From f69a3140913bfb3c08b9be7f8631cf4971e571bb Mon Sep 17 00:00:00 2001
From: Lameus <1vany1@mail.ru>
Date: Mon, 11 Dec 2023 14:03:41 +0300
Subject: [PATCH] add flag for results of congr, fix token for diar

---
 expert/core/congruence/congruence_analysis.py | 27 ++++++++++++-------
 expert/data/annotation/speech_to_text.py      |  2 +-
 .../data/diarization/speaker_diarization.py   |  2 +-
 3 files changed, 19 insertions(+), 12 deletions(-)

diff --git a/expert/core/congruence/congruence_analysis.py b/expert/core/congruence/congruence_analysis.py
index d985ad8..b3c84c2 100644
--- a/expert/core/congruence/congruence_analysis.py
+++ b/expert/core/congruence/congruence_analysis.py
@@ -65,6 +65,7 @@ def __init__(
         sr: int = 44100,
         device: torch.device | None = None,
         output_dir: str | PathLike | None = None,
+        return_path: bool = False,
     ):
         if lang not in ["en", "ru"]:
             raise NotImplementedError("'lang' must be 'en' or 'ru'.")
@@ -92,6 +93,8 @@ def __init__(
         if not os.path.exists(self.temp_path):
             os.makedirs(self.temp_path)
 
+        self.return_path = return_path
+
     @property
     def device(self) -> torch.device:
         """Check the device type.
@@ -194,15 +197,19 @@ def get_congruence(self):
         emotions_data["audio"] = audio_data.to_dict(orient="records")
         emotions_data["text"] = text_data.to_dict(orient="records")
 
-        with open(
-            os.path.join(self.temp_path, "emotions.json"), "w"
-        ) as filename:
-            json.dump(emotions_data, filename)
+        if self.return_path:
+            with open(
+                os.path.join(self.temp_path, "emotions.json"), "w"
+            ) as filename:
+                json.dump(emotions_data, filename)
 
-        cong_data[["video_path", "time_sec", "congruence"]].to_json(
-            os.path.join(self.temp_path, "congruence.json"), orient="records"
-        )
+            cong_data[["video_path", "time_sec", "congruence"]].to_json(
+                os.path.join(self.temp_path, "congruence.json"),
+                orient="records",
+            )
 
-        return os.path.join(self.temp_path, "emotions.json"), os.path.join(
-            self.temp_path, "congruence.json"
-        )
+            return os.path.join(self.temp_path, "emotions.json"), os.path.join(
+                self.temp_path, "congruence.json"
+            )
+        else:
+            return {"emotions": emotions_data, "congruence": cong_data}
diff --git a/expert/data/annotation/speech_to_text.py b/expert/data/annotation/speech_to_text.py
index e139cba..86a63ae 100644
--- a/expert/data/annotation/speech_to_text.py
+++ b/expert/data/annotation/speech_to_text.py
@@ -13,7 +13,7 @@
 def transcribe_video(
     video_path: Union[str, PathLike],
     lang: Optional[str] = "en",
-    model: Optional[str] = "server",
+    model: Optional[str] = "local",
     device: Optional[Union[torch.device, None]] = None,
 ) -> Dict:
     """Speech recognition module from video.
diff --git a/expert/data/diarization/speaker_diarization.py b/expert/data/diarization/speaker_diarization.py
index 8867b74..cbebab6 100644
--- a/expert/data/diarization/speaker_diarization.py
+++ b/expert/data/diarization/speaker_diarization.py
@@ -56,7 +56,7 @@ def __init__(
         if device is not None:
             self._device = device
 
-        token = "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"  # FIXME убрать харкод пароля # nosec
+        token = "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"
         self.pipeline = Pipeline.from_pretrained(
             "pyannote/speaker-diarization@2.1", use_auth_token=token
         )