v0.4.0

What's Changed

• Violin pitch extraction and few improvements by @genisplaja in #98
• Zenodo integration+model versioning by @genisplaja in #100
• melodic pattern finding notebook by @thomasgnuttall in #101
• Use compmusic==0.4 for Corpora access by @genisplaja in #102
• Move to pyproject, remove tox by @genisplaja in #103
• mixer-model to separate vocals+violin by @genisplaja in #104

Full Changelog: v0.3.0...v0.4.0

v0.3.0

What's Changed

• Fix Dunya API call for raaga and taala by @genisplaja in #78 (powered by attendees in the Compmusic Workshop 2022 in Chennai, India)
• Update documentation workflow by @genisplaja in #80
• Added Kamavardini scale and fixed Svara to cent mapping by @Prospekt42 in #79
• Allow array to be input, update documentation, minor fixes by @genisplaja in #81
• Update akshara pulse tracker (rhythm extractor) and fix tool listing functions by @genisplaja in #84
• Remove prints, use logging by @genisplaja in #85
• Remove unnecesary files by @thomasgnuttall in #87
• Carnatic separation, removing large models, and minor fixes by @genisplaja in #92
• Downloadable segmentation model, variable storing dir for models by @genisplaja in #93
• Small tests and further improvements before release by @genisplaja in #94
• Final doc updates for v0.3.0 by @genisplaja in #96

New Contributors

• @Prospekt42 made their first contribution in #79

Full Changelog: v0.2.1...v0.3.0

v0.2.1

What's Changed

• load_yaml rreinstated by @thomasgnuttall in #76

Full Changelog: v0.2.0...v0.2.1

v0.2.0

What is included?

• First fully functional version of compiam after the official launch during ISMIR 22.
• Access to 8 datasets through mirdata data loaders.
• Access to Carnatic and Hindustani Corpora in Dunya.
• Access to the following tools and models: Indian Tonic Multi-Pitch, Melodia, FTANet-Carnatic, DEEPSRGM, CAE + self-similarity matrix, Akshara Pulse Tracker, Mnemonic Percussion Transcription, Dhrupad Bandish Segmentation, Mridangam Stroke Classification.
• Tools to easily browse available datasets, corpora, and tools in the package.
• Wrappers to easily access datasets, corpora, and models with available pre-trained weights.
• Pre-trained model weights for FTANet-Carnatic, DEEPSRGM, CAE + self-similarity matrix, Dhrupad Bandish Segmentation.
• Usage and contributing documentation.
• First batch of example notebooks for compiam usage and tool showcase (available in the repo).

Full Changelog: https://github.com/MTG/compIAM/commits/v0.2.0