Anchor-annotator 0.0.11__py3-none-any.whl → 0.2.0__py3-none-any.whl

anchor/widgets.py

@@ -25,6 +25,8 @@ from anchor.models import (
     CorpusSelectionModel,
     DiarizationModel,
     DictionaryTableModel,
+    FileSelectionModel,
+    FileUtterancesModel,
     OovModel,
     SpeakerModel,
     TextFilterQuery,
@@ -61,7 +63,6 @@ class MediaPlayer(QtMultimedia.QMediaPlayer):  # pragma: no cover
         self.max_time = None
         self.start_load_time = None
         self.min_time = None
-        self.corpus_model = None
         self.selection_model = None
         self.timer = QtCore.QTimer(self)
         self.timer.setInterval(1)
@@ -78,7 +79,6 @@ class MediaPlayer(QtMultimedia.QMediaPlayer):  # pragma: no cover
         self._audio_output.setDevice(self.devices.defaultAudioOutput())
         self.setAudioOutput(self._audio_output)
         self.playbackStateChanged.connect(self.reset_position)
-        self.mediaStatusChanged.connect(self.update_load)
         self.fade_in_anim = QtCore.QPropertyAnimation(self._audio_output, b"volume")
         self.fade_in_anim.setDuration(10)
         self.fade_in_anim.setStartValue(0.1)
@@ -95,11 +95,6 @@ class MediaPlayer(QtMultimedia.QMediaPlayer):  # pragma: no cover
         self.fade_out_anim.finished.connect(super().pause)
         self.file_path = None
 
-    def update_load(self, state):
-        if state == self.MediaStatus.LoadedMedia:
-            self.reset_position()
-            self.audioReady.emit(True)
-
     def handle_error(self, *args):
         print("ERROR")
         print(args)
@@ -118,12 +113,22 @@ class MediaPlayer(QtMultimedia.QMediaPlayer):  # pragma: no cover
         self.fade_in_anim.start()
 
     def startTime(self):
-        if self.selection_model.selected_min_time is not None:
+        if (
+            self.selection_model.selected_min_time is not None
+            and self.selection_model.min_time
+            <= self.selection_model.selected_min_time
+            <= self.selection_model.max_time
+        ):
             return self.selection_model.selected_min_time
         return self.selection_model.min_time
 
     def maxTime(self):
-        if self.selection_model.selected_max_time is not None:
+        if (
+            self.selection_model.selected_max_time is not None
+            and self.selection_model.min_time
+            <= self.selection_model.selected_max_time
+            <= self.selection_model.max_time
+        ):
             return self.selection_model.selected_max_time
         return self.selection_model.max_time
 
@@ -149,14 +154,10 @@ class MediaPlayer(QtMultimedia.QMediaPlayer):  # pragma: no cover
                 break
         self._audio_output.setDevice(o)
 
-    def set_corpus_models(
-        self, corpus_model: Optional[CorpusModel], selection_model: Optional[CorpusSelectionModel]
-    ):
-        self.corpus_model = corpus_model
-        self.selection_model = selection_model
-        if corpus_model is None:
+    def set_models(self, selection_model: Optional[FileSelectionModel]):
+        if selection_model is None:
             return
-        # self.selection_model.fileAboutToChange.connect(self.unload_file)
+        self.selection_model = selection_model
         self.selection_model.fileChanged.connect(self.loadNewFile)
         self.selection_model.viewChanged.connect(self.update_times)
         self.selection_model.selectionAudioChanged.connect(self.update_selection_times)
@@ -187,29 +188,27 @@ class MediaPlayer(QtMultimedia.QMediaPlayer):  # pragma: no cover
         self.setCurrentTime(self.startTime())
 
     def update_times(self):
-        if (
-            self.playbackState() == QtMultimedia.QMediaPlayer.PlaybackState.StoppedState
-            or self.currentTime() < self.startTime()
-            or self.currentTime() > self.maxTime()
-        ):
+        if self.currentTime() < self.startTime() or self.currentTime() > self.maxTime():
+            self.stop()
+        if self.playbackState() != QtMultimedia.QMediaPlayer.PlaybackState.PlayingState:
             self.setCurrentTime(self.startTime())
 
     def loadNewFile(self, *args):
         self.audioReady.emit(False)
         self.stop()
         try:
-            new_file = self.selection_model.current_file.sound_file.sound_file_path
+            new_file = self.selection_model.model().file.sound_file.sound_file_path
         except Exception:
             self.setSource(QtCore.QUrl())
             return
         if (
             self.selection_model.max_time is None
-            or self.selection_model.current_file is None
-            or self.selection_model.current_file.duration is None
+            or self.selection_model.model().file is None
+            or self.selection_model.model().file.duration is None
         ):
             self.setSource(QtCore.QUrl())
             return
-        self.channels = self.selection_model.current_file.num_channels
+        self.channels = self.selection_model.model().file.num_channels
         self.setSource(f"file:///{new_file}")
         self.setPosition(0)
         self.audioReady.emit(True)
@@ -218,19 +217,6 @@ class MediaPlayer(QtMultimedia.QMediaPlayer):  # pragma: no cover
         pos = self.position()
         return pos / 1000
 
-    def setMaxTime(self, max_time):
-        if max_time is None:
-            return
-        self.max_time = max_time * 1000
-
-    def setMinTime(
-        self, min_time
-    ):  # Positions for MediaPlayer are in milliseconds, no SR required
-        if min_time is None:
-            min_time = 0
-        self.min_time = int(min_time * 1000)
-        self.setCurrentTime(min_time)
-
     def setCurrentTime(self, time):
         if time is None:
             time = 0
@@ -245,7 +231,7 @@ class MediaPlayer(QtMultimedia.QMediaPlayer):  # pragma: no cover
             self.stop()
             self.setSource(
                 QtCore.QUrl.fromLocalFile(
-                    self.selection_model.current_file.sound_file.sound_file_path
+                    self.selection_model.model().file.sound_file.sound_file_path
                 )
             )
             self.play()
@@ -557,6 +543,14 @@ class PaginationWidget(QtWidgets.QToolBar):
 
     def __init__(self, *args):
         super(PaginationWidget, self).__init__(*args)
+        w = QtWidgets.QWidget(self)
+        w.setSizePolicy(
+            QtWidgets.QSizePolicy.Policy.Expanding, QtWidgets.QSizePolicy.Policy.Expanding
+        )
+        w2 = QtWidgets.QWidget(self)
+        w2.setSizePolicy(
+            QtWidgets.QSizePolicy.Policy.Expanding, QtWidgets.QSizePolicy.Policy.Expanding
+        )
         self.current_page = 0
         self.limit = 1
         self.num_pages = 1
@@ -567,10 +561,12 @@ class PaginationWidget(QtWidgets.QToolBar):
         self.previous_page_action = QtGui.QAction(
             icon=QtGui.QIcon(":caret-left.svg"), text="Previous page"
         )
+        self.addWidget(w)
         self.page_label = QtWidgets.QLabel("Page 1 of 1")
         self.addAction(self.previous_page_action)
         self.addWidget(self.page_label)
         self.addAction(self.next_page_action)
+        self.addWidget(w2)
         self.next_page_action.triggered.connect(self.next_page)
         self.previous_page_action.triggered.connect(self.previous_page)
 
@@ -783,6 +779,7 @@ class UtteranceDetailWidget(QtWidgets.QWidget):  # pragma: no cover
         self.settings = AnchorSettings()
         self.setAttribute(QtCore.Qt.WidgetAttribute.WA_StyledBackground, True)
         self.corpus_model = None
+        self.file_model = None
         self.selection_model = None
         self.dictionary_model = None
         self.plot_widget = UtteranceView(self)
@@ -819,24 +816,28 @@ class UtteranceDetailWidget(QtWidgets.QWidget):  # pragma: no cover
     def set_models(
         self,
         corpus_model: CorpusModel,
-        selection_model: CorpusSelectionModel,
+        file_model: FileUtterancesModel,
+        selection_model: FileSelectionModel,
         dictionary_model: DictionaryTableModel,
     ):
         self.corpus_model = corpus_model
+        self.file_model = file_model
         self.selection_model = selection_model
         self.dictionary_model = dictionary_model
         self.corpus_model.textFilterChanged.connect(self.plot_widget.set_search_term)
         self.selection_model.viewChanged.connect(self.update_to_slider)
         self.selection_model.fileChanged.connect(self.update_to_slider)
-        self.plot_widget.set_models(corpus_model, selection_model, self.dictionary_model)
+        self.plot_widget.set_models(
+            corpus_model, file_model, selection_model, self.dictionary_model
+        )
 
