ocr/models.py

"""OCR result data models."""

from __future__ import annotations

from dataclasses import dataclass, field


@dataclass
class BoundingRect:
    """Quadrilateral bounding box (4 corners)."""
    x1: float = 0.0
    y1: float = 0.0
    x2: float = 0.0
    y2: float = 0.0
    x3: float = 0.0
    y3: float = 0.0
    x4: float = 0.0
    y4: float = 0.0


@dataclass
class OcrWord:
    """Single recognized word with position and confidence."""
    text: str
    bounding_rect: BoundingRect | None = None
    confidence: float = 0.0


@dataclass
class OcrLine:
    """Single line of recognized text."""
    text: str
    bounding_rect: BoundingRect | None = None
    words: list[OcrWord] = field(default_factory=list)


@dataclass
class OcrResult:
    """Complete OCR recognition result."""
    text: str = ""
    text_angle: float | None = None
    lines: list[OcrLine] = field(default_factory=list)
    error: str | None = None

    @property
    def is_empty(self) -> bool:
        return not self.text.strip()

    @property
    def average_confidence(self) -> float:
        """Average word-level confidence across all lines."""
        all_words = [w for line in self.lines for w in line.words if w.confidence > 0]
        if not all_words:
            return 0.0
        return sum(w.confidence for w in all_words) / len(all_words)