import copy import os import statistics import time from typing import List import torch from loguru import logger from magic_pdf.config.enums import SupportedPdfParseMethod from magic_pdf.data.dataset import Dataset, PageableData from magic_pdf.libs.boxbase import calculate_overlap_area_in_bbox1_area_ratio from magic_pdf.libs.clean_memory import clean_memory from magic_pdf.libs.commons import fitz, get_delta_time from magic_pdf.libs.config_reader import get_local_layoutreader_model_dir from magic_pdf.libs.convert_utils import dict_to_list from magic_pdf.libs.drop_reason import DropReason from magic_pdf.libs.hash_utils import compute_md5 from magic_pdf.libs.local_math import float_equal from magic_pdf.libs.ocr_content_type import ContentType, BlockType from magic_pdf.model.magic_model import MagicModel from magic_pdf.para.para_split_v3 import para_split from magic_pdf.pre_proc.citationmarker_remove import remove_citation_marker from magic_pdf.pre_proc.construct_page_dict import \ ocr_construct_page_component_v2 from magic_pdf.pre_proc.cut_image import ocr_cut_image_and_table from magic_pdf.pre_proc.equations_replace import ( combine_chars_to_pymudict, remove_chars_in_text_blocks, replace_equations_in_textblock) from magic_pdf.pre_proc.ocr_detect_all_bboxes import \ ocr_prepare_bboxes_for_layout_split_v2 from magic_pdf.pre_proc.ocr_dict_merge import (fill_spans_in_blocks, fix_block_spans, fix_discarded_block, fix_block_spans_v2) from magic_pdf.pre_proc.ocr_span_list_modify import ( get_qa_need_list_v2, remove_overlaps_low_confidence_spans, remove_overlaps_min_spans) from magic_pdf.pre_proc.resolve_bbox_conflict import \ check_useful_block_horizontal_overlap def remove_horizontal_overlap_block_which_smaller(all_bboxes): useful_blocks = [] for bbox in all_bboxes: useful_blocks.append({'bbox': bbox[:4]}) is_useful_block_horz_overlap, smaller_bbox, bigger_bbox = ( check_useful_block_horizontal_overlap(useful_blocks) ) if is_useful_block_horz_overlap: logger.warning( f'skip this page, reason: {DropReason.USEFUL_BLOCK_HOR_OVERLAP}, smaller bbox is {smaller_bbox}, bigger bbox is {bigger_bbox}' ) # noqa: E501 for bbox in all_bboxes.copy(): if smaller_bbox == bbox[:4]: all_bboxes.remove(bbox) return is_useful_block_horz_overlap, all_bboxes def __replace_STX_ETX(text_str: str): """Replace \u0002 and \u0003, as these characters become garbled when extracted using pymupdf. In fact, they were originally quotation marks. Drawback: This issue is only observed in English text; it has not been found in Chinese text so far. Args: text_str (str): raw text Returns: _type_: replaced text """ # noqa: E501 if text_str: s = text_str.replace('\u0002', "'") s = s.replace('\u0003', "'") return s return text_str def txt_spans_extract(pdf_page, inline_equations, interline_equations): text_raw_blocks = pdf_page.get_text('dict', flags=fitz.TEXTFLAGS_TEXT)['blocks'] char_level_text_blocks = pdf_page.get_text('rawdict', flags=fitz.TEXTFLAGS_TEXT)[ 'blocks' ] text_blocks = combine_chars_to_pymudict(text_raw_blocks, char_level_text_blocks) text_blocks = replace_equations_in_textblock( text_blocks, inline_equations, interline_equations ) text_blocks = remove_citation_marker(text_blocks) text_blocks = remove_chars_in_text_blocks(text_blocks) spans = [] for v in text_blocks: for line in v['lines']: for span in line['spans']: bbox = span['bbox'] if float_equal(bbox[0], bbox[2]) or float_equal(bbox[1], bbox[3]): continue if span.get('type') not in ( ContentType.InlineEquation, ContentType.InterlineEquation, ): spans.append( { 'bbox': list(span['bbox']), 'content': __replace_STX_ETX(span['text']), 'type': ContentType.Text, 'score': 1.0, } ) return