     def update_to_slider(self):
         with QtCore.QSignalBlocker(self.scroll_bar):
-            if self.selection_model.current_file is None or self.selection_model.min_time is None:
+            if self.selection_model.model().file is None or self.selection_model.min_time is None:
                 return
             if (
                 self.selection_model.min_time == 0
-                and self.selection_model.max_time == self.selection_model.current_file.duration
+                and self.selection_model.max_time == self.selection_model.model().file.duration
             ):
                 self.scroll_bar.setPageStep(10)
                 self.scroll_bar.setEnabled(False)
@@ -844,7 +845,7 @@ class UtteranceDetailWidget(QtWidgets.QWidget):  # pragma: no cover
                 self.pan_right_button.setEnabled(False)
                 self.scroll_bar.setMaximum(0)
                 return
-            duration_ms = int(self.selection_model.current_file.duration * 1000)
+            duration_ms = int(self.selection_model.model().file.duration * 1000)
             begin = self.selection_model.min_time * 1000
             end = self.selection_model.max_time * 1000
             window_size_ms = int(end - begin)
@@ -2194,7 +2195,7 @@ class DictionaryTableView(AnchorTableView):
             )
             self.dictionary_model.change_word_type(word_id, WordType[current_word_type])
 
-        elif index.column() == 1:
+        elif index.column() == 2:
             word_index = self.dictionary_model.index(index.row(), 0)
             word = self.dictionary_model.data(word_index, QtCore.Qt.ItemDataRole.DisplayRole)
             query = TextFilterQuery(word, False, True, False)
@@ -2245,7 +2246,9 @@ class SpeakerTableView(AnchorTableView):
         speaker_id = self.speaker_model.speakerAt(index.row())
         speaker_name = self.speaker_model.data(index, QtCore.Qt.ItemDataRole.DisplayRole)
         dialog = ConfirmationDialog(
-            f"Break up {speaker_name}", f"Are you sure you want to break up {speaker_name}?", self
+            "Break up speaker confirmation",
+            f"Are you sure you want to break up {speaker_name}?",
+            self,
         )
         if dialog.exec_():
             self.speaker_model.break_up_speaker(speaker_id)
@@ -2680,9 +2683,13 @@ class DiarizationTable(AnchorTableView):
             with self.diarization_model.corpus_model.corpus.session() as session:
                 c = session.query(Corpus).first()
                 try:
-                    file_id, begin, end, channel = (
+                    utterance_id, file_id, begin, end, channel = (
                         session.query(
-                            Utterance.file_id, Utterance.begin, Utterance.end, Utterance.channel
+                            Utterance.id,
+                            Utterance.file_id,
+                            Utterance.begin,
+                            Utterance.end,
+                            Utterance.channel,
                         )
                         .join(Utterance.speaker)
                         .filter(Utterance.speaker_id == speaker_id)
@@ -2694,6 +2701,7 @@ class DiarizationTable(AnchorTableView):
                 except TypeError:
                     self.selection_model.clearSelection()
                     return
+        self.selection_model.set_current_utterance(utterance_id)
         self.selection_model.set_current_file(
             file_id,
             begin,

anchor/workers.py

@@ -19,7 +19,8 @@ import dataclassy
 import librosa
 import numpy as np
 import psycopg2.errors
-import resampy
+import scipy
+import scipy.signal
 import soundfile
 import sqlalchemy
 import tqdm
@@ -28,6 +29,7 @@ from _kalpy.feat import compute_pitch
 from _kalpy.ivector import Plda, ivector_normalize_length
 from _kalpy.matrix import DoubleVector, FloatVector
 from kalpy.feat.pitch import PitchComputer
+from line_profiler_pycharm import profile
 from montreal_forced_aligner import config
 from montreal_forced_aligner.alignment import PretrainedAligner
 from montreal_forced_aligner.config import IVECTOR_DIMENSION, XVECTOR_DIMENSION
@@ -372,7 +374,7 @@ class ClosestSpeakerThread(threading.Thread):
                             )
                             .filter(
                                 c.speaker_ivector_column.cosine_distance(u_ivector)
-                                < min(distance - 0.09, self.threshold),
+                                < min(distance - 0.05, self.threshold),
                                 # self.threshold,
                                 Speaker.id != s_id,
                             )
@@ -2000,25 +2002,24 @@ def file_utterances_function(
     stopped: typing.Optional[threading.Event] = None,
     **kwargs,
 ):
-    with Session() as session:
-        utterances = (
-            session.query(Utterance)
-            .options(
-                selectinload(Utterance.phone_intervals).options(
-                    joinedload(PhoneInterval.phone, innerjoin=True),
-                    joinedload(PhoneInterval.workflow, innerjoin=True),
-                ),
-                selectinload(Utterance.word_intervals).options(
-                    joinedload(WordInterval.word, innerjoin=True),
-                    joinedload(WordInterval.workflow, innerjoin=True),
-                ),
-                joinedload(Utterance.speaker, innerjoin=True),
-            )
-            .filter(Utterance.file_id == file_id)
-            .order_by(Utterance.begin)
-            .all()
+    utterances = (
+        Session.query(Utterance)
+        .options(
+            selectinload(Utterance.phone_intervals).options(
+                joinedload(PhoneInterval.phone, innerjoin=True),
+                joinedload(PhoneInterval.workflow, innerjoin=True),
+            ),
+            selectinload(Utterance.word_intervals).options(
+                joinedload(WordInterval.word, innerjoin=True),
+                joinedload(WordInterval.workflow, innerjoin=True),
+            ),
+            joinedload(Utterance.speaker, innerjoin=True),
         )
-        return utterances, file_id
+        .filter(Utterance.file_id == file_id)
+        .order_by(Utterance.begin)
+        .all()
+    )
+    return utterances, file_id
 
