def fetch_html(self, url, encoding='', errors='strict'):
if not errors:
errors = 'strict'
bytes_data, bytes_encoding = self.fetch_bytes_encoding(url)
# get encoding
if not encoding:
# server send
encoding = bytes_encoding
if not encoding:
# chardect
encoding = Fetcher.d(url, bytes_data)
if not encoding:
# default: utf-8
print('fetcher:打开%s时，使用utf-8作默认' % url)
encoding = 'utf-8'
# decode
try:
return bytes_data.decode(encoding, errors)
except:
print('下载器<解文本编码>失败')
s = '字节长度:%d.使用编码:%s.网址:%s' % \
(len(bytes_data), encoding, url)
raise c_worker_exception('解文本编码(decode)时出现异常', url, s)
def fetch_html(self, url, encoding='', errors='strict'):
if not errors:
errors = 'strict'
bytes_data, bytes_encoding = self.fetch_bytes_encoding(url)
# get encoding
if not encoding:
# server send
encoding = bytes_encoding
if not encoding:
# chardect
encoding = Fetcher.d(url, bytes_data)
if not encoding:
# default: utf-8
print('fetcher:打开%s时，使用utf-8作默认' % url)
encoding = 'utf-8'
# decode
try:
return bytes_data.decode(encoding, errors)
except:
print('下载器<解文本编码>失败')
s = '字节长度:%d.使用编码:%s.网址:%s' % \
(len(bytes_data), encoding, url)
raise c_worker_exception('解文本编码(decode)时出现异常', url, s)
def pattern_error(blocknum, isblock=True):
if isblock:
s = '第%d个block的blockre编译失败' % (blocknum + 1)
else:
s = '第%d个block的itemre编译失败' % (blocknum + 1)
raise c_worker_exception('正则表达式编译失败',
'',
s)
def do_process(data_dict, worker_dict):
package = data_dict['package']
url = 'https://pypi.python.org/pypi/%s/' % package
f = Fetcher()
string = f.fetch_html(url)
if not string:
raise c_worker_exception('无法下载url', url)
# single page
single_re = (r'<span class="breadcrumb-separator">.*?'
r'<span class="breadcrumb-separator">.*?'
r'<a href="(/pypi/([^/]+)/([^"]+))">.*?'
r'class="odd".*?'
r'<td>(\d{4}-\d{1,2}-\d{1,2})</td>'
)
prog = red.d(single_re, red.DOTALL)
m = prog.search(string)
if m:
info = c_info()
info.title = m.group(2) + ' ' + m.group(3)
info.url = url
info.pub_date = m.group(4)
info.suid = info.title
return [info]
# table page
table_re = (r'<tr class="(?:odd|even)">.*?'
r'<a href="(/pypi/([^/]+)/([^"]+))">'
r'')
prog = red.d(table_re, red.DOTALL)
miter = prog.finditer(string)
lst = list()
for m in miter:
info = c_info()
info.title = m.group(2) + ' ' + m.group(3)
info.url = 'https://pypi.python.org' + m.group(1)
info.suid = info.title
lst.append(info)
return lst
def parse_xml(data_dict, xml):
if not xml:
raise c_worker_exception('xml为空字符串', data_dict['url'], '')
r = red.d(r'^\s*$')
if r.match(xml) is not None:
raise c_worker_exception('xml只有空白', data_dict['url'], '')
# remove namespace of atom
xml = red.sub(r'''<feed\s+(?:"[^"]*"|'[^']*'|[^"'>])*>''',
r'<feed>',
xml,
count=1,
flags=red.IGNORECASE | red.A)
# ElementTree
try:
doc = ET.fromstring(xml)
except Exception as e:
doc = None
if not etree:
raise c_worker_exception('解析XML失败，可以尝试安装lxml模块', data_dict['url'],
str(e))
# lxml
if doc is None:
try:
parser = etree.XMLParser(recover=True, encoding='utf-8')
doc = etree.fromstring(xml.encode('utf-8'), parser=parser)
print('使用lxml解析%s' % data_dict['url'])
if doc is None:
raise Exception('lxml模块也无法解析此XML')
except Exception as e:
raise c_worker_exception('lxml解析XML失败', data_dict['url'], str(e))
# get type of the feed
if doc.tag == 'rss' and doc.get('version', '') == '1.0':
feedtype = 0
elif doc.tag == 'rss' and doc.get('version', '') == '2.0':
feedtype = 1
elif doc.tag == 'feed':
feedtype = 2
else:
raise c_worker_exception('无法识别XML的feed类型', data_dict['url'], '')
# get feed author
if 'use_feed_author' in data_dict:
f_author = de_html_char(doc.findtext(tagnames['f_author'][feedtype]))
item_iter = doc.findall(tagnames['f_items'][feedtype])
if item_iter is None:
return []
ret = []
for item in item_iter:
# ------- info -------
one = c_info()
# title
one.title = de_html_char(item.findtext(tagnames['title'][feedtype]))
