Spaces:
Paused
Paused
File size: 13,514 Bytes
f66643d | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 | """Data models for the scanned PDF pipeline.
This module defines the core dataclasses used throughout Stage A:
- CellData: Individual table cell with bbox, row/col indices, and text
- ElementData: A layout element (text block, figure, table, etc.)
- PageData: A single page with dimensions, elements, and metadata
- ChapterInfo: Chapter metadata (filled by Stage B)
- ParsedDocument: The complete parsed document structure
"""
from __future__ import annotations
import json
from dataclasses import dataclass, field
from pathlib import Path
from typing import Any
from PIL import Image
from pdf2zh.parser.enums import ElementCategory
@dataclass
class CellData:
"""A single cell within a TABLE element.
Attributes:
bbox_pdf: [x0, y0, x1, y1] in absolute PDF points (page-level coordinates)
row_id: 0-based row index
col_id: 0-based column index
source_text: OCR text content; empty string for empty cells
translated_text: Empty string after Stage A; filled by Stage C
"""
bbox_pdf: list[float]
bbox_text: list[float]
source_text: str = ""
translated_text: str = ""
def to_dict(self) -> dict[str, Any]:
"""Serialize this cell to a JSON-compatible dictionary.
Returns:
Dict with keys ``bbox_pdf``, ``bbox_text``, ``source_text``, and ``translated_text``.
"""
return {
"bbox_pdf": self.bbox_pdf,
"bbox_text": self.bbox_text,
"source_text": self.source_text,
"translated_text": self.translated_text,
}
@classmethod
def from_dict(cls, data: dict[str, Any]) -> CellData:
"""Deserialize a :class:`CellData` from a plain dictionary.
Args:
data: Dictionary as produced by :meth:`to_dict`. ``translated_text``
defaults to ``""`` if absent (backward compatibility).
Returns:
New :class:`CellData` instance.
"""
return cls(
bbox_pdf=data["bbox_pdf"],
bbox_text=data.get("bbox_text", data["bbox_pdf"]),
source_text=data["source_text"],
translated_text=data.get("translated_text", ""),
)
@dataclass
class ElementData:
"""A layout element detected by Surya.
Attributes:
label: Raw Surya label (e.g., "Text", "Section-header", "Table")
category: One of the 5 ElementCategory values determining handling
bbox_pdf: [x0, y0, x1, y1] in PDF points; x0 < x1, y0 < y1
source_text: OCR text; always "" for BYPASS and optional for EQUATION
translated_text: Empty string after Stage A; filled by Stage C
cells: Non-empty only for TABLE category; empty list otherwise
"""
label: str
category: ElementCategory
bbox_pdf: list[float]
source_text: str
translated_text: str = ""
cells: list[CellData] = field(default_factory=list)
def to_dict(self) -> dict[str, Any]:
"""Serialize this element to a JSON-compatible dictionary."""
return {
"label": self.label,
"category": (
self.category.value
if isinstance(self.category, ElementCategory)
else self.category
),
"bbox_pdf": self.bbox_pdf,
"source_text": self.source_text,
"translated_text": self.translated_text,
"cells": [c.to_dict() for c in self.cells],
}
@classmethod
def from_dict(cls, data: dict[str, Any]) -> ElementData:
"""Deserialize an :class:`ElementData` from a plain dictionary."""
return cls(
label=data["label"],
category=ElementCategory(data["category"]),
bbox_pdf=data["bbox_pdf"],
source_text=data["source_text"],
translated_text=data.get("translated_text", ""),
cells=[CellData.from_dict(c) for c in data.get("cells", [])],
)
@dataclass
class PageData:
"""Data for a single PDF page.
Attributes:
page_index: 0-based page number
page_width: Width in PDF points (from page.rect.width)
page_height: Height in PDF points (from page.rect.height)
elements: Layout elements in top-to-bottom reading order
raw_text: Joined source_text of FLOWING_TEXT and IN_PLACE elements
chapter_id: Empty string after Stage A; filled by Stage B
"""
page_index: int
page_width: float
page_height: float
elements: list[ElementData] = field(default_factory=list)
raw_text: str = ""
chapter_id: str = ""
def to_dict(self) -> dict[str, Any]:
"""Serialize this page to a JSON-compatible dictionary.
Returns:
Dict with keys ``page_index``, ``page_width``, ``page_height``,
``elements``, ``raw_text``, and ``chapter_id``.
"""
return {
"page_index": self.page_index,
"page_width": self.page_width,
"page_height": self.page_height,
"elements": [e.to_dict() for e in self.elements],
"raw_text": self.raw_text,
"chapter_id": self.chapter_id,
}
@classmethod
def from_dict(cls, data: dict[str, Any]) -> PageData:
"""Deserialize a :class:`PageData` from a plain dictionary.
Args:
data: Dictionary as produced by :meth:`to_dict`. Optional keys
``elements``, ``raw_text``, and ``chapter_id`` default to
``[]``, ``""``, and ``""`` respectively.
Returns:
New :class:`PageData` instance.
"""
return cls(
page_index=data["page_index"],
page_width=data["page_width"],
page_height=data["page_height"],
elements=[ElementData.from_dict(e) for e in data.get("elements", [])],
raw_text=data.get("raw_text", ""),
chapter_id=data.get("chapter_id", ""),
)
@dataclass
class ChapterInfo:
"""Chapter metadata (empty after Stage A, filled by Stage B).
Attributes:
chapter_id: Identifier like "ch_0", "ch_1", etc.
title: Chapter heading text; empty if not found
start_page: 0-based inclusive start page
end_page: 0-based inclusive end page (end_page >= start_page)
summary: LLM-generated summary; empty initially
glossary: {term: definition}; empty initially
"""
chapter_id: str
title: str
start_page: int
end_page: int
summary: str = ""
glossary: dict[str, str] = field(default_factory=dict)
def to_dict(self) -> dict[str, Any]:
"""Serialize this chapter info to a JSON-compatible dictionary.
Returns:
Dict with keys ``chapter_id``, ``title``, ``start_page``,
``end_page``, ``summary``, and ``glossary``.
"""
return {
"chapter_id": self.chapter_id,
"title": self.title,
"start_page": self.start_page,
"end_page": self.end_page,
"summary": self.summary,
"glossary": self.glossary,
}
@classmethod
def from_dict(cls, data: dict[str, Any]) -> ChapterInfo:
"""Deserialize a :class:`ChapterInfo` from a plain dictionary.
Args:
data: Dictionary as produced by :meth:`to_dict`. Optional keys
``summary`` and ``glossary`` default to ``""`` and ``{}``.
Returns:
New :class:`ChapterInfo` instance.
"""
return cls(
chapter_id=data["chapter_id"],
title=data["title"],
start_page=data["start_page"],
end_page=data["end_page"],
summary=data.get("summary", ""),
glossary=data.get("glossary", {}),
)
@dataclass
class ParsedDocument:
"""Complete parsed document from Stage A.
Attributes:
pdf_path: Path to the source PDF file
pages: List of PageData, one per page, 0-based order
chapters: Empty list after Stage A; filled by Stage B
glossary: Empty dict after Stage A; filled by Stage B
"""
pdf_path: str
pages: list[PageData] = field(default_factory=list)
chapters: list[ChapterInfo] = field(default_factory=list)
glossary: dict[str, str] = field(default_factory=dict)
def to_dict(self) -> dict[str, Any]:
"""Serialize the full document to a JSON-compatible dictionary.
Returns:
Dict with keys ``pdf_path``, ``pages``, ``chapters``,
and ``glossary``.
"""
return {
"pdf_path": self.pdf_path,
"pages": [p.to_dict() for p in self.pages],
"chapters": [c.to_dict() for c in self.chapters],
"glossary": self.glossary,
}
@classmethod
def from_dict(cls, data: dict[str, Any]) -> ParsedDocument:
"""Deserialize a :class:`ParsedDocument` from a plain dictionary.
Args:
data: Dictionary as produced by :meth:`to_dict`. Optional keys
``pages``, ``chapters``, and ``glossary`` default to
``[]``, ``[]``, and ``{}`` respectively.
