目录重构

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

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

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

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

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

 import re
 
-from pdf_tools.libs.boxbase import _is_in_or_part_overlap
+from magic_pdf.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,

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

 import math
 
-from pdf_tools.libs.boxbase import is_vbox_on_side
+from magic_pdf.libs.boxbase import is_vbox_on_side
 
 
 def detect_non_horizontal_texts(result_dict):

pdf_tools/pre_proc/resolve_bbox_conflict.py → magic_pdf/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 pdf_tools.libs.boxbase import _is_in, _is_in_or_part_overlap, _is_left_overlap
+from magic_pdf.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):

othoers/check_inline_formula.py

 # 最终版：把那种text_block有重叠，且inline_formula位置在重叠部分的，认定整个页面都有问题，所有的inline_formula都改成no_check
-from pdf_tools.libs import fitz
+from magic_pdf.libs import fitz
 
 
 def check_inline_formula(page, inline_formula_boxes):

othoers/pdf2json_infer.py

@@ -3,7 +3,7 @@ from typing import Tuple
 import os
 import boto3, json
 from botocore.config import Config
-from pdf_tools.libs import fitz
+from magic_pdf.libs import fitz
 from loguru import logger
 from pathlib import Path
 from tqdm import tqdm
@@ -22,13 +22,13 @@ from validation import cal_edit_distance, format_gt_bbox, label_match, detect_va
 # from pdf2text_recogPara import parse_blocks_per_page    
 # from bbox_sort import bbox_sort, CONTENT_IDX, CONTENT_TYPE_IDX
 
-from pdf_tools.layout.bbox_sort import bbox_sort, CONTENT_IDX, CONTENT_TYPE_IDX
-from pdf_tools.pre_proc import parse_images          # 获取figures的bbox
-from pdf_tools.pre_proc.detect_tables import parse_tables           # 获取tables的bbox
-from pdf_tools.pre_proc import parse_equations     # 获取equations的bbox
+from magic_pdf.layout.bbox_sort import bbox_sort, CONTENT_IDX, CONTENT_TYPE_IDX
+from magic_pdf.pre_proc import parse_images          # 获取figures的bbox
+from magic_pdf.pre_proc.detect_tables import parse_tables           # 获取tables的bbox
+from magic_pdf.pre_proc import parse_equations     # 获取equations的bbox
 # from pdf2text_recogFootnote import parse_footnotes     # 获取footnotes的bbox
-from pdf_tools.post_proc.detect_para import process_blocks_per_page
-from pdf_tools.libs import parse_aws_param, parse_bucket_key, read_file, join_path
+from magic_pdf.post_proc.detect_para import process_blocks_per_page
+from magic_pdf.libs import parse_aws_param, parse_bucket_key, read_file, join_path
 
 
 def cut_image(bbox: Tuple, page_num: int, page: fitz.Page, save_parent_path: str, s3_profile: str):

othoers/pdf2text_evaluatePdfLayout.py

-from pdf_tools.libs import fitz             # pyMuPDF库
+from magic_pdf.libs import fitz             # pyMuPDF库
 
 
 def calculate_overlapRatio_between_rect1_and_rect2(L1: float, U1: float, R1: float, D1: float, L2: float, U2: float, R2: float, D2: float) -> (float, float):

othoers/pdf2text_getNumberOfColumn.py

-from pdf_tools.libs import fitz
+from magic_pdf.libs import fitz
 from typing import List
 
 

othoers/pdf2text_recogFootnoteLine.py

 import re
-from pdf_tools.libs import _is_in_or_part_overlap
-from pdf_tools.libs import fitz
+from magic_pdf.libs import _is_in_or_part_overlap
+from magic_pdf.libs import fitz
 import collections
 
 

othoers/pdf2text_recogPara_v2.py

@@ -11,8 +11,8 @@ import numpy as np
 from termcolor import cprint
 
 
-from pdf_tools.libs import fitz
-from pdf_tools.libs import NLPModels
+from magic_pdf.libs import fitz
+from magic_pdf.libs import NLPModels
 
 
 if sys.version_info[0] >= 3:

othoers/pdf2text_recogTitle.py

-from pdf_tools.libs.commons import fitz             # pyMuPDF库
+from magic_pdf.libs.commons import fitz             # pyMuPDF库
 
 
 def parse_titles(page_ID: int, page: fitz.Page, json_from_DocXchain_obj: dict, exclude_bboxes):

othoers/vali_bbox_sort.py

@@ -2,7 +2,7 @@ import numpy as np
 import tqdm
 import json
 from validation import cal_edit_distance, format_gt_bbox
-from pdf_tools.layout.layout_sort import sort_with_layout
+from magic_pdf.layout.layout_sort import sort_with_layout
 
 with open('/mnt/petrelfs/share_data/ouyanglinke/OCR/OCR_validation_dataset_final_rotated_formulafix_highdpi_scihub.json', 'r') as f:
     samples = json.load(f)

tests/test_commons.py

 import io
 import json
 import os
-from pdf_tools.libs import fitz
+from magic_pdf.libs import fitz
 
 from app.common.s3 import get_s3_config, get_s3_client
-from pdf_tools.libs import join_path, json_dump_path, read_file, parse_bucket_key
+from magic_pdf.libs import join_path, json_dump_path, read_file, parse_bucket_key
 from loguru import logger
 
 test_pdf_dir_path = "s3://llm-pdf-text/unittest/pdf/"

tests/test_metascan_classify/test_classify.py

@@ -2,9 +2,9 @@ import os
 
 import pytest
 
-from pdf_tools.filter import classify_by_area, classify_by_text_len, classify_by_avg_words, \
+from magic_pdf.filter import classify_by_area, classify_by_text_len, classify_by_avg_words, \
     classify_by_img_num, classify_by_text_layout, classify_by_img_narrow_strips
-from pdf_tools.filter.pdf_meta_scan import get_pdf_page_size_pts, get_pdf_textlen_per_page, get_imgs_per_page
+from magic_pdf.filter.pdf_meta_scan import get_pdf_page_size_pts, get_pdf_textlen_per_page, get_imgs_per_page
 from test.test_commons import get_docs_from_test_pdf, get_test_json_data
 
 # 获取当前目录