重构目录结构

pre_proc/detect_footer_by_model.py → pdf_tools/pre_proc/detect_footer_by_model.py

-import os                   
-import collections      # 统计库
-import re               # 正则
-from libs.commons import fitz             # pyMuPDF库
-import json             # json
+from pdf_tools.libs.commons import fitz             # pyMuPDF库
 
 
 def parse_footers(page_ID: int, page: fitz.Page, json_from_DocXchain_obj: dict):

pre_proc/detect_footer_header_by_statistics.py → pdf_tools/pre_proc/detect_footer_header_by_statistics.py

 from collections import defaultdict
 
-from loguru import logger
-
-from libs.boxbase import _is_in, calculate_iou
+from pdf_tools.libs.boxbase import calculate_iou
 
 
 def compare_bbox_with_list(bbox, bbox_list, tolerance=1):

pre_proc/detect_footnote.py → pdf_tools/pre_proc/detect_footnote.py

-import os
 from collections import Counter
-import re               # 正则
-from libs.commons import fitz             # pyMuPDF库
-import json             # json
+from pdf_tools.libs.commons import fitz             # pyMuPDF库
 
 
 def parse_footnotes_by_model(page_ID: int, page: fitz.Page, json_from_DocXchain_obj: dict, md_bookname_save_path, debug_mode=False):

pre_proc/detect_header.py → pdf_tools/pre_proc/detect_header.py

-import os                   
-import collections      # 统计库
-import re               # 正则
-from libs.commons import fitz             # pyMuPDF库
-import json             # json
+from pdf_tools.libs.commons import fitz             # pyMuPDF库
 
 
 def parse_headers(page_ID: int, page: fitz.Page, json_from_DocXchain_obj: dict):

pre_proc/detect_images.py → pdf_tools/pre_proc/detect_images.py

-import os                   
 import collections      # 统计库
 import re
-from libs.boxbase import _is_in_or_part_overlap               # 正则
-from libs.commons import fitz             # pyMuPDF库
-import json             # json
+from pdf_tools.libs.commons import fitz             # pyMuPDF库
 
 
 #--------------------------------------- Tool Functions --------------------------------------#

pre_proc/detect_page_number.py → pdf_tools/pre_proc/detect_page_number.py

-import os                   
-import collections      # 统计库
-import re               # 正则
-from libs.commons import fitz             # pyMuPDF库
-import json             # json
+from pdf_tools.libs.commons import fitz             # pyMuPDF库
 
 
 def parse_pageNos(page_ID: int, page: fitz.Page, json_from_DocXchain_obj: dict):

pre_proc/detect_tables.py → pdf_tools/pre_proc/detect_tables.py

-import os                   
-import collections      # 统计库
-import re               # 正则
-from libs.commons import fitz             # pyMuPDF库
-import json             # json
+from pdf_tools.libs.commons import fitz             # pyMuPDF库
 
 
 def parse_tables(page_ID: int, page: fitz.Page, json_from_DocXchain_obj: dict):

pre_proc/equations_replace.py → pdf_tools/pre_proc/equations_replace.py

 """
 对pymupdf返回的结构里的公式进行替换，替换为模型识别的公式结果
 """
-from libs.commons import fitz
+from pdf_tools.libs.commons import fitz
 import json
 import os
 from pathlib import Path

pre_proc/fix_image.py → pdf_tools/pre_proc/fix_image.py

@@ -2,10 +2,9 @@
 
 
 import re    
-from libs.boxbase import  _is_in_or_part_overlap, _is_part_overlap, _is_in, find_bottom_nearest_text_bbox, find_left_nearest_text_bbox, find_right_nearest_text_bbox, find_top_nearest_text_bbox
-from loguru import logger
+from pdf_tools.libs.boxbase import  _is_in_or_part_overlap, _is_part_overlap, find_bottom_nearest_text_bbox, find_left_nearest_text_bbox, find_right_nearest_text_bbox, find_top_nearest_text_bbox
 