 
 def query_dictionary_function(
@@ -3082,6 +3083,7 @@ class SpeakerTierWorker(FunctionWorker):  # pragma: no cover
         super().__init__("Generating speaker tier", *args)
         self.query_alignment = False
         self.session = None
+        self.file_id = None
 
     def set_params(self, file_id):
         with self.lock:
@@ -3091,21 +3093,31 @@ class SpeakerTierWorker(FunctionWorker):  # pragma: no cover
         if self.session is None:
             return
         self.stopped.clear()
-        with self.lock:
-            utterances = self.session.query(Utterance).options(
-                joinedload(Utterance.speaker, innerjoin=True),
+        with self.lock, self.session() as session:
+            show_phones = (
+                self.settings.value(self.settings.TIER_ALIGNED_PHONES_VISIBLE)
+                or self.settings.value(self.settings.TIER_TRANSCRIBED_PHONES_VISIBLE)
+                or self.settings.value(self.settings.TIER_REFERENCE_PHONES_VISIBLE)
             )
+            show_words = self.settings.value(
+                self.settings.TIER_ALIGNED_WORDS_VISIBLE
+            ) or self.settings.value(self.settings.TIER_TRANSCRIBED_WORDS_VISIBLE)
+            utterances = session.query(Utterance)
             if self.query_alignment:
-                utterances = utterances.options(
-                    selectinload(Utterance.phone_intervals).options(
-                        joinedload(PhoneInterval.phone, innerjoin=True),
-                        joinedload(PhoneInterval.workflow, innerjoin=True),
-                    ),
-                    selectinload(Utterance.word_intervals).options(
-                        joinedload(WordInterval.word, innerjoin=True),
-                        joinedload(WordInterval.workflow, innerjoin=True),
-                    ),
-                )
+                if show_phones:
+                    utterances = utterances.options(
+                        selectinload(Utterance.phone_intervals).options(
+                            joinedload(PhoneInterval.phone, innerjoin=True),
+                            joinedload(PhoneInterval.workflow, innerjoin=True),
+                        )
+                    )
+                if show_words:
+                    utterances = utterances.options(
+                        selectinload(Utterance.word_intervals).options(
+                            joinedload(WordInterval.word, innerjoin=True),
+                            joinedload(WordInterval.workflow, innerjoin=True),
+                        ),
+                    )
             utterances = utterances.filter(Utterance.file_id == self.file_id).order_by(
                 Utterance.begin
             )
@@ -3139,6 +3151,7 @@ class SpectrogramWorker(FunctionWorker):  # pragma: no cover
             self.end = end
             self.channel = channel
 