spans def replace_text_span(pymu_spans, ocr_spans): return list(filter(lambda x: x['type'] != ContentType.Text, ocr_spans)) + pymu_spans def model_init(model_name: str): from transformers import LayoutLMv3ForTokenClassification if torch.cuda.is_available(): device = torch.device('cuda') if torch.cuda.is_bf16_supported(): supports_bfloat16 = True else: supports_bfloat16 = False else: device = torch.device('cpu') supports_bfloat16 = False if model_name == 'layoutreader': # 检测modelscope的缓存目录是否存在 layoutreader_model_dir = get_local_layoutreader_model_dir() if os.path.exists(layoutreader_model_dir): model = LayoutLMv3ForTokenClassification.from_pretrained( layoutreader_model_dir ) else: logger.warning( 'local layoutreader model not exists, use online model from huggingface' ) model = LayoutLMv3ForTokenClassification.from_pretrained( 'hantian/layoutreader' ) # 检查设备是否支持 bfloat16 if supports_bfloat16: model.bfloat16() model.to(device).eval() else: logger.error('model name not allow') exit(1) return model class ModelSingleton: _instance = None _models = {} def __new__(cls, *args, **kwargs): if cls._instance is None: cls._instance = super().__new__(cls) return cls._instance def get_model(self, model_name: str): if model_name not in self._models: self._models[model_name] = model_init(model_name=model_name) return self._models[model_name] def do_predict(boxes: List[List[int]], model) -> List[int]: from magic_pdf.model.v3.helpers import (boxes2inputs, parse_logits, prepare_inputs) inputs = boxes2inputs(boxes) inputs = prepare_inputs(inputs, model) logits = model(**inputs).logits.cpu().squeeze(0) return parse_logits(logits, len(boxes)) def cal_block_index(fix_blocks, sorted_bboxes): for block in fix_blocks: line_index_list = [] if len(block['lines']) == 0: block['index'] = sorted_bboxes.index(block['bbox']) else: for line in block['lines']: line['index'] = sorted_bboxes.index(line['bbox']) line_index_list.append(line['index']) median_value = statistics.median(line_index_list) block['index'] = median_value # 删除图表body block中的虚拟line信息, 并用real_lines信息回填 if block['type'] in [BlockType.ImageBody, BlockType.TableBody]: block['virtual_lines'] = copy.deepcopy(block['lines']) block['lines'] = copy.deepcopy(block['real_lines']) del block['real_lines'] return fix_blocks def insert_lines_into_block(block_bbox, line_height, page_w, page_h): # block_bbox是一个元组(x0, y0, x1, y1)，其中(x0, y0)是左下角坐标，(x1, y1)是右上角坐标 x0, y0, x1, y1 = block_bbox block_height = y1 - y0 block_weight = x1 - x0 # 如果block高度小于n行正文，则直接返回block的bbox if line_height * 3 < block_height: if ( block_height > page_h * 0.25 and page_w * 0.5 > block_weight > page_w * 0.25 ): # 可能是双列结构，可以切细点 lines = int(block_height / line_height) + 1 else: # 如果block的宽度超过0.4页面宽度，则将block分成3行(是一种复杂布局，图不能切的太细) if block_weight > page_w * 0.4: line_height = (y1 - y0) / 3 lines = 3 elif block_weight > page_w * 0.25: # （可能是三列结构，也切细点） lines = int(block_height / line_height) + 1 else: # 判断长宽比 if block_height / block_weight > 1.2: # 细长的不分 return [[x0, y0, x1, y1]] else: # 不细长的还是分成两行 line_height = (y1 - y0) / 2 lines = 2 # 确定从哪个y位置开始绘制线条 current_y = y0 # 用于存储线条的位置信息[(x0, y), ...] lines_positions = [] for i in range(lines): lines_positions.append([x0, current_y, x1, current_y + line_height]) current_y += line_height return lines_positions else: return [[x0, y0, x1, y1]] def sort_lines_by_model(fix_blocks, page_w, page_h, line_height): page_line_list = [] for block in fix_blocks: if block['type'] in [ BlockType.Text, BlockType.Title, BlockType.InterlineEquation, BlockType.ImageCaption, BlockType.ImageFootnote, BlockType.TableCaption, BlockType.TableFootnote ]: if len(block['lines']) == 0: bbox = block['bbox'] lines = insert_lines_into_block(bbox, line_height, page_w, page_h) for line in lines: block['lines'].append({'bbox': line, 'spans': []}) page_line_list.extend(lines) else: for line in block['lines']: bbox = line['bbox'] page_line_list.append(bbox) elif block['type'] in [BlockType.ImageBody, BlockType.TableBody]: bbox = block['bbox'] block["real_lines"] = copy.deepcopy(block['lines']) lines = insert_lines_into_block(bbox, line_height, page_w, page_h) block['lines'] = [] for line in lines: block['lines'].append({'bbox': line, 'spans': []}) page_line_list.extend(lines) # 使用layoutreader排序 x_scale = 1000.0 / page_w y_scale = 1000.0 / page_h boxes = [] # logger.info(f"Scale: {x_scale}, {y_scale}, Boxes len: {len(page_line_list)}") for left, top, right, bottom in page_line_list: if left < 0: logger.warning( f'left < 0, left: {left}, right: {right}, top: {top}, bottom: {bottom}, page_w: {page_w}, page_h: {page_h}' ) # noqa: E501 left = 0 if right > page_w: logger.warning( f'right > page_w, left: {left}, right: {right}, top: {top}, bottom: {bottom}, page_w: {page_w}, page_h: {page_h}' ) # noqa: E501 right = page_w if top < 0: logger.warning( f'top < 0, left: {left}, right: {right}, top: {top}, bottom: {bottom}, page_w: {page_w}, page_h: {page_h}' ) # noqa: E501 top = 0 if bottom > page_h: logger.warning( f'bottom > page_h, left: {left}, right: {right}, top: {top}, bottom: {bottom}, page_w: {page_w}, page_h: {page_h}' ) # noqa: E501 bottom = page_h left = round(left * x_scale) top = round(top * y_scale) right = round(right * x_scale) bottom = round(bottom * y_scale) assert ( 1000 >= right >= left >= 0 and 1000 >= bottom >= top >= 0 ), f'Invalid box. right: {right}, left: {left}, bottom: {bottom}, top: {top}' # noqa: E126, E121 boxes.append([left, top, right, bottom]) model_manager = ModelSingleton() model = model_manager.get_model('layoutreader') with torch.no_grad(): orders = do_predict(boxes, model) sorted_bboxes = [page_line_list[i] for i in orders] return sorted_bboxes def get_line_height(blocks): page_line_height_list = [] for block in blocks: if block['type'] in [ BlockType.Text, BlockType.Title, BlockType.ImageCaption, BlockType.ImageFootnote, BlockType.TableCaption, BlockType.TableFootnote ]: for line in block['lines']: bbox = line['bbox'] page_line_height_list.append(int(bbox[3] - bbox[1])) if len(page_line_height_list) > 0: return statistics.median(page_line_height_list) else: return 10 def process_groups(groups, body_key, caption_key, footnote_key): body_blocks = [] caption_blocks = [] footnote_blocks = [] for i, group in enumerate(groups): group[body_key]['group_id'] = i body_blocks.append(group[body_key]) for caption_block in group[caption_key]: caption_block['group_id'] = i caption_blocks.append(caption_block) for footnote_block in group[footnote_key]: footnote_block['group_id'] = i footnote_blocks.append(footnote_block) return body_blocks, caption_blocks, footnote_blocks def process_block_list(blocks, body_type, block_type): indices = [block['index'] for block in blocks] median_index = statistics.median(indices) body_bbox = next((block['bbox'] for block in blocks if block.get('type') == body_type), []) return { 'type': block_type, 'bbox': body_bbox, 'blocks': blocks, 'index': median_index, } def revert_group_blocks(blocks): image_groups = {} table_groups = {} new_blocks = [] for block in blocks: if block['type'] in [BlockType.ImageBody, BlockType.ImageCaption, BlockType.ImageFootnote]: group_id = block['group_id'] if group_id not in image_groups: image_groups[group_id] = [] image_groups[group_id].append(block) elif block['type'] in [BlockType.TableBody, BlockType.TableCaption, BlockType.TableFootnote]: