mirror of
https://github.com/opendatalab/MinerU.git
synced 2026-09-24 23:10:23 +08:00
add version_name to middle json
This commit is contained in:
@@ -13,19 +13,6 @@ import fitz # 1.23.9中已经切换到rebase
|
||||
# import fitz_old as fitz # 使用1.23.9之前的pymupdf库
|
||||
|
||||
|
||||
def get_version():
|
||||
command = ["git", "describe", "--tags"]
|
||||
try:
|
||||
version = subprocess.check_output(command).decode().strip()
|
||||
version_parts = version.split("-")
|
||||
if len(version_parts) > 1 and version_parts[0].startswith("magic_pdf"):
|
||||
return version_parts[1]
|
||||
else:
|
||||
raise ValueError(f"Invalid version tag {version}. Expected format is magic_pdf-<version>-released.")
|
||||
except Exception as e:
|
||||
print(e)
|
||||
return "0.0.0"
|
||||
|
||||
def get_delta_time(input_time):
|
||||
return round(time.time() - input_time, 2)
|
||||
|
||||
|
||||
@@ -0,0 +1 @@
|
||||
__version__ = "0.0.0"
|
||||
@@ -16,7 +16,7 @@ import re
|
||||
|
||||
from loguru import logger
|
||||
|
||||
from magic_pdf.libs.commons import get_version
|
||||
from magic_pdf.libs.version import __version__
|
||||
from magic_pdf.rw import AbsReaderWriter
|
||||
from magic_pdf.pdf_parse_by_ocr_v2 import parse_pdf_by_ocr
|
||||
from magic_pdf.pdf_parse_by_txt_v2 import parse_pdf_by_txt
|
||||
@@ -39,7 +39,7 @@ def parse_txt_pdf(pdf_bytes: bytes, pdf_models: list, imageWriter: AbsReaderWrit
|
||||
|
||||
pdf_info_dict["_parse_type"] = PARSE_TYPE_TXT
|
||||
|
||||
pdf_info_dict["_version_name"] = get_version()
|
||||
pdf_info_dict["_version_name"] = __version__
|
||||
|
||||
return pdf_info_dict
|
||||
|
||||
@@ -59,7 +59,7 @@ def parse_ocr_pdf(pdf_bytes: bytes, pdf_models: list, imageWriter: AbsReaderWrit
|
||||
|
||||
pdf_info_dict["_parse_type"] = PARSE_TYPE_OCR
|
||||
|
||||
pdf_info_dict["_version_name"] = get_version()
|
||||
pdf_info_dict["_version_name"] = __version__
|
||||
|
||||
return pdf_info_dict
|
||||
|
||||
@@ -122,6 +122,6 @@ def parse_union_pdf(pdf_bytes: bytes, pdf_models: list, imageWriter: AbsReaderWr
|
||||
else:
|
||||
pdf_info_dict["_parse_type"] = PARSE_TYPE_TXT
|
||||
|
||||
pdf_info_dict["_version_name"] = get_version()
|
||||
pdf_info_dict["_version_name"] = __version__
|
||||
|
||||
return pdf_info_dict
|
||||
|
||||
@@ -1,6 +1,7 @@
|
||||
from setuptools import setup, find_packages
|
||||
import os
|
||||
import subprocess
|
||||
|
||||
from magic_pdf.libs.commons import get_version
|
||||
from setuptools import setup, find_packages
|
||||
|
||||
|
||||
def parse_requirements(filename):
|
||||
@@ -19,14 +20,36 @@ def parse_requirements(filename):
|
||||
return requires
|
||||
|
||||
|
||||
def get_version():
|
||||
command = ["git", "describe", "--tags"]
|
||||
try:
|
||||
version = subprocess.check_output(command).decode().strip()
|
||||
version_parts = version.split("-")
|
||||
if len(version_parts) > 1 and version_parts[0].startswith("magic_pdf"):
|
||||
return version_parts[1]
|
||||
else:
|
||||
raise ValueError(f"Invalid version tag {version}. Expected format is magic_pdf-<version>-released.")
|
||||
except Exception as e:
|
||||
print(e)
|
||||
return "0.0.0"
|
||||
|
||||
setup(
|
||||
name="magic_pdf", # 项目名
|
||||
version=get_version(), # 自动从tag中获取版本号
|
||||
packages=find_packages(), # 包含所有的包
|
||||
install_requires=parse_requirements('requirements.txt'), # 项目依赖的第三方库
|
||||
python_requires=">=3.9", # 项目依赖的 Python 版本
|
||||
# entry_points={"console_scripts": ["my_command=my_project.main:run"]}, # 项目提供的可执行命令
|
||||
include_package_data=True, # 是否包含非代码文件,如数据文件、配置文件等
|
||||
zip_safe=False, # 是否使用 zip 文件格式打包,一般设为 False
|
||||
)
|
||||
|
||||
def write_version_to_commons(version):
|
||||
commons_path = os.path.join(os.path.dirname(__file__), 'magic_pdf', 'libs', 'version.py')
|
||||
with open(commons_path, 'w') as f:
|
||||
f.write(f'__version__ = "{version}"\n')
|
||||
|
||||
|
||||
if __name__ == '__main__':
|
||||
version_name = get_version()
|
||||
write_version_to_commons(version_name)
|
||||
setup(
|
||||
name="magic_pdf", # 项目名
|
||||
version=version_name, # 自动从tag中获取版本号
|
||||
packages=find_packages(), # 包含所有的包
|
||||
install_requires=parse_requirements('requirements.txt'), # 项目依赖的第三方库
|
||||
python_requires=">=3.9", # 项目依赖的 Python 版本
|
||||
# entry_points={"console_scripts": ["my_command=my_project.main:run"]}, # 项目提供的可执行命令
|
||||
include_package_data=True, # 是否包含非代码文件,如数据文件、配置文件等
|
||||
zip_safe=False, # 是否使用 zip 文件格式打包,一般设为 False
|
||||
)
|
||||
|
||||
Reference in New Issue
Block a user