from fastapi import FastAPI, Query from pydantic import BaseModel from typing import Optional import pandas as pd from EntitySentiment import entity_sentiment_single, load_pipeline import spacy try: nlp = spacy.load("en_core_web_sm") except OSError: import os os.system("python -m spacy download en_core_web_sm") nlp = spacy.load("en_core_web_sm") # Initialize FastAPI app = FastAPI( title="Entity-Level Financial Sentiment API", description="Runs entity-based sentiment analysis on financial texts using FinBERT + spaCy", version="1.0.0" ) # Load pipeline ONCE at startup (so models don’t reload every request) nlp, tokenizer, model, device, id2label = load_pipeline() class SentimentRequest(BaseModel): text: str entity_filter: Optional[str] = None @app.post("/analyze") def analyze_sentiment(req: SentimentRequest): df = entity_sentiment_single( text=req.text, entity_filter=req.entity_filter, finbert_model_name="yiyanghkust/finbert-tone", spacy_model_name="en_core_web_trf" ) return df.to_dict(orient="records") @app.get("/") def root(): return {"message": "Entity-Level Financial Sentiment API is running!"}