重构目录结构

layout/layout_det_utils.py → pdf_tools/layout/layout_det_utils.py

-from layout.bbox_sort import X0_EXT_IDX, X0_IDX, X1_EXT_IDX, X1_IDX, Y0_EXT_IDX, Y0_IDX, Y1_EXT_IDX, Y1_IDX
-from libs.boxbase import _is_bottom_full_overlap, _left_intersect, _right_intersect
+from pdf_tools.layout.bbox_sort import X0_EXT_IDX, X0_IDX, X1_EXT_IDX, X1_IDX, Y0_IDX, Y1_EXT_IDX, Y1_IDX
+from pdf_tools.libs.boxbase import _is_bottom_full_overlap, _left_intersect, _right_intersect
 
 
 def find_all_left_bbox_direct(this_bbox, all_bboxes) -> list:

layout/layout_sort.py → pdf_tools/layout/layout_sort.py

@@ -2,11 +2,10 @@
 对pdf上的box进行layout识别，并对内部组成的box进行排序
 """
 
-import json
 from loguru import logger
-from layout.bbox_sort import CONTENT_IDX, CONTENT_TYPE_IDX, X0_EXT_IDX, X0_IDX, X1_EXT_IDX, X1_IDX, Y0_EXT_IDX, Y0_IDX, Y1_EXT_IDX, Y1_IDX, paper_bbox_sort
-from layout.layout_det_utils import find_all_left_bbox_direct, find_all_right_bbox_direct, find_bottom_bbox_direct_from_left_edge, find_bottom_bbox_direct_from_right_edge, find_top_bbox_direct_from_left_edge, find_top_bbox_direct_from_right_edge, find_all_top_bbox_direct, find_all_bottom_bbox_direct, get_left_edge_bboxes, get_right_edge_bboxes
-from libs.boxbase import get_bbox_in_boundry
+from pdf_tools.layout.bbox_sort import CONTENT_IDX, CONTENT_TYPE_IDX, X0_EXT_IDX, X0_IDX, X1_EXT_IDX, X1_IDX, Y0_EXT_IDX, Y0_IDX, Y1_EXT_IDX, Y1_IDX, paper_bbox_sort
+from pdf_tools.layout.layout_det_utils import find_all_left_bbox_direct, find_all_right_bbox_direct, find_bottom_bbox_direct_from_left_edge, find_bottom_bbox_direct_from_right_edge, find_top_bbox_direct_from_left_edge, find_top_bbox_direct_from_right_edge, find_all_top_bbox_direct, find_all_bottom_bbox_direct, get_left_edge_bboxes, get_right_edge_bboxes
+from pdf_tools.libs.boxbase import get_bbox_in_boundry
 
 
 LAYOUT_V = "V"

layout/layout_spiler_recog.py → pdf_tools/layout/layout_spiler_recog.py

@@ -3,8 +3,8 @@
 """
 
 import os
-from libs.commons import fitz
-from libs.boxbase import _is_in_or_part_overlap
+from pdf_tools.libs.commons import fitz
+from pdf_tools.libs.boxbase import _is_in_or_part_overlap
 
 
 def __rect_filter_by_width(rect, page_w, page_h):

layout/mcol_sort.py → pdf_tools/layout/mcol_sort.py

@@ -49,9 +49,8 @@ Usage
       print(page.get_text(clip=rect, sort=True))
   ----------------------------------------------------------------------------------
 """
-import os
 import sys
-from libs.commons import fitz
+from pdf_tools.libs.commons import fitz
 
 
 def column_boxes(page, footer_margin=50, header_margin=50, no_image_text=True):

libs/nlp_utils.py → pdf_tools/libs/nlp_utils.py

@@ -10,7 +10,7 @@ import spacy
 import en_core_web_sm
 import zh_core_web_sm
 
-from libs.language import detect_lang
+from pdf_tools.libs.language import detect_lang
 
 
 class NLPModels:

libs/pdf_image_tools.py → pdf_tools/libs/pdf_image_tools.py

@@ -4,9 +4,9 @@ from typing import Tuple
 import io
 
 # from app.common.s3 import get_s3_client
-from libs.commons import fitz
+from pdf_tools.libs.commons import fitz
 from loguru import logger
-from libs.commons import parse_bucket_key, join_path
+from pdf_tools.libs.commons import parse_bucket_key, join_path
 
 
 def cut_image(bbox: Tuple, page_num: int, page: fitz.Page, save_parent_path: str, s3_return_path=None, img_s3_client=None, upload_switch=True):

libs/vis_utils.py → pdf_tools/libs/vis_utils.py

-from libs.commons import fitz
+from pdf_tools.libs.commons import fitz
 import os
-from loguru import logger
-from layout.bbox_sort import CONTENT_TYPE_IDX
 
 
 def draw_bbox_on_page(raw_pdf_doc: fitz.Document, paras_dict:dict, save_path: str):

para/block_continuation_processor.py → pdf_tools/para/block_continuation_processor.py

 import os
-import sys
 import unicodedata
 
-from para.commons import *
+from pdf_tools.para.commons import *
 
 
 if sys.version_info[0] >= 3: