Update app.py
Browse files
app.py
CHANGED
@@ -55,13 +55,20 @@ def convert_mp4_to_mp3(mp4_path, mp3_path):
|
|
55 |
# Transcribe audio
|
56 |
def transcribe_audio(audio_path):
|
57 |
try:
|
|
|
|
|
|
|
|
|
58 |
inputs = ffmpeg_read(audio_path, whisper_pipeline.feature_extractor.sampling_rate)
|
59 |
inputs = {"array": inputs, "sampling_rate": whisper_pipeline.feature_extractor.sampling_rate}
|
|
|
|
|
60 |
result = whisper_pipeline(inputs, batch_size=BATCH_SIZE, return_timestamps=False)
|
61 |
return result["text"]
|
62 |
except Exception as e:
|
63 |
return f"Error during transcription: {e}"
|
64 |
|
|
|
65 |
# Classify the sentence to the correct SOAP section
|
66 |
def classify_sentence(sentence):
|
67 |
similarities = {section: util.pytorch_cos_sim(embedder.encode(sentence), soap_embeddings[section]) for section in soap_prompts.keys()}
|
|
|
55 |
# Transcribe audio
|
56 |
def transcribe_audio(audio_path):
|
57 |
try:
|
58 |
+
if not os.path.exists(audio_path):
|
59 |
+
raise FileNotFoundError(f"Audio file not found: {audio_path}")
|
60 |
+
|
61 |
+
# Read the audio file and prepare inputs for Whisper
|
62 |
inputs = ffmpeg_read(audio_path, whisper_pipeline.feature_extractor.sampling_rate)
|
63 |
inputs = {"array": inputs, "sampling_rate": whisper_pipeline.feature_extractor.sampling_rate}
|
64 |
+
|
65 |
+
# Perform transcription using Whisper
|
66 |
result = whisper_pipeline(inputs, batch_size=BATCH_SIZE, return_timestamps=False)
|
67 |
return result["text"]
|
68 |
except Exception as e:
|
69 |
return f"Error during transcription: {e}"
|
70 |
|
71 |
+
|
72 |
# Classify the sentence to the correct SOAP section
|
73 |
def classify_sentence(sentence):
|
74 |
similarities = {section: util.pytorch_cos_sim(embedder.encode(sentence), soap_embeddings[section]) for section in soap_prompts.keys()}
|