|
|
import logging |
|
|
import sys |
|
|
from fastapi import FastAPI, UploadFile, File, HTTPException |
|
|
from pydantic import BaseModel |
|
|
import config |
|
|
from models import load_whisper, load_summarizer, load_spacy |
|
|
from services import process_transcription, process_summary |
|
|
|
|
|
|
|
|
logger = logging.getLogger(__name__) |
|
|
|
|
|
app = FastAPI( |
|
|
title="Transcription and Summarization API", |
|
|
description="API using Faster-Whisper, spaCy, and Hugging Face Transformers", |
|
|
version="1.0.0" |
|
|
) |
|
|
|
|
|
logger.info("Application starting up - loading models...") |
|
|
whisper_model = load_whisper(config) |
|
|
summarizer_pipeline = load_summarizer(config) |
|
|
nlp_spacy = load_spacy(config) |
|
|
logger.info("Model loading complete.") |
|
|
|
|
|
if not whisper_model: |
|
|
logger.critical("Whisper model failed to load. Transcription endpoint will be unavailable.") |
|
|
if not summarizer_pipeline: |
|
|
logger.critical("Summarizer pipeline failed to load. Summarization endpoint will be unavailable.") |
|
|
if not nlp_spacy: |
|
|
logger.warning("SpaCy model failed to load. Summarization will proceed without spaCy preprocessing.") |
|
|
|
|
|
|
|
|
class TranscriptInput(BaseModel): |
|
|
transcript: str |
|
|
|
|
|
|
|
|
@app.get("/health") |
|
|
def health(): |
|
|
return {"status": "ok", |
|
|
"whisper_loaded": whisper_model is not None, |
|
|
"summarizer_loaded": summarizer_pipeline is not None, |
|
|
"spacy_loaded": nlp_spacy is not None |
|
|
} |
|
|
|
|
|
|
|
|
@app.post("/transcribe") |
|
|
async def transcription(audio_file : UploadFile = File(...)): |
|
|
if whisper_model is None: |
|
|
raise HTTPException(status_code=503, detail="Transcription service unavailable.") |
|
|
|
|
|
try: |
|
|
content = await audio_file.read() |
|
|
transcript, info = process_transcription(content, whisper_model) |
|
|
logger.info(f"Transcription successful. Language: {info.language}") |
|
|
return {"transcript": transcript} |
|
|
except ValueError as ve: |
|
|
logger.error(f"Value error during transcription processing: {ve}") |
|
|
raise HTTPException(status_code=400, detail=str(ve)) |
|
|
except Exception as e: |
|
|
logger.error(f"Unhandled error during transcription: {e}", exc_info=True) |
|
|
raise HTTPException(status_code=500, detail="Internal server error during transcription.") |
|
|
|
|
|
|
|
|
@app.post("/summarize") |
|
|
def summarize(input: TranscriptInput): |
|
|
|
|
|
if summarizer_pipeline is None: |
|
|
raise HTTPException(status_code=503, detail="Summarization service unavailable.") |
|
|
if not input.transcript: |
|
|
raise HTTPException(status_code=400, detail="Transcript cannot be empty.") |
|
|
|
|
|
try: |
|
|
summary = process_summary(input.transcript, summarizer_pipeline, nlp_spacy, config) |
|
|
return {"summary": summary} |
|
|
|
|
|
except ValueError as ve: |
|
|
logger.error(f"Value error during summary processing: {ve}") |
|
|
raise HTTPException(status_code=400, detail=str(ve)) |
|
|
|
|
|
except Exception as e: |
|
|
logger.error(f"Unhandled error during summarization: {e}", exc_info=True) |
|
|
raise HTTPException(status_code=500, detail="Internal server error during summarization.") |
