app.py

# pulmoprobe_backend/app.py

from fastapi import FastAPI, HTTPException
from fastapi.middleware.cors import CORSMiddleware
from pydantic import BaseModel
import joblib
import pandas as pd
from huggingface_hub import hf_hub_download
import os
import logging

# ------------------------------------------------------------
# Setup Logging
# ------------------------------------------------------------
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)

# ------------------------------------------------------------
# FastAPI Initialization
# ------------------------------------------------------------
app = FastAPI(title="PulmoProbe AI API")

# Allow CORS for frontend communication
app.add_middleware(
    CORSMiddleware,
    allow_origins=["*"],
    allow_credentials=True,
    allow_methods=["*"],
    allow_headers=["*"],
)

# ------------------------------------------------------------
# Hugging Face Model Setup
# ------------------------------------------------------------
os.environ['HF_HOME'] = '/tmp/huggingface'
os.makedirs(os.environ['HF_HOME'], exist_ok=True)

MODEL_REPO_ID = "costaspinto/PulmoProbe"
MODEL_FILENAME = "best_model.joblib"

try:
    model_path = hf_hub_download(
        repo_id=MODEL_REPO_ID,
        filename=MODEL_FILENAME,
        cache_dir=os.environ['HF_HOME']
    )
    model = joblib.load(model_path)
    logger.info("✅ Model loaded successfully")
except Exception as e:
    logger.error(f"❌ Failed to load model: {str(e)}")
    raise RuntimeError(f"Model loading failed: {str(e)}")

# ------------------------------------------------------------
# Define Input Schema (One-Hot Encoded)
# ------------------------------------------------------------
class OneHotPatientData(BaseModel):
    age: float
    bmi: float
    cholesterol_level: float
    hypertension: int
    asthma: int
    cirrhosis: int
    other_cancer: int
    gender_Male: int
    family_history_Yes: int

    country_Belgium: int
    country_Bulgaria: int
    country_Croatia: int
    country_Cyprus: int
    country_Czech_Republic: int
    country_Denmark: int
    country_Estonia: int
    country_Finland: int
    country_France: int
    country_Germany: int
    country_Greece: int
    country_Hungary: int
    country_Ireland: int
    country_Italy: int
    country_Latvia: int
    country_Lithuania: int
    country_Luxembourg: int
    country_Malta: int
    country_Netherlands: int
    country_Poland: int
    country_Portugal: int
    country_Romania: int
    country_Slovakia: int
    country_Slovenia: int
    country_Spain: int
    country_Sweden: int

    cancer_stage_Stage_II: int
    cancer_stage_Stage_III: int
    cancer_stage_Stage_IV: int

    smoking_status_Former_Smoker: int
    smoking_status_Never_Smoked: int
    smoking_status_Passive_Smoker: int

    treatment_type_Combined: int
    treatment_type_Radiation: int
    treatment_type_Surgery: int

# ------------------------------------------------------------
# Root Endpoint
# ------------------------------------------------------------
@app.get("/")
def read_root():
    return {"message": "Welcome to PulmoProbe AI API"}

# ------------------------------------------------------------
# Prediction Endpoint
# ------------------------------------------------------------
@app.post("/predict")
def predict(data: OneHotPatientData):
    try:
        input_dict = data.dict()
        logger.info(f"Incoming data: {input_dict}")

        feature_order = [
            'age', 'bmi', 'cholesterol_level', 'hypertension', 'asthma',
            'cirrhosis', 'other_cancer', 'gender_Male',
            'country_Belgium', 'country_Bulgaria', 'country_Croatia', 'country_Cyprus',
            'country_Czech Republic', 'country_Denmark', 'country_Estonia', 'country_Finland',
            'country_France', 'country_Germany', 'country_Greece', 'country_Hungary',
            'country_Ireland', 'country_Italy', 'country_Latvia', 'country_Lithuania',
            'country_Luxembourg', 'country_Malta', 'country_Netherlands', 'country_Poland',
            'country_Portugal', 'country_Romania', 'country_Slovakia', 'country_Slovenia',
            'country_Spain', 'country_Sweden',
            'cancer_stage_Stage Ii', 'cancer_stage_Stage Iii', 'cancer_stage_Stage Iv',
            'family_history_Yes',
            'smoking_status_Former Smoker', 'smoking_status_Never Smoked', 'smoking_status_Passive Smoker',
            'treatment_type_Combined', 'treatment_type_Radiation', 'treatment_type_Surgery'
        ]

        input_dict_complete = {col: input_dict.get(col, 0) for col in feature_order}
        input_df = pd.DataFrame([input_dict_complete], columns=feature_order)

        probabilities = model.predict_proba(input_df)[0]
        confidence_high_risk = probabilities[1]
        risk_level = "High Risk of Non-Survival" if confidence_high_risk > 0.5 else "Low Risk of Non-Survival"

        result = {
            "risk": risk_level,
            "confidence": f"{confidence_high_risk*100:.1f}%"
        }
        return result

    except Exception as e:
        logger.error(f"Prediction error: {str(e)}")
        raise HTTPException(status_code=500, detail=str(e))