Returns:
New :class:`ParsedDocument` instance.
"""
return cls(
pdf_path=data["pdf_path"],
pages=[PageData.from_dict(p) for p in data.get("pages", [])],
chapters=[ChapterInfo.from_dict(c) for c in data.get("chapters", [])],
glossary=data.get("glossary", {}),
)
def to_json(self, indent: int = 2) -> str:
"""Serialize the document to a JSON string.
Args:
indent: Number of spaces for JSON indentation (default 2).
Returns:
Pretty-printed JSON string with UTF-8 characters unescaped
(``ensure_ascii=False``).
"""
return json.dumps(self.to_dict(), indent=indent, ensure_ascii=False)
@classmethod
def from_json(cls, json_str: str) -> ParsedDocument:
"""Deserialize a :class:`ParsedDocument` from a JSON string.
Args:
json_str: JSON string as produced by :meth:`to_json`.
Returns:
New :class:`ParsedDocument` instance.
"""
return cls.from_dict(json.loads(json_str))
def save(self, path: str | Path) -> None:
"""Save the document to a JSON file on disk.
Creates parent directories if they do not exist.
Args:
path: Destination file path (``str`` or :class:`~pathlib.Path`).
"""
path = Path(path)
path.write_text(self.to_json(), encoding="utf-8")
@classmethod
def load(cls, path: str | Path) -> ParsedDocument:
"""Load a :class:`ParsedDocument` from a JSON file on disk.
Args:
path: Source file path (``str`` or :class:`~pathlib.Path`).
Returns:
New :class:`ParsedDocument` instance parsed from the file.
"""
path = Path(path)
return cls.from_json(path.read_text(encoding="utf-8"))
@dataclass(slots=True)
class LayoutBlockResult:
"""A layout block with stable IDs and coordinates in all required spaces."""
block_id: str
page_index: int
position: int
label: str
category: ElementCategory
bbox_layout: list[float]
bbox_image: list[float]
bbox_pdf: list[float]
@dataclass(slots=True)
class LayoutPageResult:
"""Layout output for one page."""
page_index: int
page_width: float
page_height: float
layout_image_bbox: list[float]
image_bbox: list[float]
blocks: list[LayoutBlockResult] = field(default_factory=list)
@dataclass(slots=True)
class LayoutParseResult:
"""Full layout phase output."""
pdf_path: str
pages: list[LayoutPageResult] = field(default_factory=list)
def page_map(self) -> dict[int, LayoutPageResult]:
return {page.page_index: page for page in self.pages}
def block_map(self) -> dict[str, LayoutBlockResult]:
return {block.block_id: block for page in self.pages for block in page.blocks}
@dataclass(slots=True)
class OCRPageResult:
"""OCR output for one full page."""
page_index: int
image_bbox: list[float]
ocr_result: Any
@property
def image_width(self) -> float:
return self.image_bbox[2] - self.image_bbox[0]
@property
def image_height(self) -> float:
return self.image_bbox[3] - self.image_bbox[1]
@dataclass(slots=True)
class OCRParseResult:
"""Full-page OCR phase output."""
pdf_path: str
pages: list[OCRPageResult] = field(default_factory=list)
def page_map(self) -> dict[int, OCRPageResult]:
return {page.page_index: page for page in self.pages}
@dataclass(slots=True)
class TableBlockResult:
"""Merged table output for one layout table block."""
block_id: str
cells_bbox: list[list[float]]
crop_size: tuple[float, float]
@dataclass(slots=True)
class TableParseResult:
"""Table phase output indexed by layout block id."""
pdf_path: str
tables: dict[str, TableBlockResult] = field(default_factory=dict)
@dataclass(slots=True)
class _DocumentContext:
"""Immutable page selection and geometry for one PDF parse request."""
pdf_path: Path
page_indices: list[int]
page_dims: dict[int, tuple[float, float]]
@dataclass(slots=True)
class _TableJob:
"""Bookkeeping for one table crop inside a batch."""
block: LayoutBlockResult
page_width: float
page_height: float
table_crop: Image.Image
|