Remove dataclassy for Python 3.14 compatibility (#947)

Author: mmcauliffe

ci/docker_environment.yaml

@@ -10,7 +10,6 @@ dependencies:
   - tqdm
   - requests
   - pyyaml
-  - dataclassy
   - kaldi=*=*cpu*
   - pynini
   - openfst

environment.yml

@@ -9,7 +9,6 @@ dependencies:
   - tqdm
   - requests
   - pyyaml
-  - dataclassy
   - kaldi=*=*cpu*
   - scipy
   - pynini

github_environment.yml

@@ -2,13 +2,12 @@ channels:
   - conda-forge
 dependencies:
   - python=3.12
-  - numpy<2.1.0
+  - numpy<2.4.0
   - librosa
   - pysoundfile
   - tqdm
   - requests
   - pyyaml
-  - dataclassy
   - kaldi=*=*cpu*
   - scipy
   - pynini

montreal_forced_aligner/acoustic_modeling/lda.py

@@ -1,6 +1,7 @@
 """Class definitions for LDA trainer"""
 from __future__ import annotations
 
+import dataclasses
 import logging
 import os
 import shutil
@@ -35,6 +36,7 @@
 logger = logging.getLogger("mfa")
 
 
+@dataclasses.dataclass(slots=True)
 class LdaAccStatsArguments(MfaArguments):
     """Arguments for :func:`~montreal_forced_aligner.acoustic_modeling.lda.LdaAccStatsFunction`"""
 
@@ -43,6 +45,7 @@ class LdaAccStatsArguments(MfaArguments):
     lda_options: MetaDict
 
 
+@dataclasses.dataclass(slots=True)
 class CalcLdaMlltArguments(MfaArguments):
     """Arguments for :func:`~montreal_forced_aligner.acoustic_modeling.lda.CalcLdaMlltFunction`"""
 

montreal_forced_aligner/acoustic_modeling/monophone.py

@@ -1,6 +1,7 @@
 """Class definitions for Monophone trainer"""
 from __future__ import annotations
 
+import dataclasses
 import logging
 import typing
 from pathlib import Path
@@ -27,6 +28,7 @@
 logger = logging.getLogger("mfa")
 
 
+@dataclasses.dataclass(slots=True)
 class MonoAlignEqualArguments(MfaArguments):
     """Arguments for :func:`~montreal_forced_aligner.acoustic_modeling.monophone.MonoAlignEqualFunction`"""
 

montreal_forced_aligner/acoustic_modeling/sat.py

@@ -1,6 +1,7 @@
 """Class definitions for Speaker Adapted Triphone trainer"""
 from __future__ import annotations
 
+import dataclasses
 import logging
 import os
 import shutil
@@ -36,6 +37,7 @@
 logger = logging.getLogger("mfa")
 
 
+@dataclasses.dataclass(slots=True)
 class AccStatsTwoFeatsArguments(MfaArguments):
     """Arguments for :func:`~montreal_forced_aligner.acoustic_modeling.sat.AccStatsTwoFeatsFunction`"""
 

montreal_forced_aligner/acoustic_modeling/trainer.py

@@ -8,6 +8,7 @@
 import shutil
 import time
 import typing
+from dataclasses import dataclass
 from pathlib import Path
 from typing import TYPE_CHECKING, Any, Dict, List, Optional, Tuple
 
@@ -42,14 +43,10 @@
 from montreal_forced_aligner.utils import log_kaldi_errors, run_kaldi_function
 
 if TYPE_CHECKING:
-    from dataclasses import dataclass
-
     from montreal_forced_aligner.acoustic_modeling.base import AcousticModelTrainingMixin
     from montreal_forced_aligner.acoustic_modeling.pronunciation_probabilities import (
         PronunciationProbabilityTrainer,
     )
-else:
-    from dataclassy import dataclass
 
 __all__ = ["TrainableAligner", "TransitionAccFunction", "TransitionAccArguments"]
 

montreal_forced_aligner/acoustic_modeling/triphone.py

@@ -1,6 +1,7 @@
 """Class definitions for TriphoneTrainer"""
 from __future__ import annotations
 
+import dataclasses
 import logging
 import os
 import typing
@@ -34,13 +35,15 @@
 logger = logging.getLogger("mfa")
 
 
+@dataclasses.dataclass(slots=True)
 class TreeStatsArguments(MfaArguments):
     """Arguments for :func:`~montreal_forced_aligner.acoustic_modeling.triphone.tree_stats_func`"""
 
     working_directory: Path
     model_path: Path
 
 
+@dataclasses.dataclass(slots=True)
 class ConvertAlignmentsArguments(MfaArguments):
     """Arguments for :func:`~montreal_forced_aligner.acoustic_modeling.triphone.ConvertAlignmentsFunction`"""
 

montreal_forced_aligner/alignment/adapting.py

@@ -287,6 +287,7 @@ def meta(self) -> MetaDict:
                 "clitic_marker": self.clitic_marker,
                 "position_dependent_phones": self.position_dependent_phones,
             },
+            "language": str(self.language),
             "oov_phone": self.oov_phone,
             "optional_silence_phone": self.optional_silence_phone,
             "silence_probability": self.silence_probability,

montreal_forced_aligner/alignment/base.py

@@ -282,9 +282,7 @@ def analyze_alignments(self, calculate_duration_statistics=True):
                     )
                     .join(Utterance.file)
                     .join(Utterance.speaker)
-                    .order_by(
-                        sqlalchemy.desc(Utterance.duration_deviation)
-                    )
+                    .order_by(sqlalchemy.desc(Utterance.duration_deviation))
                 )
                 for row in utterances:
                     writer.writerow([*row])
@@ -1080,7 +1078,7 @@ def collect_alignments(self) -> None:
                         self.db_path.as_posix(),
                         "--cmd",
                         ".mode csv",
-                        f".import {word_csv_path.as_posix()} word_interval_temp",
+                        f'.import "{word_csv_path.as_posix()}" word_interval_temp',
                     ]
                 )
             subprocess.check_call(
@@ -1089,7 +1087,7 @@ def collect_alignments(self) -> None:
                     self.db_path.as_posix(),
                     "--cmd",
                     ".mode csv",
-                    f".import {phone_csv_path.as_posix()} phone_interval_temp",
+                    f'.import "{phone_csv_path.as_posix()}" phone_interval_temp',
                 ]
             )
         with self.session() as session:
@@ -1419,7 +1417,7 @@ def evaluate_alignments(
             to_comp = []
             score_func = functools.partial(
                 align_phones,
-                silence_phone=self.optional_silence_phone,
+                silence_phones={self.optional_silence_phone},
                 custom_mapping=mapping,
                 debug=config.DEBUG,
             )
@@ -1472,7 +1470,8 @@ def evaluate_alignments(
                 to_comp.append((reference_phones, comparison_phones))
             with ThreadPool(1) as pool:
                 gen = pool.starmap(score_func, to_comp)
-                for i, (score, phone_error_rate, errors) in enumerate(gen):
+                for i, result in enumerate(gen):
+                    score, phone_error_rate, errors = result[:3]
                     if score is None:
                         continue
                     u = indices[i]