+    @profile
     def run(self):
         self.stopped.clear()
         dynamic_range = self.settings.value(self.settings.SPEC_DYNAMIC_RANGE)
@@ -3147,12 +3160,18 @@ class SpectrogramWorker(FunctionWorker):  # pragma: no cover
         window_size = self.settings.value(self.settings.SPEC_WINDOW_SIZE)
         pre_emph_coeff = self.settings.value(self.settings.SPEC_PREEMPH)
         max_freq = self.settings.value(self.settings.SPEC_MAX_FREQ)
+        if self.y.shape[0] == 0:
+            return
+        duration = self.y.shape[0] / self.sample_rate
+        if duration > 30:
+            return
         with self.lock:
-            if self.y.shape[0] == 0:
-                return
             max_sr = 2 * max_freq
             if self.sample_rate > max_sr:
-                self.y = resampy.resample(self.y, self.sample_rate, max_sr)
+                self.y = scipy.signal.resample(
+                    self.y, int(self.y.shape[0] * max_sr / self.sample_rate)
+                )
+                # self.y = resampy.resample(self.y, self.sample_rate, max_sr, filter='kaiser_fast')
                 self.sample_rate = max_sr
             self.y = librosa.effects.preemphasis(self.y, coef=pre_emph_coeff)
             if self.stopped.is_set():
@@ -3226,6 +3245,10 @@ class PitchWorker(FunctionWorker):  # pragma: no cover
                 max_f0=self.max_f0,
                 penalty_factor=self.penalty_factor,
                 delta_pitch=self.delta_pitch,
+                add_pov_feature=True,
+                add_normalized_log_pitch=False,
+                add_delta_pitch=False,
+                add_raw_log_pitch=True,
             )
 
     def run(self):
@@ -3233,6 +3256,8 @@ class PitchWorker(FunctionWorker):  # pragma: no cover
         with self.lock:
             if self.y.shape[0] == 0:
                 return
+            if self.end - self.begin < 0.1:
+                return
             pitch_track = compute_pitch(
                 self.y, self.pitch_computer.extraction_opts, self.pitch_computer.process_opts
             ).numpy()

Anchor_annotator-0.0.11.dist-info/RECORD

@@ -1,21 +0,0 @@
-anchor/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-anchor/__main__.py,sha256=5ufG8lcx2x1am-04xI991AG7saJd24dxPw5JzjmB878,45
-anchor/_version.py,sha256=Ht_kK4WoZp7blBjGQJ5CR8Vli99GsMsH_I96HrAX0WE,413
-anchor/command_line.py,sha256=xvuCWaPWNVZTg5Ic28qbOYsOLaFbodhBsoZHKJSBazs,482
-anchor/db.py,sha256=ef4lO6HtCKoxC9CorIc0ZbPxKpjHa576a0ZIBOWNU9E,4956
-anchor/main.py,sha256=-_EF3q7LJIEXBgkM1UnxszkKrelQsKfeSZoC6WAv9KM,113303
-anchor/models.py,sha256=ub94jARUDBjy2SKzqX_WrktldjvJ7I8K5NurgoLcCyk,74619
-anchor/plot.py,sha256=FizOExN7UBCc40p7C9RZZ1PJBMXvXJ1oWfQXxZ-I0EI,105575
-anchor/resources_rc.py,sha256=sQ6GvAK3NTVR5bvgR9jCWDeepSYOhEph2mg1ECxnMOs,3560262
-anchor/settings.py,sha256=SMqliO60BfR1Z-J5G1BwIfdq1v3wFIw7l2C-ad21lZ8,46359
-anchor/ui_error_dialog.py,sha256=c_QS0s1VaJEV9AhcrQZQyWHHpUPudWjJY1NI7Ytipio,3832
-anchor/ui_main_window.py,sha256=aEABdKi1Eb1c2MKUsbCbufEp1lKSLFNsC9TPX244UPI,64618
-anchor/ui_preferences.py,sha256=uQFPF_4cCRWwHhUQqMOdonIVyy_BjdSqya7YRhhGz3w,38948
-anchor/undo.py,sha256=rVus-7HC9wPIiab3dUxIeNGK7jWOMSVmDvCFEwU-408,33163
-anchor/widgets.py,sha256=Voj92E4y-SazmkIWEOO63tKaFdT7A4LWpcLPf049Fg0,135087
-anchor/workers.py,sha256=VmSqnW8jaaKY0SphualiUZW-a3TCs2gR82sKyQvK3A8,169802
-Anchor_annotator-0.0.11.dist-info/LICENSE,sha256=C0oIsblENEgWQ7XMNdYoXyXsIA5wa3YF0I9lK3H7A1s,1076
-Anchor_annotator-0.0.11.dist-info/METADATA,sha256=H2TgJ9GtQBdWTD32ygJJiovdOmKGw7fxxR8USQt2ca4,1501
-Anchor_annotator-0.0.11.dist-info/WHEEL,sha256=Xo9-1PvkuimrydujYJAjF7pCkriuXBpUPEjma1nZyJ0,92
-Anchor_annotator-0.0.11.dist-info/top_level.txt,sha256=wX6ZKxImGRZKFQjs3f6XYw_TfbAp6Xs3SmbLfLbFAJ0,7
-Anchor_annotator-0.0.11.dist-info/RECORD,,