2.0.0b4

Aditya514 · Nov 9, 2021 · decbacf · decbacf
1 parent 31bda03
commit decbacf
Show file tree

Hide file tree

Showing 21 changed files with 197 additions and 172 deletions.
diff --git a/MANIFEST.in b/MANIFEST.in
diff --git a/README.md b/README.md
@@ -1,9 +1,9 @@
 # Montreal Forced Aligner
 
-[![Build Status](https://travis-ci.com/MontrealCorpusTools/Montreal-Forced-Aligner.svg?branch=main)](https://travis-ci.com/MontrealCorpusTools/Montreal-Forced-Aligner)
-[![codecov](https://codecov.io/gh/MontrealCorpusTools/Montreal-Forced-Aligner/branch/master/graph/badge.svg?token=GgfM9GXFJ4)](https://codecov.io/gh/MontrealCorpusTools/Montreal-Forced-Aligner)
+![Continuous Integration](https://github.com/MontrealCorpusTools/Montreal-Forced-Aligner/actions/workflows/main.yml/badge.svg)
+[![codecov](https://codecov.io/gh/MontrealCorpusTools/Montreal-Forced-Aligner/branch/main/graph/badge.svg?token=GgfM9GXFJ4)](https://codecov.io/gh/MontrealCorpusTools/Montreal-Forced-Aligner)
 [![Documentation Status](https://readthedocs.org/projects/montreal-forced-aligner/badge/?version=latest)](http://montreal-forced-aligner.readthedocs.io/en/latest/?badge=latest)
-[![Interrogate Status](https://github.com/MontrealCorpusTools/montreal-forced-aligner/docs/source/_static/interrogate_badge.svg)](https://github.com/MontrealCorpusTools/montreal-forced-aligner/docs/source/_static/interrogate_badge.svg)
+[![Interrogate Status](https://montreal-forced-aligner.readthedocs.io/en/latest/_static/interrogate_badge.svg)](https://github.com/MontrealCorpusTools/montreal-forced-aligner/)
 [![DOI](https://zenodo.org/badge/44983969.svg)](https://zenodo.org/badge/latestdoi/44983969)
 
 The Montreal Forced Aligner is a command line utility for performing forced alignment of speech datasets using Kaldi (http://kaldi-asr.org/).

diff --git a/docs/source/api_aligner.rst → docs/source/api_reference/api_aligner.rst b/docs/source/api_aligner.rst → docs/source/api_reference/api_aligner.rst
diff --git a/docs/source/api_command_line.rst → ...source/api_reference/api_command_line.rst b/docs/source/api_command_line.rst → ...source/api_reference/api_command_line.rst
diff --git a/docs/source/api_config.rst → docs/source/api_reference/api_config.rst b/docs/source/api_config.rst → docs/source/api_reference/api_config.rst
diff --git a/docs/source/api_corpus.rst → docs/source/api_reference/api_corpus.rst b/docs/source/api_corpus.rst → docs/source/api_reference/api_corpus.rst
diff --git a/docs/source/api_g2p.rst → docs/source/api_reference/api_g2p.rst b/docs/source/api_g2p.rst → docs/source/api_reference/api_g2p.rst
diff --git a/docs/source/api_index.rst → docs/source/api_reference/api_index.rst b/docs/source/api_index.rst → docs/source/api_reference/api_index.rst
@@ -16,42 +16,66 @@ API Reference
    api_multiprocessing.rst
    api_trainers.rst
 
-.. _dictionary_api:
 
-.. automodapi:: montreal_forced_aligner.dictionary
+.. _speaker_classifier_api:
+Speaker Classifier API
+======================
 
-.. _exceptions_api:
+.. automodapi:: montreal_forced_aligner.speaker_classifier
 
-.. automodapi:: montreal_forced_aligner.exceptions
+.. _transcriber_api:
+Transcriber API
+===============
 
