Bug fixes (MontrealCorpusTools#369)

Resolves MontrealCorpusTools#368
Resolves MontrealCorpusTools#359
Resolves MontrealCorpusTools#363

.github/workflows/main.yml

@@ -20,7 +20,7 @@ jobs:
       - uses: "actions/checkout@v2"
       - uses: "actions/setup-python@v2"
         with:
-          python-version: "3.9"
+          python-version: "3.8"
       - name: Install dependencies
         run: |
           python -m pip install --upgrade pip

ci/mfa_publish.yml

@@ -3,6 +3,7 @@ channels:
   - conda-forge
   - defaults
 dependencies:
+  - python>=3.8
   - numpy
   - librosa
   - tqdm

docs/source/_templates/sidebar-nav-bs.html

@@ -1,7 +1,6 @@
 <nav class="bd-links" id="bd-docs-nav" aria-label="{{ _('Main navigation') }}">
   <div class="bd-toc-item active">
     {{ generate_nav_html("sidebar",
-                         show_nav_level=theme_show_nav_level|int,
                          maxdepth=theme_navigation_depth|int,
                          collapse=theme_collapse_navigation|tobool,
                          includehidden=True,

docs/source/changelog/changelog_2.0.rst

@@ -10,6 +10,21 @@
 Beta releases
 =============
 
+2.0.0b9
+-------
+
+- Fixed a bug where unknown word phones were showing up as blank
+- Fixed a bug where TextGrid export would hang
+- Fixed compatibility issues with Python 3.8
+- Added logging for when configuration parameters are ignored
+- Added some functionality from the LibriSpeech recipe for triphone training with Arpabet
+
+  - Not sure if it'll improve anything, but I'll run some tests and maybe extend it to other phone sets
+
+- Added better logging to TextGrid export
+- Added new classes for managing collections of utterances, speakers, and files
+- Fixed a bug where oovs were not being properly reported by the validation tool
+
 2.0.0b8
 -------
 

docs/source/external_links.py

@@ -17,7 +17,7 @@
     :license: BSD, see LICENSE for details.
 """
 
-from typing import Any
+from typing import Any, Dict, List, Tuple
 
 import sphinx
 from docutils import nodes, utils
@@ -42,8 +42,8 @@ def model_role(
     lineno: int,
     inliner: Inliner,
     options: dict = None,
-    content: list[str] = None,
-) -> tuple[list[Node], list[system_message]]:
+    content: List[str] = None,
+) -> Tuple[List[Node], List[system_message]]:
     text = utils.unescape(text)
     model_type, model_name = text.split("/")
     full_url = f"https://github.com/MontrealCorpusTools/mfa-models/raw/main/{model_type}/{model_name.lower()}.zip"
@@ -59,8 +59,8 @@ def kaldi_steps_role(
     lineno: int,
     inliner: Inliner,
     options: dict = None,
-    content: list[str] = None,
-) -> tuple[list[Node], list[system_message]]:
+    content: List[str] = None,
+) -> Tuple[List[Node], List[system_message]]:
     text = utils.unescape(text)
     full_url = f"https://github.com/kaldi-asr/kaldi/tree/master/egs/wsj/s5/steps/{text}.sh"
     title = f"{text}.sh"
@@ -75,8 +75,8 @@ def kaldi_utils_role(
     lineno: int,
     inliner: Inliner,
     options: dict = None,
-    content: list[str] = None,
-) -> tuple[list[Node], list[system_message]]:
+    content: List[str] = None,
+) -> Tuple[List[Node], List[system_message]]:
     filename = utils.unescape(text)
     full_url = f"https://github.com/kaldi-asr/kaldi/tree/master/egs/wsj/s5/utils/{filename}"
     title = f"{text}"
@@ -91,8 +91,8 @@ def kaldi_steps_sid_role(
     lineno: int,
     inliner: Inliner,
     options: dict = None,
-    content: list[str] = None,
-) -> tuple[list[Node], list[system_message]]:
+    content: List[str] = None,
+) -> Tuple[List[Node], List[system_message]]:
     text = utils.unescape(text)
     full_url = f"https://github.com/kaldi-asr/kaldi/tree/cbed4ff688a172a7f765493d24771c1bd57dcd20/egs/sre08/v1/sid/{text}.sh"
     title = f"sid/{text}.sh"
@@ -107,8 +107,8 @@ def kaldi_docs_role(
     lineno: int,
     inliner: Inliner,
     options: dict = None,
-    content: list[str] = None,
-) -> tuple[list[Node], list[system_message]]:
+    content: List[str] = None,
+) -> Tuple[List[Node], List[system_message]]:
     text = utils.unescape(text)
     t = text.split("#")
     text = t[0]
@@ -130,8 +130,8 @@ def openfst_src_role(
     lineno: int,
     inliner: Inliner,
     options: dict = None,
-    content: list[str] = None,
-) -> tuple[list[Node], list[system_message]]:
+    content: List[str] = None,
+) -> Tuple[List[Node], List[system_message]]:
     text = utils.unescape(text)
     full_url = f"https://www.openfst.org/doxygen/fst/html/{text}-main_8cc_source.html"
     title = f"OpenFst {text} source"
@@ -146,8 +146,8 @@ def kaldi_src_role(
     lineno: int,
     inliner: Inliner,
     options: dict = None,
-    content: list[str] = None,
-) -> tuple[list[Node], list[system_message]]:
+    content: List[str] = None,
+) -> Tuple[List[Node], List[system_message]]:
     text = utils.unescape(text)
     mapping = {
         "bin": set(
@@ -379,8 +379,8 @@ def xref(
     lineno: int,
     inliner: Inliner,
     options: dict = None,
-    content: list[str] = None,
-) -> tuple[list[Node], list[system_message]]:
+    content: List[str] = None,
+) -> Tuple[List[Node], List[system_message]]:
 
