Spaces:
Running
Running
File size: 15,086 Bytes
8b9f7d9 643b2e1 8b9f7d9 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 |
"""
Learn Pathophysiology - FastAPI Backend
Serves RAG + LLM API endpoints and Vue frontend static files.
Deploy: HuggingFace Spaces (Docker) or run locally.
"""
import os
import secrets
import logging
from pathlib import Path
from datetime import datetime, timezone, timedelta
import jwt
from fastapi import FastAPI, UploadFile, File, Form, HTTPException, Depends, Request
from fastapi.middleware.cors import CORSMiddleware
from fastapi.staticfiles import StaticFiles
from fastapi.responses import FileResponse
from pydantic import BaseModel
from dotenv import load_dotenv
from google import genai
from google.genai import types
from google.oauth2 import id_token as google_id_token
from google.auth.transport import requests as google_requests
import chromadb
load_dotenv()
logger = logging.getLogger(__name__)
# =============================================================================
# CONFIGURATION
# =============================================================================
GEMINI_API_KEY = os.environ.get("GEMINI_API_KEY", "")
CHROMA_DIR = os.environ.get("CHROMA_DIR", "../chroma_db")
COLLECTION_NAME = "pathophysiology"
EMBEDDING_MODEL = "gemini-embedding-001"
DEFAULT_MODEL = "gemini-3-flash-preview"
RAG_TOP_K = 5
# Auth
GOOGLE_CLIENT_ID = os.environ.get("GOOGLE_CLIENT_ID", "")
JWT_SECRET = os.environ.get("JWT_SECRET", secrets.token_urlsafe(32))
JWT_ALGORITHM = "HS256"
JWT_EXPIRY_DAYS = 7
AUTH_ENABLED = bool(GOOGLE_CLIENT_ID) # disable auth if no client ID
AVAILABLE_MODELS = {
"gemini-3-flash-preview": {
"name": "Gemini 3 Flash",
"description": "Najnoviji i najbrzi model",
"icon": "swords",
"wc3_name": "Blademaster",
"tier": "fast",
},
"gemini-2.5-flash": {
"name": "Gemini 2.5 Flash",
"description": "Brz i pouzdan",
"icon": "bow",
"wc3_name": "Shadow Hunter",
"tier": "fast",
},
"gemini-2.5-pro": {
"name": "Gemini 2.5 Pro",
"description": "Najpametniji za kompleksne zadatke",
"icon": "mage",
"wc3_name": "Archmage",
"tier": "smart",
},
}
SYSTEM_PROMPT = """Ti si "Learn Pathophysiology AI", strucni asistent za ucenje patofiziologije
za studente medicine.
ULOGA:
- Objasnjavaš patofiziološke koncepte jasno i precizno
- Koristiš primjere i analogije kad je moguce
- Povezuješ koncepte s klinickom praksom
- Odgovaraš na hrvatskom jeziku
KONTEKST IZ BAZE ZNANJA:
{rag_context}
PRAVILA:
1. Uvijek citiraj izvor kad koristiš informacije iz konteksta
2. Ako nisi siguran, reci to otvoreno
3. Koristi medicinsku terminologiju, ali objasni kompleksne termine
4. Budi koncizan ali potpun u odgovorima
5. Odgovaraj na hrvatskom jeziku"""
# =============================================================================
# SINGLETONS
# =============================================================================
_genai_client = None
_chroma_collection = None
def get_client():
global _genai_client
if _genai_client is None:
if not GEMINI_API_KEY:
raise HTTPException(status_code=500, detail="GEMINI_API_KEY not configured")
_genai_client = genai.Client(api_key=GEMINI_API_KEY)
return _genai_client
def get_collection():
global _chroma_collection
if _chroma_collection is None:
chroma_path = Path(CHROMA_DIR)
if not chroma_path.exists():
# Try relative to this file
alt_path = Path(__file__).parent.parent.parent / "chroma_db"
if alt_path.exists():
chroma_path = alt_path
else:
return None
try:
client = chromadb.PersistentClient(path=str(chroma_path))
_chroma_collection = client.get_collection(COLLECTION_NAME)
except Exception as e:
logger.error(f"ChromaDB error: {e}")
return None
return _chroma_collection
# =============================================================================
# AUTH HELPERS
# =============================================================================
def create_jwt(email: str, name: str, picture: str = "") -> str:
payload = {
"sub": email,
"name": name,
"picture": picture,
"iat": datetime.now(timezone.utc),
"exp": datetime.now(timezone.utc) + timedelta(days=JWT_EXPIRY_DAYS),
}
return jwt.encode(payload, JWT_SECRET, algorithm=JWT_ALGORITHM)
def decode_jwt(token: str) -> dict | None:
try:
return jwt.decode(token, JWT_SECRET, algorithms=[JWT_ALGORITHM])
except jwt.ExpiredSignatureError:
return None
except jwt.InvalidTokenError:
return None
async def require_auth(request: Request):
"""FastAPI dependency — returns user dict or raises 401."""