-.. _helper_api:
+.. automodapi:: montreal_forced_aligner.transcriber
 
-.. automodapi:: montreal_forced_aligner.helper
+.. _segmenter_api:
+Segmenter API
+=============
 
-.. _models_api:
+.. automodapi:: montreal_forced_aligner.segmenter
 
-.. automodapi:: montreal_forced_aligner.models
+.. _validator_api:
+Validator API
+=============
 
-.. _segmenter_api:
+.. automodapi:: montreal_forced_aligner.validator
 
-.. automodapi:: montreal_forced_aligner.segmenter
 
-.. _speaker_classifier_api:
+.. _dictionary_api:
+Dictionary API
+==============
 
-.. automodapi:: montreal_forced_aligner.speaker_classifier
+.. automodapi:: montreal_forced_aligner.dictionary
 
-.. _transcriber_api:
 
-.. automodapi:: montreal_forced_aligner.transcriber
+.. _models_api:
+Models API
+==========
+
+.. automodapi:: montreal_forced_aligner.models
+
 
 .. _textgrid_api:
+TextGrid API
+============
 
 .. automodapi:: montreal_forced_aligner.textgrid
 
 .. _utils_api:
 
+Utils API
+=========
 .. automodapi:: montreal_forced_aligner.utils
 
-.. _validator_api:
+.. _helper_api:
+Helper API
+==========
 
-.. automodapi:: montreal_forced_aligner.validator
+.. automodapi:: montreal_forced_aligner.helper
+
+.. _exceptions_api:
+Exceptions API
+==============
+
+.. automodapi:: montreal_forced_aligner.exceptions
diff --git a/docs/source/api_lm.rst → docs/source/api_reference/api_lm.rst b/docs/source/api_lm.rst → docs/source/api_reference/api_lm.rst
diff --git a/docs/source/api_multiprocessing.rst → ...rce/api_reference/api_multiprocessing.rst b/docs/source/api_multiprocessing.rst → ...rce/api_reference/api_multiprocessing.rst
diff --git a/docs/source/api_trainers.rst → docs/source/api_reference/api_trainers.rst b/docs/source/api_trainers.rst → docs/source/api_reference/api_trainers.rst
diff --git a/docs/source/changelog/changelog_1.0.rst b/docs/source/changelog/changelog_1.0.rst
@@ -0,0 +1,97 @@
+
+.. _changelog_1.0:
+
+1.X Changelog
+=============
+
+1.1.0
+-----
+
+Major changes to system, see :ref:`whats_new_1_1`.
+
+1.0.0
+-----
+
+- Added Grapheme-to-Phoneme capabilities
+- Acoustic models no longer contain the dictionary they were trained with
+- Dictionaries must be specified when aligning using pretrained models
+- The aligner now automatically cleans the temporary directory when the previous run failed
+- Added validation for types of command line arguments
+- Catch and list files that could not be read using UTF-8
+- Update Kaldi version to 5.1 and OpenFST version to 1.6.2 on Mac and Linux
+- Add support for specifying custom non-speech annotations in pronunciation dictionary with sil and spn
+- Made command line flags more consistent in spelling
+- Made pretrained models for many languages available
+
+0.8.0
+-----
+
+- Fixed an issue where aligning using pretrained models was improperly updating the original model with sparser data
+- Added a flag to turn off speaker adaptation when aligning using a pretrained model
+- Optimized training graph generation when aligning using a pretrained model
+
+0.7.3
+-----
+
+- Added warning messages and log output when wav files are ignored because they have too low of a sampling rate or
+  no .lab or .TextGrid file associated with them
+
+0.7.2
+-----
+
+- Fixed an issue where speaker character flags were being ignored when parsing TextGrid files
+
+0.7.1
+-----
+
+- Fixed an issue where the number of gaussians was set too low for triphone training
+
+0.7.0
+-----
+
+- Fixed an issue with unicode characters not being correctly parsed when using ``--nodict``
+- Fixed an issue where short intervals in TextGrid were not being properly ignored
+- Added a command line option ``--temp_directory`` to allow for user specification of the
+  temporary directory that MFA stores all files during alignment, with the
+  default of ``~/Documents/MFA``
+- Added logging directory and some logging for when utterances are ignored
+
+0.6.3
+-----
+
+- Improved memory and time efficiency of extracting channels from stereo
+  files, particularly for long sound files
+
+0.6.2
+-----
+
+- Fixed an issue where pretrained models were not being bundled with the source code
+
+0.6.1
+-----
+
+- Fixed an issue with Linux binaries not finding Kaldi binaries
+- English models now use all of LibriSpeech dataset and not just clean
+  subset (increased number of accents being the primary difference between the two)
+
+0.6.0
+-----
+
+- Added commandline argument ``--clean`` to remove temporary files
+- Added support for multiple sampling rates in a single dataset
+- Fix some bugs relating to using a single process
+- Fixed a bug where spaces were being inserted into transcriptions when using ``--nodict``
+- Fixed a bug where having no out-of-vocabulary items would cause a crash at the end of aligning
+- Fixed a bug where the frozen executable could not find the included pretrained models
+- Fixed an issue where dictionaries in model outputs were binary files rather than editable text files
+- Added docstrings to main classes
+- Updated built in model ``english`` for the full 1000-hour LibriSpeech corpus
+
+0.5.0
+-----
+
+- Initial release
+- Prosodylab-aligner format supported
+- TextGrid format supported
+- Align using pretrained models supported
+- Train models and align concurrently supported
diff --git a/docs/source/changelog.rst → docs/source/changelog/changelog_2.0.rst b/docs/source/changelog.rst → docs/source/changelog/changelog_2.0.rst
@@ -3,10 +3,35 @@
 .. _`PR #288`: https://github.com/MontrealCorpusTools/Montreal-Forced-Aligner/pull/288
 .. _`PR #337`: https://github.com/MontrealCorpusTools/Montreal-Forced-Aligner/pull/337
 