     title = target = text
     # look if explicit title and target are given with `foo <bar>` syntax
@@ -409,7 +409,7 @@ def get_refs(app):
     xref.links = app.config.xref_links
 
 
-def setup(app: Sphinx) -> dict[str, Any]:
+def setup(app: Sphinx) -> Dict[str, Any]:
     app.add_config_value("xref_links", {}, "env")
     app.add_role("mfa_model", model_role)
     app.add_role("kaldi_steps", kaldi_steps_role)

docs/source/reference/corpus/index.rst

@@ -30,6 +30,19 @@ Corpora
 Helper classes and functions
 ============================
 
+Collections
+-----------
+
+.. currentmodule::  montreal_forced_aligner.corpus.classes
+
+.. autosummary::
+   :toctree: generated/
+
+   Collection
+   SpeakerCollection
+   FileCollection
+   UtteranceCollection
+
 Multiprocessing
 ---------------
 

docs/source/user_guide/configuration/global.rst

@@ -68,8 +68,7 @@ The :code:`multilingual_ipa`, :code:`strip_diacritics`, and :code:`digraphs` are
    "oov_word", "<unk>", "Internal word symbol to use for out of vocabulary items"
    "oov_phone", "spn", "Internal phone symbol to use for out of vocabulary items"
    "silence_word", "!sil", "Internal word symbol to use initial silence"
-   "nonoptional_silence_phone", "sil", "Internal phone symbol to use initial silence"
-   "optional_silence_phone", "sp", "Internal phone symbol to use optional silence in the middle of utterances"
+   "optional_silence_phone", "sp", "Internal phone symbol to use optional silence in or around utterances"
    "position_dependent_phones", "True", "Flag for whether phones should mark their position in the word as part of the phone symbol internally"
    "num_silence_states", "5", "Number of states to use for silence phones"
    "num_non_silence_states", "3", "Number of states to use for non-silence phones"

environment.yml

@@ -2,7 +2,7 @@ name: mfa
 channels:
   - conda-forge
 dependencies:
-  - python>=3.9
+  - python>=3.8
   - numpy
   - librosa
   - tqdm
@@ -16,3 +16,4 @@ dependencies:
   - ngram
   - pynini
   - praatio
+  - biopython

environment_win.yml

@@ -2,7 +2,7 @@ name: montreal-forced-aligner
 channels:
   - conda-forge
 dependencies:
-  - python>=3.9
+  - python>=3.8
   - numpy
   - librosa
   - tqdm
@@ -12,3 +12,4 @@ dependencies:
   - kaldi
   - sox
   - praatio
+  - biopython