def file_tree_test(postfix='.html'):
# --- 服务器数据 220 ---
path = '/data/hadoop/yisun/data/tianchi2/train/chongzu/html/'
filename = None
# filename = ['20546245.html']
label_file = '/data/hadoop/yisun/data/tianchi2/train_label/chongzu.train'
outpath = '../data/extract_result/train_chongzu/'
# # # --- 天池服务器
# path = '/home/118_16/data/chongzu_train_html/'
# filename = None
# # filename = ['20546245.html']
# label_file = '../data/train_data/train_labels/chongzu.train'
# outpath = '../data/extract_result/train_chongzu/'
# --- 本地外部数据 thinkpad ---
path = 'E:\\天池大赛\\公告数据\\天池大赛\\announcement_extract\\复赛数据\\复赛新增类型训练数据-20180712\\资产重组\\html\\'
filename = ['20546245.html']
label_file = '../data/train_data/train_labels/chongzu.train'
outpath = '../data/extract_result/train_chongzu/'
# --- 本地外部数据 ---
path = 'D:\\TianChi_competition\\公告信息抽取\\materials\\复赛\\复赛新增类型训练数据-20180712\\资产重组\\html\\'
filename = ['9895659.html']
label_file = '../data/train_data/train_labels/chongzu.train'
outpath = '../data/extract_result/train_chongzu/'
# --- 本地数据 ---
# path = '../data/temp2/'
# filename = None
# filename = ['19223567.html']
# outpath = '../data/temp2/result/'
# label_file = '../data/extract_result/train_chongzu/'
if filename == None:
files_name = os.listdir(path)
else:
files_name = filename
file_list = [i for i in files_name if i.endswith(postfix)]
html_dict = tian_chi.read_html(filepath=path, filename=file_list)
content = {}
for index in html_dict:
# -------------------------- 核心过程 -----------------------------
# --- 获取预内容 ---
res = tian_chi.extract_pre_content(html_dict[index])
tag = res[3]
_major_promption = res[2]
# --- 获取天池 text+table ---
text_list = tian_chi.get_content(tag)
# --- text to text_list
_content = content_format.tags_format(tags_list=text_list)[0]
major_promption = content_format.tags_format(_major_promption)[0]
# --- 转换为文档树 ---
t = content_format.FileTree(content_list=_content,
zhongdashixiangtishi=major_promption)
t.get_tree_list()
t.get_file_tree()
# result = t.get_tree_content(strcture=['第二章', '实际控制人'], method='content')
content[index] = t.titles
return t
def t2(drop_table=False):
path = '../data/temp2/'
filename = '23599.html'
outpath = '../data/temp2/'
html_dict = tc.read_html(filepath=path, filename=filename)
contents = {}
data = {}
for n, index in enumerate(html_dict):
print('processing {0} -- total: {1} this: {2}'.format(
index, len(html_dict), n))
# 是否去掉表格tag
h = html_dict[index]
if drop_table:
for i in h.find_all('table'):
i.decompose()
mulu, shiyi_dict, major_promption, tag = tc.extract_pre_content(h)
_content = tc.get_content(tag)
_content, part_table, table_failed = cf.tags_format(_content)
content = {
'mulu': mulu,
'shiyi': shiyi_dict,
'major_promption': major_promption,
'content': _content
}
contents[index] = content
res = cf.file_tree(content['mulu'], content['shiyi'],
content['major_promption'], content['content'])
# data = res.get_file_tree()
data[index] = res.get_tree_list()
return data
def html2file_tree(html, drop_table=False):
mulu, shiyi_dict, major_promption, tag = tian_chi.extract_pre_content(html)
_content = tian_chi.get_content(tag)
_content, part_table, table_failed = content_format.tags_format(_content)
content = {
'mulu': mulu,
'shiyi': shiyi_dict,
'major_promption': major_promption,
'content': _content
}
res = content_format.FileTree(content['mulu'], content['shiyi'],
content['major_promption'],
content['content'])
data = res.get_tree_list()
return data
def get_pre_content(path,
filename,
drop_table=False,
keys=['mulu', 'shiyi', 'major_promption', 'content'],
text_trans=False,
df_json=True):
"""
获取文件的预信息
:param path:
:param filename:
:param drop_table:
:param keys:
:param text_trans:
:return:
"""
html_dict = tian_chi.read_html(filepath=path, filename=filename)
contents = {}
for n, index in enumerate(html_dict):
print('processing {0} -- total: {1} this: {2}'.format(
index, len(html_dict), n))
# 是否去掉表格tag
h = html_dict[index]
if drop_table:
for i in h.find_all('table'):
i.decompose()
mulu, shiyi_dict, major_promption, tag = tian_chi.extract_pre_content(
h)
_content = tian_chi.get_content(tag)
# 是否转换 tags ---> text list
if text_trans:
_content, part_table, table_failed = content_format.tags_format(
_content, df_json=True)
content = {
'mulu': mulu,
'shiyi': shiyi_dict,
'major_promption': major_promption,
'content': _content
}
res = {}
for i in keys:
res[i] = content[i]
contents[index] = res
return contents
def get_file_tree(path,
filename,
drop_table=False,
method='list',
df_json=False):
# --- pre_main ---
html_dict = tian_chi.read_html(filepath=path, filename=filename)
contents = {}
data = {}
for n, index in enumerate(html_dict):
print('processing {0} -- total: {1} this: {2}'.format(
index, len(html_dict), n))
# 是否去掉表格tag
h = html_dict[index]
if drop_table:
for i in h.find_all('table'):
i.decompose()
mulu, shiyi_dict, major_promption, tag = tian_chi.extract_pre_content(
h)
_content = tian_chi.get_content(tag)
_content, part_table, table_failed = content_format.tags_format(
_content, df_json=df_json)
content = {
'mulu': mulu,
'shiyi': shiyi_dict,
'major_promption': major_promption,
'content': _content
}
contents[index] = content
res = content_format.FileTree(content['mulu'], content['shiyi'],
content['major_promption'],
content['content'])
# data = res.get_file_tree()
if method == 'tree':
res.get_file_tree()
if method == 'list':
res.get_tree_list()
if method == 'both':
res.get_file_tree()
res.get_tree_list()
data[index] = res
return data