-.. _changelog:
+.. _changelog_2.0:
+
+*************
+2.0 Changelog
+*************
+
+.. _2.0b:
+
+Beta releases
+=============
+
+2.0.0b4
+-------
+
+- Massive refactor to a proper class-based API for interacting with MFA corpora
+
+  - Sorry, I really do hope this is the last big refactor of 2.0
+  - :class:`~montreal_forced_aligner.corpus.classes.Speakers`, :class:`~montreal_forced_aligner.corpus.classes.Files`, and :class:`~montreal_forced_aligner.corpus.classes.Utterances` have dedicated classes rather than having their information split across dictionaries mimicking Kaldi files, so they should be more useful for interacting with outside of MFA
+  - Added :class:`~montreal_forced_aligner.multiprocessing.classes.Job` class as well to make it easier to generate and keep track of information about different processes
+- Updated installation style to be more dependent on conda-forge packages
+
+  - Kaldi is now on conda-forge! |:tada:|
+  - MFA should follow suit soon, making installation a lot simpler
+
+- Added a :code:`mfa model` command for inspecting, listing, downloading, and saving pretrained models, see :ref:`pretrained_models` for more information.
+- Fixed a bug where saving command history with errors would throw an error of its own
+- Fixed an issue where one Job could process another Job's data, result in an error
+- Updated API documentation to reflect refactor changes
 
-Changelog
-=========
 
 2.0.0b3
 -------
@@ -35,6 +60,10 @@ Beta release!
 - Refactored and optimized the TextGrid export process to use multiple processes by default, you should be significant speed ups.
 - Removed shorthand flags for ``-c`` and ``-d`` since they could represent multiple different flags/arguments.
 