# url
if feedtype < 2:
url = de_html_char(item.findtext(tagnames['url'][feedtype]))
else:
url = ''
link_iter = item.findall('link')
if link_iter is not None:
for tag_link in link_iter:
if tag_link.get('rel') == 'alternate':
if tag_link.get('type') == 'text/html':
url = tag_link.get('href')
break
url = url or tag_link.get('href')
one.url = url
# author, summary, pub_date
if 'use_feed_author' in data_dict:
one.author = de_html_char(
item.findtext(tagnames['author'][feedtype], f_author))
one.summary = de_html_char(item.findtext(
tagnames['summary'][feedtype]))
one.pub_date = de_html_char(
item.findtext(tagnames['pub_date'][feedtype]))
# suid
guid = item.findtext(tagnames['guid'][feedtype])
one.suid = guid or one.url or one.title
ret.append(one)
return ret
def fetch_bytes_encoding(self, url):
'''return (byte data, encoding)'''
def get_encoding(r):
contenttype = r.getheader('Content-Type', '')
if not contenttype:
return ''
matcher = re_contenttype.search(contenttype)
if matcher:
return Fetcher.lookup_encoding(matcher.group(1))
else:
return ''
# --------------主体开始-------------
# request对象
req = urllib.request.Request(url)
req.add_header('User-Agent', self.info.ua)
e = None
# 重试用的循环
for i in range(self.info.retry_count):
try:
# r是HTTPResponse对象
r = self.opener.open(req,
timeout=self.info.open_timeout
)
ret_data = r.read()
encoding = get_encoding(r)
# decompress
contentenc = r.getheader('Content-Encoding', '')
if contentenc:
contentenc = contentenc.lower()
if 'gzip' in contentenc:
ret_data = gzip.decompress(ret_data)
elif 'deflate' in contentenc:
try:
# first try: zlib
ret_data = zlib.decompress(ret_data, 15)
except:
# second try: raw deflate
ret_data = zlib.decompress(ret_data, -15)
# get encoding from bytes content
if not encoding:
matcher = re_meta.search(ret_data)
if matcher:
try:
extract = matcher.group(1).decode('ascii')
except:
encoding = ''
else:
encoding = Fetcher.lookup_encoding(extract)
return ret_data, encoding
except Exception as ee:
e = ee
if i < self.info.retry_count - 1:
time.sleep(self.info.retry_interval)
else:
print('fetcher:异常,下载%s失败' % url, '\n异常信息:', e)
s = '%s (下载%s失败，重试了%d次，连接超时限制%d秒)' % \
(str(e), url, self.info.retry_count, self.info.open_timeout)
raise c_worker_exception('下载url失败', url, s)
def parse_html(data_dict, base_url, html):
if not html:
raise c_worker_exception('html为空字符串', data_dict['url'], '')
r = red.d(r'^\s*$')
if r.match(html) is not None:
raise c_worker_exception('html只有空白', data_dict['url'], '')
# extract json string
re = red.d(data_dict['re_pattern'], data_dict['re_flags'])
if re is None:
raise c_worker_exception('正则表达式编译失败', '', '用于提取json字符串的正则表达式编译失败')
m = re.search(html)
if m is None:
raise c_worker_exception('无法用re(正则表达式)提取json字符', data_dict['url'], '')
json_str = m.group(1)
# replace
if 'repl' in data_dict:
r = red.d(data_dict['repl_pattern'], data_dict['repl_flags'])
if r is None:
raise c_worker_exception('replace正则表达式编译失败')
json_str = r.sub(data_dict['repl'], json_str)
# parse json
try:
json_obj = json.loads(json_str)
except Exception as e:
raise c_worker_exception('解析json时出错', data_dict['url'], str(e))
# blocks
json_lst = data_dict['blocks_list']
ret = list()
for i, block in enumerate(json_lst):
# travel path
path = block[0]
block_j = json_obj
# path必定为tuple
for ii, path_item in enumerate(path):
try:
block_j = block_j[path_item]
except:
s = '第%d个block, block_path的第%d个路径元素%s无效'
raise c_worker_exception(
s % (i + 1, ii + 1, str(path_item)), data_dict['url'],
'path:%s 可能是网站改变了json的设计结构' % str(path))
# extract
if type(block_j) == list:
pass
elif type(block_j) == dict:
block_j = block_j.values()
else:
s = '第%d个block, block_path找到的不是列表或字典'
raise c_worker_exception(s % (i + 1), data_dict['url'],
'path:%s 可能是网站改变了json的设计结构' % str(path))
for block_item_j in block_j:
info = c_info()
