from fastapi import FastAPI, HTTPException from fastapi.middleware.cors import CORSMiddleware from pydantic import BaseModel import os import sys BASE_DIR = os.path.dirname(os.path.abspath(__file__)) sys.path.append(BASE_DIR) from model import SentiTaxEngine, SentiTaxRLM app = FastAPI( title="SentiTax RLM API", description="SML and RLM service for taxation analysis and compliance checking.", version="2.0.0" ) app.add_middleware( CORSMiddleware, allow_origins=["*"], allow_methods=["*"], allow_headers=["*"], ) WEIGHTS_PATH = os.path.join(BASE_DIR, "weights", "tax_model.pt") engine = SentiTaxEngine(WEIGHTS_PATH) rlm_engine = SentiTaxRLM() class RequestBody(BaseModel): text: str tier: str = "A" @app.get("/") async def root(): return {"status": "active", "service": "sentitax", "port": 9206} @app.get("/health") async def health(): rlm_health = await rlm_engine.engine.health_check() return { "status": "ok", "rlm_health": rlm_health } @app.post("/api/v1/tax/compliance") async def predict_endpoint(body: RequestBody, deep: bool = False): try: if deep or body.tier in ("C", "D"): return await rlm_engine.predict_deep(body.text, body.tier) return engine.predict(body.text) except Exception as e: raise HTTPException(status_code=500, detail=str(e)) @app.post("/api/v1/tax/score") async def score_endpoint(body: RequestBody, deep: bool = False): # Support legacy /score path mapped in sml_client.py try: if deep or body.tier in ("C", "D"): return await rlm_engine.predict_deep(body.text, body.tier) return engine.predict(body.text) except Exception as e: raise HTTPException(status_code=500, detail=str(e)) @app.post("/api/v1/tax/reason") async def tax_reason(body: RequestBody): try: return await rlm_engine.predict_deep(body.text, body.tier) except Exception as e: raise HTTPException(status_code=500, detail=str(e)) @app.on_event("shutdown") async def shutdown_event(): from senti.core.engines.superpacks.rlm_engine import RLMEngine await RLMEngine.shutdown() @app.post("/api/v1/tax/predict") async def predict_endpoint_legacy_alias(body: RequestBody): try: # Try different possible engines if 'engine' in globals(): return engine.predict(body.text) elif 'rlm_engine' in globals(): return await rlm_engine.predict_deep(body.text, "A") else: return {"status": "ok", "service": "sentitax"} except Exception as e: raise HTTPException(status_code=500, detail=str(e)) if __name__ == "__main__": import uvicorn uvicorn.run(app, host="0.0.0.0", port=9206)