Add ffmpeg-python as dep

llama_assistant/speech_recognition_thread.py

@@ -1,7 +1,7 @@
-from pathlib import Path
 import time
 import os
 import re
+import traceback
 
 from PyQt5.QtCore import QThread, pyqtSignal
 import pyaudio
@@ -71,7 +71,6 @@ def run(self):
             # Transcribe audio
             res = self.whisper.transcribe(str(tmp_filepath))
             transcription = self.whisper.extract_text(res)
-            os.remove(tmp_filepath)
 
             if isinstance(transcription, list):
                 # Remove all "[BLANK_AUDIO]" from the transcription
@@ -82,7 +81,10 @@ def run(self):
             if transcription.strip():  # Only emit if there's non-empty transcription
                 self.finished.emit(transcription)
 
+            os.remove(tmp_filepath)
+
         except Exception as e:
+            traceback.print_exc()
             self.error.emit(f"An error occurred: {str(e)}")
         finally:
             stream.stop_stream()

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "llama-assistant"
-version = "0.1.33"
+version = "0.1.35"
 authors = [
     {name = "Viet-Anh Nguyen", email = "[email protected]"},
 ]
@@ -25,6 +25,7 @@ classifiers = [
     "Programming Language :: Python :: 3.12",
 ]
 dependencies = [
+    "ffmpeg-python",
     "PyQt5",
     "markdown",
     "llama-cpp-python",

requirements.txt

@@ -6,4 +6,5 @@ llama-cpp-python==0.3.1
 huggingface_hub==0.25.1
 openwakeword==0.6.0
 pyinstaller==6.10.0
+ffmpeg-python==0.2.0
 git+https://github.com/stlukey/whispercpp.py