Merge pull request #6 from HuangPuStar/feat/version-upgrade

fix: 统一使用相对导入rag_tokenizer，确保所有文件都能正确访问rag_tokenizer对象

Author: sqhyz55

deepdoc/parser/docx_parser.py

@@ -18,7 +18,7 @@
 import re
 import pandas as pd
 from collections import Counter
-from deepdoc.depend import rag_tokenizer
+from ..depend import rag_tokenizer
 from io import BytesIO
 
 

deepdoc/parser/html_parser.py

@@ -16,7 +16,7 @@
 #
 
 from ..depend.find_codec import find_codec
-from deepdoc.depend import rag_tokenizer
+from ..depend import rag_tokenizer
 import uuid
 import chardet
 from bs4 import BeautifulSoup, NavigableString, Tag, Comment

deepdoc/parser/pdf_parser.py

@@ -39,7 +39,7 @@
 from ..common.file_utils import get_project_base_directory
 from ..common.misc_utils import pip_install_torch
 from deepdoc.vision import OCR, AscendLayoutRecognizer, LayoutRecognizer, Recognizer, TableStructureRecognizer
-from deepdoc.depend import rag_tokenizer
+from ..depend import rag_tokenizer
 from ..depend.prompts import vision_llm_describe_prompt
 from ..common import settings
 

deepdoc/parser/resume/step_two.py

@@ -24,7 +24,7 @@
 import signal
 import numpy as np
 from .entities import degrees, schools, corporations
-from ...depend.rag_tokenizer import tokenize, fine_grained_tokenize, tag, freq, loadUserDict, addUserDict, tradi2simp, strQ2B
+from ...depend import rag_tokenizer
 from ...depend.surname import surname
 from xpinyin import Pinyin
 from contextlib import contextmanager

deepdoc/vision/table_structure_recognizer.py

@@ -22,7 +22,7 @@
 from huggingface_hub import snapshot_download
 
 from ..common.file_utils import get_project_base_directory
-from ..depend.rag_tokenizer import tokenize
+from ..depend import rag_tokenizer
 
 from .recognizer import Recognizer