for key, sub_path in block[1].items():
temp_jj = block_item_j
for sub_path_item in sub_path:
try:
temp_jj = temp_jj[sub_path_item]
except Exception as e:
print('异常：', e)
s1 = '处理第%d个block的映射时异常' % (i + 1)
s2 = 'path:%s,key:%s,map:%s,无法找到指定元素%s.' % \
(str(path), key, str(sub_path),
str(sub_path_item))
raise c_worker_exception(s1, '', s2)
ss = item_process(temp_jj)
if key == 'title':
info.title = ss
elif key == 'url':
info.url = ss
elif key == 'summary':
info.summary = ss
elif key == 'author':
info.author = ss
elif key == 'pub_date':
info.pub_date = ss
elif key == 'urljoin':
info.url = urljoin(base_url, ss)
elif key == 'suid':
info.suid = ss
elif key == 'temp':
info.temp = ss
else:
print('无法处理map_rule', key, sub_path)
if not info.suid:
info.suid = info.url
ret.append(info)
return ret
def fetch_bytes_encoding(self, url):
'''return (byte data, encoding)'''
def get_encoding(r):
contenttype = r.getheader('Content-Type', '')
if not contenttype:
return ''
matcher = re_contenttype.search(contenttype)
if matcher:
return Fetcher.lookup_encoding(matcher.group(1))
else:
return ''
# --------------主体开始-------------
# request对象
req = urllib.request.Request(url)
req.add_header('User-Agent', self.info.ua)
e = None
# 重试用的循环
for i in range(self.info.retry_count):
try:
# r是HTTPResponse对象
r = self.opener.open(req,
timeout=self.info.open_timeout
)
ret_data = r.read()
encoding = get_encoding(r)
# decompress
contentenc = r.getheader('Content-Encoding', '')
if contentenc:
contentenc = contentenc.lower()
if 'gzip' in contentenc:
ret_data = gzip.decompress(ret_data)
elif 'deflate' in contentenc:
try:
# first try: zlib
ret_data = zlib.decompress(ret_data, 15)
except:
# second try: raw deflate
ret_data = zlib.decompress(ret_data, -15)
# get encoding from bytes content
if not encoding:
matcher = re_meta.search(ret_data)
if matcher:
try:
extract = matcher.group(1).decode('ascii')
except:
encoding = ''
else:
encoding = Fetcher.lookup_encoding(extract)
return ret_data, encoding
except Exception as ee:
e = ee
if i < self.info.retry_count-1:
time.sleep(self.info.retry_interval)
else:
print('fetcher:异常,下载%s失败' % url, '\n异常信息:', e)
s = '%s (下载%s失败，重试了%d次，连接超时限制%d秒)' % \
(str(e), url, self.info.retry_count, self.info.open_timeout)
raise c_worker_exception('下载url失败', url, s)
def parse_html(data_dict, base_url, html):
if not html:
raise c_worker_exception('html为空字符串', data_dict['url'], '')
r = red.d(r'^\s*$')
if r.match(html) is not None:
raise c_worker_exception('html只有空白', data_dict['url'], '')
re_lst = data_dict['blocks_list']
ret = list()
for i, block in enumerate(re_lst):
# block re
block_prog = red.d(block[0][0], block[0][1])
if block_prog is None:
pattern_error(i)
itr = block_prog.finditer(html)
matches = list(itr)
if len(matches) != 1:
s = '第%d个block的block_re找到的结果为%d，应为1' % \
(i + 1, len(matches))
raise c_worker_exception(s, '',
'可能是网页改版、服务器显示错误信息')
subhtml = matches[0].group(1)
# item re
item_prog = red.d(block[1][0], block[1][1])
if item_prog is None:
pattern_error(i, False)
itr = item_prog.finditer(subhtml)
matches = list(itr)
if not matches:
s = '第%d个block的item_re找到的结果为0，应大于0' % (i + 1)
raise c_worker_exception(s, '', '可能是网页改版')
for m in matches:
info = c_info()
for k, v in block[2].items():
try:
ss = map_attrs(m, v)
except Exception as e:
s1 = '处理第%d个block的map_rule时异常' % (i + 1)
s2 = '赋值%s给%s时出错，%s' % (str(v), str(k), str(e))
raise c_worker_exception(s1, '', s2)
if k == 'title':
info.title = ss
elif k == 'url':
info.url = ss
elif k == 'urljoin':
info.url = urljoin(base_url, ss)
elif k == 'summary':
info.summary = ss
elif k == 'author':
info.author = ss
elif k == 'pub_date':
info.pub_date = ss
elif k == 'suid':
info.suid = ss
elif k == 'temp':
info.temp = ss
else:
print('无法处理map_rule', k, v)
if not info.suid:
info.suid = info.url
ret.append(info)
return ret
