3.0 bug fixes (#768)

* 3.0 bug fixes * Fix praatio version in setup.cfg * Fix for pool size bug
MontrealCorpusTools · Feb 27, 2024 · 4657517 · 4657517
1 parent 95c701e
commit 4657517
Show file tree

Hide file tree

Showing 7 changed files with 12 additions and 11 deletions.
diff --git a/.github/ISSUE_TEMPLATE/bug_report.md b/.github/ISSUE_TEMPLATE/bug_report.md
@@ -10,7 +10,7 @@ assignees: mmcauliffe
 **Debugging checklist**
 
 [ ] Have you read the troubleshooting page (https://montreal-forced-aligner.readthedocs.io/en/latest/user_guide/troubleshooting.html) and searched the documentation to ensure that your issue is not addressed there?
-[ ] Have you updated to latest MFA version? What is the output of `mfa version`?
+[ ] Have you updated to latest MFA version (check https://montreal-forced-aligner.readthedocs.io/en/latest/changelog/changelog_3.0.html)? What is the output of `mfa version`?
 [ ] Have you tried rerunning the command with the `--clean` flag?
 
 **Describe the issue**

diff --git a/docs/source/changelog/changelog_3.0.rst b/docs/source/changelog/changelog_3.0.rst
@@ -5,6 +5,12 @@
 3.0 Changelog
 *************
 
+3.0.1
+=====
+
+- Fixed an issue where pool size would be too low for number of jobs
+- Fixed an issue with specifying :code:`--phone_groups_path` causing a crash
+
 3.0.0
 =====
 

diff --git a/docs/source/user_guide/troubleshooting.rst b/docs/source/user_guide/troubleshooting.rst
@@ -11,7 +11,7 @@ Errors aligning single files
 
 In general, MFA is not intended to align single files, particularly if they are long, have noise in the background, a different style such a singing etc.  Some aspects of these can be improved by aligning a larger set of utterances per speaker, as successful alignments on the first pass can aid second pass alignments through feature-space transforms.
 
-If you must align a single file and run into an error, you can increase the beam width via `--beam 100`, this will result in less optimal alignments being generated.  The beam width is intentionally set low so that the fMLLR feature space transform for speaker adaptation is using quality alignments and not potential noise or bad alignments.
+If you must align a single file and run into an error, you can increase the beam width via :code:`--beam 100`, this will result in less optimal alignments being generated.  The beam width is intentionally set low so that the fMLLR feature space transform for speaker adaptation is using quality alignments and not potential noise or bad alignments.
 
 .. seealso::
 

diff --git a/montreal_forced_aligner/abc.py b/montreal_forced_aligner/abc.py
@@ -402,8 +402,8 @@ def construct_engine(self, **kwargs) -> sqlalchemy.engine.Engine:
         if not config.USE_POSTGRES:
             if kwargs.pop("read_only", False):
                 db_string += "?mode=ro&nolock=1&uri=true"
-        kwargs["pool_size"] = 10
-        kwargs["max_overflow"] = 10
+        kwargs["pool_size"] = config.NUM_JOBS + 10
+        kwargs["max_overflow"] = config.NUM_JOBS + 10
         e = sqlalchemy.create_engine(
             db_string,
             **kwargs,

diff --git a/montreal_forced_aligner/command_line/train_acoustic_model.py b/montreal_forced_aligner/command_line/train_acoustic_model.py
@@ -123,7 +123,6 @@ def train_acoustic_model_cli(context, **kwargs) -> None:
     corpus_directory = kwargs["corpus_directory"].absolute()
     dictionary_path = kwargs["dictionary_path"]
     g2p_model_path = kwargs.get("g2p_model_path", None)
-    phone_groups_path = kwargs.get("phone_groups_path", None)
     if kwargs.get("phone_set_type", "UNKNOWN") != "UNKNOWN":
         import warnings
 
@@ -139,7 +138,6 @@ def train_acoustic_model_cli(context, **kwargs) -> None:
         corpus_directory=corpus_directory,
         dictionary_path=dictionary_path,
         g2p_model_path=g2p_model_path,
-        phone_groups_path=phone_groups_path,
         **TrainableAligner.parse_parameters(config_path, context.params, context.args),
     )
     try:

diff --git a/requirements.txt b/requirements.txt
@@ -1,4 +1,4 @@
-praatio>=6.0
+praatio>=6.0.0
 tqdm
 pyyaml
 librosa

diff --git a/setup.cfg b/setup.cfg
@@ -17,9 +17,6 @@ classifiers = Development Status :: 3 - Alpha
     Programming Language :: Python
     Programming Language :: Python :: 3
     Programming Language :: Python :: 3 :: Only
-    Programming Language :: Python :: 3.8
-    Programming Language :: Python :: 3.9
-    Programming Language :: Python :: 3.10
     Programming Language :: Python :: Implementation :: CPython
     Topic :: Multimedia :: Sound/Audio :: Speech
     Topic :: Scientific/Engineering
@@ -43,7 +40,7 @@ install_requires =
     librosa
     matplotlib
     numpy
-    praatio>=5.0
+    praatio>=6.0.0
     pyyaml
     requests
     rich