-from libs.textbase import get_text_block_base_info
+from pdf_tools.libs.textbase import get_text_block_base_info
 
 def fix_image_vertical(image_bboxes:list, text_blocks:list):
     """

pre_proc/fix_table.py → pdf_tools/pre_proc/fix_table.py

-import os                   
-import collections      # 统计库
-import re               # 正则
-from libs.commons import fitz             # pyMuPDF库
-import json
+from pdf_tools.libs.commons import fitz             # pyMuPDF库
 import re
 
-from libs.boxbase import _is_in_or_part_overlap, _is_part_overlap, find_bottom_nearest_text_bbox, find_left_nearest_text_bbox, find_right_nearest_text_bbox, find_top_nearest_text_bbox             # json
+from pdf_tools.libs.boxbase import _is_in_or_part_overlap, _is_part_overlap, find_bottom_nearest_text_bbox, find_left_nearest_text_bbox, find_right_nearest_text_bbox, find_top_nearest_text_bbox             # json
 
 
 ## version 2

pre_proc/pdf_pre_filter.py → pdf_tools/pre_proc/pdf_pre_filter.py

-from libs.commons import fitz
-from libs.boxbase import _is_in, _is_in_or_part_overlap
-from libs.drop_reason import DropReason
+from pdf_tools.libs.commons import fitz
+from pdf_tools.libs.boxbase import _is_in, _is_in_or_part_overlap
+from pdf_tools.libs.drop_reason import DropReason
 
 
 def __area(box):

pre_proc/remove_colored_strip_bbox.py → pdf_tools/pre_proc/remove_colored_strip_bbox.py

-from libs.boxbase import _is_in, _is_in_or_part_overlap, calculate_overlap_area_2_minbox_area_ratio
+from pdf_tools.libs.boxbase import _is_in, _is_in_or_part_overlap, calculate_overlap_area_2_minbox_area_ratio
 from loguru import logger
 
-from libs.drop_tag import COLOR_BG_HEADER_TXT_BLOCK
+from pdf_tools.libs.drop_tag import COLOR_BG_HEADER_TXT_BLOCK
 
 
 def __area(box):

pre_proc/remove_footer_header.py → pdf_tools/pre_proc/remove_footer_header.py

 import re
 
-from libs.boxbase import _is_in_or_part_overlap
+from pdf_tools.libs.boxbase import _is_in_or_part_overlap
 
 
 def remove_headder_footer_one_page(text_raw_blocks, image_bboxes, table_bboxes, header_bboxs, footer_bboxs,

pre_proc/remove_rotate_bbox.py → pdf_tools/pre_proc/remove_rotate_bbox.py

-
-import json
 import math
 
-from libs.boxbase import is_vbox_on_side
+from pdf_tools.libs.boxbase import is_vbox_on_side
 
 
 def detect_non_horizontal_texts(result_dict):
@@ -84,7 +82,7 @@ def detect_non_horizontal_texts(result_dict):
 1. 当一个block里全部文字都不是dir=(1,0)，这个block整体去掉
 2. 当一个block里全部文字都是dir=(1,0)，但是每行只有一个字，这个block整体去掉。这个block必须出现在页面的四周，否则不去掉
 """
-import string, re
+import re
 
 def __is_a_word(sentence):
     # 如果输入是中文并且长度为1，则返回True

pre_proc/resolve_bbox_conflict.py → pdf_tools/pre_proc/resolve_bbox_conflict.py

@@ -5,7 +5,7 @@
 2. 然后去掉出现在文字blcok上的图片bbox
 """
 
-from libs.boxbase import _is_in, _is_in_or_part_overlap, _is_left_overlap, calculate_iou, calculate_overlap_area_2_minbox_area_ratio
+from pdf_tools.libs.boxbase import _is_in, _is_in_or_part_overlap, _is_left_overlap
 
 
 def resolve_bbox_overlap_conflict(images:list, tables:list, interline_equations:list, inline_equations:list, text_raw_blocks:list):