include_dirs.append(d)
return include_dirs
ext_modules = [
Extension('dragnet._weninger',
sources=["dragnet/_weninger.pyx"],
include_dirs = [get_include()],
language="c++"),
Extension('dragnet.lcs',
sources=["dragnet/lcs.pyx"],
include_dirs = [get_include()],
language="c++"),
Extension('dragnet.blocks',
sources=["dragnet/blocks.pyx"],
include_dirs = lxml.get_include() + find_libxml2_include(),
language="c++",
libraries=['xml2']),
Extension('dragnet.readability',
sources=["dragnet/readability.pyx"],
include_dirs = [get_include()],
extra_compile_args=['-std=c++0x'],
language="c++"),
]
setup(
name = 'dragnet',
version = '1.0.1',
description = 'Extract the main article content (and optionally comments) from a web page',
author = 'Matt Peters, Dan Lecocq',
include_dirs.append(d)
return include_dirs
ext_modules = [
Extension('dragnet._weninger',
sources=["dragnet/_weninger.pyx"],
include_dirs=[get_include()],
language="c++"),
Extension('dragnet.lcs',
sources=["dragnet/lcs.pyx"],
include_dirs=[get_include()],
language="c++"),
Extension('dragnet.blocks',
sources=["dragnet/blocks.pyx"],
include_dirs=lxml.get_include() + find_libxml2_include(),
language="c++",
libraries=['xml2']),
Extension('dragnet.readability',
sources=["dragnet/readability.pyx"],
include_dirs=[get_include()],
extra_compile_args=['-std=c++0x'],
language="c++"),
]
setup(
name='dragnet',
version='1.0.1',
description=
'Extract the main article content (and optionally comments) from a web page',
author='Matt Peters, Dan Lecocq',
setup(console=['D:\Users\u156726\PycharmProjects\SFR_APP\SFR_APP_BK.py'],options={
'py2exe': {
'packages' : ['matplotlib', 'pytz'],
}
},
data_files=matplotlib.get_py2exe_datafiles(),)
"""
includes = ['matplotlib','datetime','lxml.etree', 'lxml._elementpath', 'gzip','pygal','matplotlib.pyplot','ttk','lxml', 'numpy', 'Tkinter', 'Tkconstants', 'pygal','Tix','matplotlib.backends.backend_qt4agg','matplotlib.ticker','PyQt4','datetime','tkFileDialog','ttk','threading' ,'csv','re','numpy','os','tkinter.tix']
excludes = ['_gtkagg', '_tkagg', 'curses', 'pywin.debugger', 'pywin.debugger.dbgcon', 'pywin.dialogs' ]
packages = []
dll_excludes = []
import Tix
import lxml
lxml.get_include()
def files(folder):
for path in glob.glob(folder+'/*'):
if os.path.isfile(path):
yield path
class Target:
def __init__(self, **kw):
self.__dict__.update(kw)
self.version = "1.0.0.0"
self.company_name = "SFR"
self.copyright = "Copyright (c) 2017 KAID Belkacem."
self.name = "SFR_Perf_DataAnalyser"
import hashlib
from uuid import uuid4
import lxml
lxml.get_include()