+.. _2.0a:
+
+2.0 alpha releases
+==================
 
 2.0.0a24
 --------
@@ -210,96 +239,3 @@ Currently under development with major changes, see :ref:`whats_new_2_0`.
 - Removed multiprocessing from speaker adaptation, as the executables use multiple threads leading to a bottleneck in
   performance.  This change should result in faster speaker adaptation.
 - Optimized corpus parsing algorithm to be O(n log n) instead of O(n^2) (`PR #194`_)
-
-
-1.1.0
------
-
-Major changes to system, see :ref:`whats_new_1_1`.
-
-1.0.0
------
-
-- Added Grapheme-to-Phoneme capabilities
-- Acoustic models no longer contain the dictionary they were trained with
-- Dictionaries must be specified when aligning using pretrained models
-- The aligner now automatically cleans the temporary directory when the previous run failed
-- Added validation for types of command line arguments
-- Catch and list files that could not be read using UTF-8
-- Update Kaldi version to 5.1 and OpenFST version to 1.6.2 on Mac and Linux
-- Add support for specifying custom non-speech annotations in pronunciation dictionary with sil and spn
-- Made command line flags more consistent in spelling
-- Made pretrained models for many languages available
-
-0.8.0
------
-
-- Fixed an issue where aligning using pretrained models was improperly updating the original model with sparser data
-- Added a flag to turn off speaker adaptation when aligning using a pretrained model
-- Optimized training graph generation when aligning using a pretrained model
-
-0.7.3
------
-
-- Added warning messages and log output when wav files are ignored because they have too low of a sampling rate or
-  no .lab or .TextGrid file associated with them
-
-0.7.2
------
-
-- Fixed an issue where speaker character flags were being ignored when parsing TextGrid files
-
-0.7.1
------
-
-- Fixed an issue where the number of gaussians was set too low for triphone training
-
-0.7.0
------
-
-- Fixed an issue with unicode characters not being correctly parsed when using ``--nodict``
-- Fixed an issue where short intervals in TextGrid were not being properly ignored
-- Added a command line option ``--temp_directory`` to allow for user specification of the
-  temporary directory that MFA stores all files during alignment, with the
-  default of ``~/Documents/MFA``
-- Added logging directory and some logging for when utterances are ignored
-
-0.6.3
------
-
-- Improved memory and time efficiency of extracting channels from stereo
-  files, particularly for long sound files
-
-0.6.2
------
-
-- Fixed an issue where pretrained models were not being bundled with the source code
-
-0.6.1
------
-
-- Fixed an issue with Linux binaries not finding Kaldi binaries
-- English models now use all of LibriSpeech dataset and not just clean
-  subset (increased number of accents being the primary difference between the two)
-
-0.6.0
------
-
-- Added commandline argument ``--clean`` to remove temporary files
-- Added support for multiple sampling rates in a single dataset
-- Fix some bugs relating to using a single process
-- Fixed a bug where spaces were being inserted into transcriptions when using ``--nodict``
-- Fixed a bug where having no out-of-vocabulary items would cause a crash at the end of aligning
-- Fixed a bug where the frozen executable could not find the included pretrained models
-- Fixed an issue where dictionaries in model outputs were binary files rather than editable text files
-- Added docstrings to main classes
-- Updated built in model ``english`` for the full 1000-hour LibriSpeech corpus
-
-0.5.0
------
-
-- Initial release
-- Prosodylab-aligner format supported
-- TextGrid format supported
-- Align using pretrained models supported
-- Train models and align concurrently supported
diff --git a/docs/source/changelog/changelog_index.rst b/docs/source/changelog/changelog_index.rst
@@ -0,0 +1,12 @@
+
+.. _changelog:
+
+*********
+Changelog
+*********
+
+.. toctree::
+   :maxdepth: 3
+
+   changelog_2.0.rst
+   changelog_1.0.rst
diff --git a/docs/source/conf.py b/docs/source/conf.py
@@ -41,6 +41,7 @@
     "sphinx_automodapi.automodapi",
     "sphinx_automodapi.smart_resolver",
     "sphinx.ext.intersphinx",
+    "sphinxemoji.sphinxemoji",
 ]
 
 intersphinx_mapping = {"python": ("https://docs.python.org/3", None)}