group_id = block['group_id'] if group_id not in table_groups: table_groups[group_id] = [] table_groups[group_id].append(block) else: new_blocks.append(block) for group_id, blocks in image_groups.items(): new_blocks.append(process_block_list(blocks, BlockType.ImageBody, BlockType.Image)) for group_id, blocks in table_groups.items(): new_blocks.append(process_block_list(blocks, BlockType.TableBody, BlockType.Table)) return new_blocks def remove_outside_spans(spans, all_bboxes, all_discarded_blocks): def get_block_bboxes(blocks, block_type_list): return [block[0:4] for block in blocks if block[7] in block_type_list] image_bboxes = get_block_bboxes(all_bboxes, [BlockType.ImageBody]) table_bboxes = get_block_bboxes(all_bboxes, [BlockType.TableBody]) other_block_type = [] for block_type in BlockType.__dict__.values(): if not isinstance(block_type, str): continue if block_type not in [BlockType.ImageBody, BlockType.TableBody]: other_block_type.append(block_type) other_block_bboxes = get_block_bboxes(all_bboxes, other_block_type) discarded_block_bboxes = get_block_bboxes(all_discarded_blocks, [BlockType.Discarded]) new_spans = [] for span in spans: span_bbox = span['bbox'] span_type = span['type'] if any(calculate_overlap_area_in_bbox1_area_ratio(span_bbox, block_bbox) > 0.4 for block_bbox in discarded_block_bboxes): new_spans.append(span) continue if span_type == ContentType.Image: if any(calculate_overlap_area_in_bbox1_area_ratio(span_bbox, block_bbox) > 0.5 for block_bbox in image_bboxes): new_spans.append(span) elif span_type == ContentType.Table: if any(calculate_overlap_area_in_bbox1_area_ratio(span_bbox, block_bbox) > 0.5 for block_bbox in table_bboxes): new_spans.append(span) else: if any(calculate_overlap_area_in_bbox1_area_ratio(span_bbox, block_bbox) > 0.5 for block_bbox in other_block_bboxes): new_spans.append(span) return new_spans def parse_page_core( page_doc: PageableData, magic_model, page_id, pdf_bytes_md5, imageWriter, parse_mode ): need_drop = False drop_reason = [] """从magic_model对象中获取后面会用到的区块信息""" # img_blocks = magic_model.get_imgs(page_id) # table_blocks = magic_model.get_tables(page_id) img_groups = magic_model.get_imgs_v2(page_id) table_groups = magic_model.get_tables_v2(page_id) img_body_blocks, img_caption_blocks, img_footnote_blocks = process_groups( img_groups, 'image_body', 'image_caption_list', 'image_footnote_list' ) table_body_blocks, table_caption_blocks, table_footnote_blocks = process_groups( table_groups, 'table_body', 'table_caption_list', 'table_footnote_list' ) discarded_blocks = magic_model.get_discarded(page_id) text_blocks = magic_model.get_text_blocks(page_id) title_blocks = magic_model.get_title_blocks(page_id) inline_equations, interline_equations, interline_equation_blocks = ( magic_model.get_equations(page_id) ) page_w, page_h = magic_model.get_page_size(page_id) """将所有区块的bbox整理到一起""" # interline_equation_blocks参数不够准，后面切换到interline_equations上 interline_equation_blocks = [] if len(interline_equation_blocks) > 0: all_bboxes, all_discarded_blocks = ocr_prepare_bboxes_for_layout_split_v2( img_body_blocks, img_caption_blocks, img_footnote_blocks, table_body_blocks, table_caption_blocks, table_footnote_blocks, discarded_blocks, text_blocks, title_blocks, interline_equation_blocks, page_w, page_h, ) else: all_bboxes, all_discarded_blocks = ocr_prepare_bboxes_for_layout_split_v2( img_body_blocks, img_caption_blocks, img_footnote_blocks, table_body_blocks, table_caption_blocks, table_footnote_blocks, discarded_blocks, text_blocks, title_blocks, interline_equations, page_w, page_h, ) spans = magic_model.get_all_spans(page_id) """根据parse_mode，构造spans""" if parse_mode == SupportedPdfParseMethod.TXT: """ocr 中文本类的 span 用 pymu spans 替换！""" pymu_spans = txt_spans_extract(page_doc, inline_equations, interline_equations) spans = replace_text_span(pymu_spans, spans) elif parse_mode == SupportedPdfParseMethod.OCR: pass else: raise Exception('parse_mode must be txt or ocr') """在删除重复span之前，应该通过image_body和table_body的block过滤一下image和table的span""" """顺便删除大水印并保留abandon的span""" spans = remove_outside_spans(spans, all_bboxes, all_discarded_blocks) """删除重叠spans中置信度较低的那些""" spans, dropped_spans_by_confidence = remove_overlaps_low_confidence_spans(spans) """删除重叠spans中较小的那些""" spans, dropped_spans_by_span_overlap = remove_overlaps_min_spans(spans) """对image和table截图""" spans = ocr_cut_image_and_table( spans, page_doc, page_id, pdf_bytes_md5, imageWriter ) """先处理不需要排版的discarded_blocks""" discarded_block_with_spans, spans = fill_spans_in_blocks( all_discarded_blocks, spans, 0.4 ) fix_discarded_blocks = fix_discarded_block(discarded_block_with_spans) """如果当前页面没有bbox则跳过""" if len(all_bboxes) == 0: logger.warning(f'skip this page, not found useful bbox, page_id: {page_id}') return ocr_construct_page_component_v2( [], [], page_id, page_w, page_h, [], [], [], interline_equations, fix_discarded_blocks, need_drop, drop_reason, ) """将span填入blocks中""" block_with_spans, spans = fill_spans_in_blocks(all_bboxes, spans, 0.5) """对block进行fix操作""" fix_blocks = fix_block_spans_v2(block_with_spans) """获取所有line并计算正文line的高度""" line_height = get_line_height(fix_blocks) """获取所有line并对line排序""" sorted_bboxes = sort_lines_by_model(fix_blocks, page_w, page_h, line_height) """根据line的中位数算block的序列关系""" fix_blocks = cal_block_index(fix_blocks, sorted_bboxes) """将image和table的block还原回group形式参与后续流程""" fix_blocks = revert_group_blocks(fix_blocks) """重排block""" sorted_blocks = sorted(fix_blocks, key=lambda b: b['index']) """获取QA需要外置的list""" images, tables, interline_equations = get_qa_need_list_v2(sorted_blocks) """构造pdf_info_dict""" page_info = ocr_construct_page_component_v2( sorted_blocks, [], page_id, page_w, page_h, [], images, tables, interline_equations, fix_discarded_blocks, need_drop, drop_reason, ) return page_info def pdf_parse_union( dataset: Dataset, model_list, imageWriter, parse_mode, start_page_id=0, end_page_id=None, debug_mode=False, ): pdf_bytes_md5 = compute_md5(dataset.data_bits()) """初始化空的pdf_info_dict""" pdf_info_dict = {} """用model_list和docs对象初始化magic_model""" magic_model = MagicModel(model_list, dataset) """根据输入的起始范围解析pdf""" # end_page_id = end_page_id if end_page_id else len(pdf_docs) - 1 end_page_id = ( end_page_id if end_page_id is not None and end_page_id >= 0 else len(dataset) - 1 ) if end_page_id > len(dataset) - 1: logger.warning('end_page_id is out of range, use pdf_docs length') end_page_id = len(dataset) - 1 """初始化启动时间""" start_time = time.time() for page_id, page in enumerate(dataset): """debug时输出每页解析的耗时.""" if debug_mode: time_now = time.time() logger.info( f'page_id: {page_id}, last_page_cost_time: {get_delta_time(start_time)}' ) start_time = time_now """解析pdf中的每一页""" if start_page_id <= page_id <= end_page_id: page_info = parse_page_core( page, magic_model, page_id, pdf_bytes_md5, imageWriter, parse_mode ) else: page_info = page.get_page_info() page_w = page_info.w page_h = page_info.h page_info = ocr_construct_page_component_v2( [], [], page_id, page_w, page_h, [], [], [], [], [], True, 'skip page' ) pdf_info_dict[f'page_{page_id}'] = page_info """分段""" para_split(pdf_info_dict, debug_mode=debug_mode) """dict转list""" pdf_info_list = dict_to_list(pdf_info_dict) new_pdf_info_dict = { 'pdf_info': pdf_info_list, } clean_memory() return new_pdf_info_dict if __name__ == '__main__': pass