if not AUTH_ENABLED:
return {"sub": "anonymous", "name": "Local User"}
auth_header = request.headers.get("Authorization", "")
if not auth_header.startswith("Bearer "):
raise HTTPException(status_code=401, detail="Not authenticated")
token = auth_header[7:]
user = decode_jwt(token)
if not user:
raise HTTPException(status_code=401, detail="Invalid or expired token")
return user
# =============================================================================
# RAG FUNCTIONS
# =============================================================================
def embed_query(text: str) -> list[float]:
c = get_client()
result = c.models.embed_content(model=EMBEDDING_MODEL, contents=text)
return result.embeddings[0].values
def query_rag(query_text: str, top_k: int = RAG_TOP_K):
coll = get_collection()
if coll is None:
return "Nema dostupnog konteksta.", []
try:
query_embedding = embed_query(query_text)
results = coll.query(
query_embeddings=[query_embedding],
n_results=top_k,
include=["documents", "metadatas", "distances"]
)
contexts = []
citations = []
if results and results["documents"] and results["documents"][0]:
for idx, (doc, meta, dist) in enumerate(zip(
results["documents"][0],
results["metadatas"][0],
results["distances"][0]
)):
contexts.append(doc)
similarity = max(0, 1 - dist / 2)
citations.append({
"text": doc[:600] + "..." if len(doc) > 600 else doc,
"score": round(similarity, 3),
"source": meta.get("source", "Baza znanja"),
"page_num": meta.get("page_num", "?"),
"rank": idx + 1,
})
formatted = "\n\n---\n\n".join(contexts) if contexts else "Nema konteksta."
return formatted, citations
except Exception as e:
logger.error(f"RAG error: {e}")
return "Nema dostupnog konteksta.", []
def generate_chat_response(message: str, history: list, model_name: str = ""):
model_name = model_name or DEFAULT_MODEL
if model_name not in AVAILABLE_MODELS:
model_name = DEFAULT_MODEL
c = get_client()
rag_context, citations = query_rag(message)
system_prompt = SYSTEM_PROMPT.format(rag_context=rag_context)
contents = [system_prompt]
for msg in (history or [])[-10:]:
role = msg.get("role", "user")
content = msg.get("content", "")
if role == "user":
contents.append(f"Student: {content}")
else:
contents.append(f"Asistent: {content}")
contents.append(f"Student: {message}")
response = c.models.generate_content(
model=model_name,
contents="\n\n".join(contents),
config=types.GenerateContentConfig(
temperature=0.7,
max_output_tokens=8192,
top_p=0.9,
)
)
return response.text, citations
def do_analyze_image(image_bytes: bytes, question: str = "", model_name: str = ""):
model_name = model_name or DEFAULT_MODEL
if model_name not in AVAILABLE_MODELS:
model_name = DEFAULT_MODEL
c = get_client()
# Extract keywords from image
extract_resp = c.models.generate_content(
model=model_name,
contents=[
types.Part.from_bytes(data=image_bytes, mime_type="image/jpeg"),
"Izvuci glavni topic i kljucne rijeci s ove stranice. Odgovori kratko."
],
config=types.GenerateContentConfig(temperature=0.3, max_output_tokens=200)
)
rag_context, citations = query_rag(extract_resp.text, top_k=3)
if question:
prompt = (
f"Analiziraj ovu stranicu iz materijala za patofiziologiju "
f"i odgovori na pitanje studenta.\n\n"
f"PITANJE: {question}\n\n"
f"KONTEKST IZ BAZE ZNANJA:\n{rag_context}\n\n"
f"Odgovori detaljno na hrvatskom jeziku."
)
else:
prompt = (
f"Analiziraj ovu stranicu iz materijala za patofiziologiju.\n\n"
f"1. Prepoznaj glavni topic\n2. Izvuci kljucne pojmove\n"
f"3. Sazmi glavne tocke\n4. Objasni klinicku vaznost\n\n"
f"KONTEKST IZ BAZE ZNANJA:\n{rag_context}\n\n"
f"Odgovori na hrvatskom jeziku."
