52 lines
1.5 KiB
Python
52 lines
1.5 KiB
Python
from __future__ import annotations
|
|
|
|
import re
|
|
from dataclasses import dataclass, asdict
|
|
|
|
from app.fuzzy_match import best_fuzzy_match
|
|
|
|
|
|
ORDER_RE = re.compile(r"\b(?P<order>\d{4}/\d{4}/(?:[1-9]|[1-9]\d))\b")
|
|
DEFAULT_MODEL_MIN_SCORE = 0.72
|
|
DEFAULT_COLOR_MIN_SCORE = 0.72
|
|
|
|
|
|
@dataclass
|
|
class ParsedLabel:
|
|
order_number: str | None
|
|
color_code: str | None
|
|
product_model: str | None
|
|
raw_text: str
|
|
color_score: float | None = None
|
|
product_model_score: float | None = None
|
|
|
|
def to_dict(self) -> dict[str, str | float | None]:
|
|
return asdict(self)
|
|
|
|
|
|
def normalize_ocr_text(text: str) -> str:
|
|
return " ".join(text.replace("\n", " ").replace("\r", " ").split())
|
|
|
|
|
|
def parse_label_text(
|
|
text: str,
|
|
known_colors: list[str],
|
|
known_models: list[str],
|
|
model_min_score: float = DEFAULT_MODEL_MIN_SCORE,
|
|
color_min_score: float = DEFAULT_COLOR_MIN_SCORE,
|
|
) -> ParsedLabel:
|
|
normalized = normalize_ocr_text(text)
|
|
order_match = ORDER_RE.search(normalized)
|
|
|
|
color_match = best_fuzzy_match(normalized, known_colors, color_min_score)
|
|
model_match = best_fuzzy_match(normalized, known_models, model_min_score)
|
|
|
|
return ParsedLabel(
|
|
order_number=order_match.group("order") if order_match else None,
|
|
color_code=color_match.value if color_match else None,
|
|
product_model=model_match.value if model_match else None,
|
|
raw_text=normalized,
|
|
color_score=color_match.score if color_match else None,
|
|
product_model_score=model_match.score if model_match else None,
|
|
)
|