From ea996b19571705fbd3db42852980e8b57a834eea Mon Sep 17 00:00:00 2001
From: ArrangingFear56 <82667894+ArrangingFear56@users.noreply.github.com>
Date: Mon, 9 Dec 2024 17:11:52 +0800
Subject: [PATCH] Delete modules/whisper_module.py

---
 modules/whisper_module.py | 56 ---------------------------------------
 1 file changed, 56 deletions(-)
 delete mode 100644 modules/whisper_module.py

diff --git a/modules/whisper_module.py b/modules/whisper_module.py
deleted file mode 100644
index 551ac89..0000000
--- a/modules/whisper_module.py
+++ /dev/null
@@ -1,56 +0,0 @@
-"""
-Speech-to-text module based on Whisper for SillyTavern Extras
-    - Whisper github: https://github.com/openai/whisper
-
-Authors:
-    - Tony Ribeiro (https://github.com/Tony-sama)
-
-Models are saved into user cache folder, example: C:/Users/toto/.cache/whisper
-
-References:
-    - Code adapted from:
-        - whisper github: https://github.com/openai/whisper
-        - oobabooga text-generation-webui github: https://github.com/oobabooga/text-generation-webui
-"""
-from flask import jsonify, abort, request
-
-from faster_whisper import WhisperModel
-
-DEBUG_PREFIX = "<stt whisper module>"
-RECORDING_FILE_PATH = "stt_test.wav"
-
-model_size = "large-v3-turbo"
-
-model = WhisperModel(model_size, device="cuda", compute_type="float16")
-
-def load_model(file_path=None):
-    """
-    Load given vosk model from file or default to en-us model.
-    Download model to user cache folder, example: C:/Users/toto/.cache/vosk
-    """
-    return WhisperModel(model_size, device="cuda", compute_type="float16")
-
-def process_audio():
-    """
-    Transcript request audio file to text using Whisper
-    """
-
-    if model is None:
-        print(DEBUG_PREFIX,"Whisper model not initialized yet.")
-        return WhisperModel(model_size, device="cuda", compute_type="float16")
-
-    try:
-        file = request.files.get('AudioFile')
-        language = request.form.get('language', default=None)
-        file.save(RECORDING_FILE_PATH)
-        segments, info = model.transcribe(RECORDING_FILE_PATH, beam_size=5)
-        transcript=""
-        for segment in segments:
-            transcript=transcript+" "+segment.text
-        print(DEBUG_PREFIX, "Transcripted from audio file (whisper):", transcript)
-
-        return jsonify({"transcript": transcript})
-
-    except Exception as e: # No exception observed during test but we never know
-        print(e)
-        abort(500, DEBUG_PREFIX+" Exception occurs while processing audio")