from __future__ import annotations

import io

import pytesseract
from fastapi import FastAPI, File, UploadFile
from PIL import Image
from pytesseract import Output

app = FastAPI(title="Local OCR Service", version="0.1.0")


@app.get("/health")
async def health() -> dict:
    return {"status": "ok"}


@app.post("/ocr")
async def ocr(file: UploadFile = File(...)) -> dict:
    data = await file.read()
    image = Image.open(io.BytesIO(data)).convert("RGB")
    structured = pytesseract.image_to_data(image, output_type=Output.DICT)
    lines = []
    width, height = image.size
    order = 0
    for idx, text in enumerate(structured["text"]):
        text = text.strip()
        if not text:
            continue
        confidence = float(structured["conf"][idx]) / 100.0
        x = structured["left"][idx] / width
        y = structured["top"][idx] / height
        w = structured["width"][idx] / width
        h = structured["height"][idx] / height
        order += 1
        lines.append(
            {
                "id": f"line_{order}",
                "text": text,
                "bbox": [x, y, x + w, y + h],
                "confidence": max(0.0, min(1.0, confidence)),
                "reading_order": order,
                "region_type": "text_line",
            }
        )
    return {
        "markdown": "\n".join(line["text"] for line in lines),
        "lines": lines,
        "tables": [],
    }