services/document-parser-api/main.py

"""
Hosted document parser API for Fresh Catch Inventory.

Deploy to Hugging Face Spaces (Docker), Fly.io, or any VM with Python + Tesseract.
"""

from __future__ import annotations

import os
import subprocess
import sys
import tempfile
from pathlib import Path

from fastapi import FastAPI, File, Header, HTTPException, Query, UploadFile
from fastapi.middleware.cors import CORSMiddleware

REPO_ROOT = Path(__file__).resolve().parents[2]
PARSE_SCRIPT = REPO_ROOT / "scripts" / "parse_vendor_document.py"
SERVICE_SECRET = os.environ.get("DOCUMENT_PARSER_SERVICE_SECRET", "").strip()

app = FastAPI(title="Fresh Catch Document Parser", version="1.0.0")

app.add_middleware(
    CORSMiddleware,
    allow_origins=os.environ.get("DOCUMENT_PARSER_CORS_ORIGINS", "*").split(","),
    allow_credentials=True,
    allow_methods=["POST", "GET"],
    allow_headers=["*"],
)


def verify_auth(authorization: str | None) -> None:
    if not SERVICE_SECRET:
        return
    if not authorization or authorization != f"Bearer {SERVICE_SECRET}":
        raise HTTPException(status_code=401, detail="Unauthorized")


@app.get("/health")
def health() -> dict[str, str]:
    return {"status": "ok"}


@app.post("/parse")
async def parse_document(
    file: UploadFile = File(...),
    type: str = Query("auto", pattern="^(auto|invoice|receipt)$"),
    authorization: str | None = Header(default=None),
) -> dict:
    verify_auth(authorization)

    if not PARSE_SCRIPT.exists():
        raise HTTPException(status_code=500, detail="parse_vendor_document.py not found")

    contents = await file.read()
    if not contents:
        raise HTTPException(status_code=400, detail="Empty file")

    suffix = Path(file.filename or "upload.png").suffix or ".png"
    with tempfile.NamedTemporaryFile(suffix=suffix, delete=False) as tmp:
        tmp.write(contents)
        image_path = tmp.name

    try:
        completed = subprocess.run(
            [sys.executable, str(PARSE_SCRIPT), "--image", image_path, "--type", type],
            capture_output=True,
            text=True,
            timeout=int(os.environ.get("DOCUMENT_PARSER_TIMEOUT_MS", "120000")) // 1000,
            cwd=str(REPO_ROOT),
        )
    finally:
        Path(image_path).unlink(missing_ok=True)

    if completed.returncode != 0:
        detail = (completed.stderr or completed.stdout or "Parse failed").strip()
        raise HTTPException(status_code=500, detail=detail[:2000])

    import json

    try:
        return json.loads(completed.stdout)
    except json.JSONDecodeError as error:
        raise HTTPException(status_code=500, detail=f"Invalid parser output: {error}") from error