
• Fix pytest memory leak by running forked (#439)
• Remove depracated pandas methods (#438)
• Convert CI to Github Actions (#436)
• Update HuggingFace BERT Extractors (#435)
• Resample upsampling and add upsampling test case (#431)
• Librosa links (#429)
• Google API Regressions (#424)
• Access string in numpy array (#422)
• Update REAME to link to correct pliers docs page (#421)
• Test regressions (#418)
• Remove --no-user-group (#414)
• Use multi-stage builds in Dockerfile to reduce final image size (#411)
• Remove instance of Python2 super (#410)
• Update librosa feature links (#409)
• If Google Face Extractor result landmark has no type, skip. (#406)
• librosa import error (numba) (#405)
• Remove soundfile as dependency (#408)