Merge pull request #1838 from myhloli/dev

refactor(magic_pdf): improve paragraph splitting logic and update dep…
This commit is contained in:
Xiaomeng Zhao
2025-03-04 19:09:55 +08:00
committed by GitHub
2 changed files with 17 additions and 14 deletions

View File

@@ -108,6 +108,22 @@ def __is_list_or_index_block(block):
):
multiple_para_flag = True
block_text = ''
for line in block['lines']:
line_text = ''
for span in line['spans']:
span_type = span['type']
if span_type == ContentType.Text:
line_text += span['content'].strip()
# 添加所有文本包括空行保持与block['lines']长度一致
lines_text_list.append(line_text)
block_text = ''.join(lines_text_list)
block_lang = detect_lang(block_text)
# logger.info(f"block_lang: {block_lang}")
for line in block['lines']:
line_mid_x = (line['bbox'][0] + line['bbox'][2]) / 2
block_mid_x = (block['bbox_fs'][0] + block['bbox_fs'][2]) / 2
@@ -119,19 +135,6 @@ def __is_list_or_index_block(block):
if abs(line_mid_x - block_mid_x) < line_height / 2:
center_close_num += 1
line_text = ''
for span in line['spans']:
span_type = span['type']
if span_type == ContentType.Text:
line_text += span['content'].strip()
# 添加所有文本包括空行保持与block['lines']长度一致
lines_text_list.append(line_text)
block_text = ''.join(lines_text_list)
block_lang = detect_lang(block_text)
# logger.info(f"block_lang: {block_lang}")
# 计算line左侧顶格数量是否大于2是否顶格用abs(block['bbox_fs'][0] - line['bbox'][0]) < line_height/2 来判断
if abs(block['bbox_fs'][0] - line['bbox'][0]) < line_height / 2:
left_close_num += 1

View File

@@ -1,7 +1,7 @@
boto3>=1.28.43
Brotli>=1.1.0
click>=8.1.7
fast-langdetect>=0.2.3
fast-langdetect>=0.2.3,<0.3.0
loguru>=0.6.0
numpy>=1.21.6,<2.0.0
pydantic>=2.7.2