)
response = c.models.generate_content(
model=model_name,
contents=[
types.Part.from_bytes(data=image_bytes, mime_type="image/jpeg"),
prompt
],
config=types.GenerateContentConfig(temperature=0.5, max_output_tokens=8192)
)
return response.text, citations
# =============================================================================
# FASTAPI APP
# =============================================================================
app = FastAPI(title="Learn Pathophysiology API", version="1.0.0")
app.add_middleware(
CORSMiddleware,
allow_origins=["*"],
allow_credentials=True,
allow_methods=["*"],
allow_headers=["*"],
)
# --- Request / Response Models ---
class ChatRequest(BaseModel):
message: str
model: str = ""
history: list = []
# --- Auth Models ---
class GoogleAuthRequest(BaseModel):
credential: str # Google ID token
# --- Auth Endpoints ---
@app.get("/api/auth/config")
async def auth_config():
"""Tell the frontend if auth is required and the Google Client ID."""
return {
"auth_enabled": AUTH_ENABLED,
"google_client_id": GOOGLE_CLIENT_ID if AUTH_ENABLED else None,
}
@app.post("/api/auth/google")
async def auth_google(req: GoogleAuthRequest):
"""Verify Google ID token and return a JWT session token."""
if not AUTH_ENABLED:
raise HTTPException(status_code=400, detail="Auth not enabled")
try:
idinfo = google_id_token.verify_oauth2_token(
req.credential,
google_requests.Request(),
GOOGLE_CLIENT_ID,
)
email = idinfo.get("email", "")
name = idinfo.get("name", email)
picture = idinfo.get("picture", "")
token = create_jwt(email, name, picture)
return {
"token": token,
"user": {"email": email, "name": name, "picture": picture},
}
except ValueError as e:
logger.error(f"Google auth failed: {e}")
raise HTTPException(status_code=401, detail="Invalid Google token")
@app.get("/api/auth/me")
async def auth_me(user=Depends(require_auth)):
"""Return the current user's info from their JWT."""
return {
"email": user.get("sub", ""),
"name": user.get("name", ""),
"picture": user.get("picture", ""),
}
# --- API Endpoints (public) ---
@app.get("/api/health")
async def health():
coll = get_collection()
return {
"status": "ok",
"chroma_docs": coll.count() if coll else 0,
"has_api_key": bool(GEMINI_API_KEY),
}
@app.get("/api/models")
async def list_models():
return {"models": AVAILABLE_MODELS, "default": DEFAULT_MODEL}
@app.get("/api/stats")
async def stats():
coll = get_collection()
return {
"documents": coll.count() if coll else 0,
"collection": COLLECTION_NAME,
}
# --- API Endpoints (protected) ---
@app.post("/api/chat")
async def chat(req: ChatRequest, user=Depends(require_auth)):
try:
model = req.model or DEFAULT_MODEL
reply, citations = generate_chat_response(req.message, req.history, model)
return {"reply": reply, "citations": citations, "model_used": model}
except Exception as e:
logger.error(f"Chat error: {e}")
raise HTTPException(status_code=500, detail=str(e))
@app.post("/api/analyze-image")
async def analyze_image_endpoint(
image: UploadFile = File(...),
question: str = Form(""),
model: str = Form(""),
user=Depends(require_auth),
):
try:
model_name = model or DEFAULT_MODEL
image_bytes = await image.read()
analysis, citations = do_analyze_image(image_bytes, question, model_name)
return {"analysis": analysis, "citations": citations, "model_used": model_name}
except Exception as e:
logger.error(f"Image analysis error: {e}")
raise HTTPException(status_code=500, detail=str(e))
# --- Google OAuth Redirect Handler ---
# When using ux_mode='redirect', Google POSTs credential as form data to login_uri
@app.post("/")
async def google_redirect_callback(request: Request):
"""Handle Google Sign-In redirect (POST with credential in form data)."""
form = await request.form()
credential = form.get("credential", "")
if credential and AUTH_ENABLED:
try:
idinfo = google_id_token.verify_oauth2_token(
str(credential),
google_requests.Request(),
GOOGLE_CLIENT_ID,
)
email = idinfo.get("email", "")
name = idinfo.get("name", email)
picture = idinfo.get("picture", "")
token = create_jwt(email, name, picture)
# Return an HTML page that stores the token and redirects to the app
html = f"""<!DOCTYPE html>
<html><head><title>Logging in...</title></head>
<body><script>
localStorage.setItem('lp_auth_token', '{token}');
window.location.href = '/';
</script><p>Logging in...</p></body></html>"""
from fastapi.responses import HTMLResponse
return HTMLResponse(content=html)
except Exception as e:
logger.error(f"Google redirect auth failed: {e}")
# Fallback: serve index.html
static_dir = Path(__file__).parent / "static"
return FileResponse(str(static_dir / "index.html"))
# --- Serve Vue Frontend (production) ---
static_dir = Path(__file__).parent / "static"
if static_dir.exists():
@app.get("/")
async def serve_index():
return FileResponse(str(static_dir / "index.html"))
app.mount("/", StaticFiles(directory=str(static_dir), html=True), name="static")
if __name__ == "__main__":
import uvicorn
uvicorn.run("api:app", host="0.0.0.0", port=7860, reload=True)
|