main.py

from fastapi import FastAPI, HTTPException
from pydantic import BaseModel
from transformers import AutoTokenizer, AutoModelForSequenceClassification, TextClassificationPipeline
import torch

app = FastAPI(title="FinBERT API")

# --------- Tải Model Global ---------
model_name = "ProsusAI/finbert"
print(f"Loading model {model_name}...")

try:
    tokenizer = AutoTokenizer.from_pretrained(model_name, use_fast=True)
    model = AutoModelForSequenceClassification.from_pretrained(model_name, num_labels=3)
    cryptobert = TextClassificationPipeline(
        model=model, 
        tokenizer=tokenizer, 
        max_length=64, 
        truncation=True, 
        padding='max_length'
    )
    print("Model loaded successfully!")
except Exception as e:
    print(f"Error loading model: {e}")
    cryptobert = None

# --------- Định nghĩa Schema ---------
class AnalyzeRequest(BaseModel):
    texts: list[str]

class AnalyzeResult(BaseModel):
    text: str
    label: str
    score: float
    numeric_score: float

class AnalyzeResponse(BaseModel):
    results: list[AnalyzeResult]
    avg_score: float

# --------- Helper Function ---------
def calculate_numeric_score(label: str, score: float) -> float:
    if label == 'positive':
        return score
    elif label == 'negative':
        return -score
    else: # Neutral
        return 0.0

# --------- API Endpoints ---------
@app.get("/")
def read_root():
    return {"status": "ok", "message": "FinBERT API is running", "model_loaded": cryptobert is not None}

@app.post("/api/sentiment", response_model=AnalyzeResponse)
def analyze_sentiment(req: AnalyzeRequest):
    if not cryptobert:
        raise HTTPException(status_code=500, detail="Model is not loaded properly.")
    
    if not req.texts:
        return {"results": [], "avg_score": 0.0}

    try:
        # Run predictions in batch
        preds = cryptobert(req.texts)
        
        results = []
        total_numeric = 0.0
        
        for text, pred in zip(req.texts, preds):
            label = pred['label']
            score = float(pred['score'])
            numeric_score = calculate_numeric_score(label, score)
            
            results.append({
                "text": text,
                "label": label,
                "score": score,
                "numeric_score": numeric_score
            })
            total_numeric += numeric_score
            
        avg_score = total_numeric / len(results) if len(results) > 0 else 0.0
        
        return {
            "results": results,
            "avg_score": avg_score
        }
    except Exception as e:
        raise HTTPException(status_code=500, detail=str(e))