import requests
from lxml import etree
# Create your tests here.
def getBlogList():
con = requests.get('https://blog.csdn.net/p571912102')
# print(len(con.content.decode('utf-8')))
content = con.content.decode('utf-8')
html_obj = etree.HTML(content)
html_str = etree.tostring(html_obj, pretty_print=True)
blog_obj_list = html_obj.xpath('//div[@data-articleid]')[1:]
for i in blog_obj_list:
blog_tag_obj = i.xpath('h4/a/span')[0]
blog_data_id = int(i.xpath('@data-articleid')[0])
print('id--->', blog_data_id)
blog_content_obj = i.xpath('p/a')[0].text
blog_title_obj = i.xpath('h4/a/text()')[1].strip()
blog_time_obj_list = i.xpath(
'//div[@data-articleid][2]/div/p/span') # 发布时间 阅读 评论
for i in blog_time_obj_list:
print('---time--', i.text)
detil_url = 'https://blog.csdn.net/p571912102/article/details/{}'.format(
blog_data_id)
def include_dirs():
import lxml
dirs = xmlsec_config()['include_dirs']
dirs.extend(lxml.get_include())
return dirs
def run(self):
ext = self.ext_map['xmlsec']
self.debug = os.environ.get('PYXMLSEC_ENABLE_DEBUG', False)
self.static = os.environ.get('PYXMLSEC_STATIC_DEPS', False)
if self.static or sys.platform == 'win32':
self.info('starting static build on {}'.format(sys.platform))
buildroot = Path('build', 'tmp')
self.prefix_dir = buildroot / 'prefix'
self.prefix_dir.mkdir(parents=True, exist_ok=True)
self.prefix_dir = self.prefix_dir.absolute()
self.build_libs_dir = buildroot / 'libs'
self.build_libs_dir.mkdir(exist_ok=True)
self.libs_dir = Path(os.environ.get('PYXMLSEC_LIBS_DIR', 'libs'))
self.libs_dir.mkdir(exist_ok=True)
if sys.platform == 'win32':
self.prepare_static_build_win()
elif 'linux' in sys.platform:
self.prepare_static_build_linux()
else:
import pkgconfig
try:
config = pkgconfig.parse('xmlsec1')
except EnvironmentError:
raise DistutilsError('Unable to invoke pkg-config.')
except pkgconfig.PackageNotFoundError:
raise DistutilsError('xmlsec1 is not installed or not in path.')
if config is None or not config.get('libraries'):
raise DistutilsError('Bad or incomplete result returned from pkg-config.')
ext.define_macros.extend(config['define_macros'])
ext.include_dirs.extend(config['include_dirs'])
ext.library_dirs.extend(config['library_dirs'])
ext.libraries.extend(config['libraries'])
import lxml
ext.include_dirs.extend(lxml.get_include())
ext.define_macros.extend(
[('MODULE_NAME', self.distribution.metadata.name), ('MODULE_VERSION', self.distribution.metadata.version)]
)
# escape the XMLSEC_CRYPTO macro value, see mehcode/python-xmlsec#141
for (key, value) in ext.define_macros:
if key == 'XMLSEC_CRYPTO' and not (value.startswith('"') and value.endswith('"')):
ext.define_macros.remove((key, value))
ext.define_macros.append((key, '"{0}"'.format(value)))
break
if sys.platform == 'win32':
ext.extra_compile_args.append('/Zi')
else:
ext.extra_compile_args.extend(
[
'-g',
'-std=c99',
'-fPIC',
'-fno-strict-aliasing',
'-Wno-error=declaration-after-statement',
'-Werror=implicit-function-declaration',
]
)
if self.debug:
ext.extra_compile_args.append('-Wall')
ext.extra_compile_args.append('-O0')
ext.define_macros.append(('PYXMLSEC_ENABLE_DEBUG', '1'))
else:
ext.extra_compile_args.append('-Os')
super(build_ext, self).run()
from setuptools import setup, Extension
from setuptools.command.build_ext import build_ext as setuptools_build_ext
from Cython.Build.Cythonize import cythonize
import lxml
MODE = 'release'
class build_ext(setuptools_build_ext):
def build_extension(self, ext):
subprocess.check_call(['cargo', 'build'] +
(['--release'] if MODE == 'release' else []))
setuptools_build_ext.build_extension(self, ext)
includes = ['/usr/include/libxml2'] + lxml.get_include()
setup(
name='htmlpyever',
ext_modules=cythonize([
Extension(
name='htmlpyever',
sources=['htmlpyever.pyx'],
libraries=['html5ever_glue', 'xml2'],
library_dirs=['target/{}'.format(MODE)],
include_dirs=includes,
depends=['target/{}/libhtml5ever_glue.a'.format(MODE)],
)
],
include_path=includes),
setup_requires=['cython'],
install_requires=['lxml'],