src/ocr_utils.py

import io, re
from typing import List, Dict, Tuple, Any
from PIL import Image, ImageDraw

try:
    import pytesseract
    from pytesseract import Output as TessOutput
except Exception:
    pytesseract = None
    TessOutput = None

PRICE_RE = re.compile(r"(\$\s*\d{1,4}(?:\.\d{2})?|\b\d{1,4}\.\d{2}\b)")

def ocr_image(image_bytes: bytes):
    """
    ALWAYS return (full_text, tokens, size) even if OCR fails.
    """
    try:
        img = Image.open(io.BytesIO(image_bytes)).convert("RGB")
    except Exception:
        return "", [], (0, 0)

    if pytesseract is None:
        return "", [], img.size

    try:
        data = pytesseract.image_to_data(img, output_type=TessOutput.DICT)
    except Exception:
        return "", [], img.size

    tokens: List[Dict[str, Any]] = []
    for i, txt in enumerate(data["text"]):
        if not txt:
            continue
        try:
            conf = float(data.get("conf", ["-1"])[i])
        except Exception:
            conf = -1.0
        x, y, w, h = data["left"][i], data["top"][i], data["width"][i], data["height"][i]
        tokens.append({"text": txt, "conf": conf, "box": (x, y, w, h)})

    full_text = " ".join(t["text"] for t in tokens)
    return full_text, tokens, img.size


def guess_price(tokens: List[Dict[str, Any]]):
    """Find smallest plausible price in tokens."""
    best = None
    for t in tokens:
        m = PRICE_RE.search(t["text"].replace(",", ""))
        if m:
            raw = m.group(0).replace("$", "").strip()
            try:
                val = float(raw)
                if 0.5 <= val <= 1000:
                    if best is None or val < best[0]:
                        best = (val, t["box"])
            except Exception:
                pass
    return best if best else (None, None)


def guess_title(text: str) -> str:
    """Make a short human-readable title from OCR text."""
    words = re.findall(r"[A-Za-z0-9\-]{3,}", text or "")
    ret