h_novel / examples /preprocess /content_preprocess.py
qgyd2021's picture
[update]process filename
06124c0
raw
history blame
7.43 kB
#!/usr/bin/python3
# -*- coding: utf-8 -*-
import argparse
from pathlib import Path
import re
import shutil
import string
import pandas as pd
from project_settings import project_path
def get_args():
parser = argparse.ArgumentParser()
parser.add_argument(
"--data_dir",
default=(project_path / "data2").as_posix(),
type=str
)
parser.add_argument(
"--output_file",
default="metadata.csv",
type=str
)
args = parser.parse_args()
return args
class SentenceType(object):
def __init__(self):
self.advertise_keywords = [
"请访问",
]
self.url_keywords = [
"www.",
]
self.pattern_predictor_map = {
"www.(?:.*?).(net|com|cn)": "url"
}
def contains_advertise(self, text: str):
for k in self.advertise_keywords:
if text.__contains__(k):
return "advertise"
return None
def contains_url(self, text: str):
for k in self.url_keywords:
if text.__contains__(k):
return "url"
return None
def pattern_predict(self, text: str):
for k, v in self.pattern_predictor_map.items():
if re.search(k, text, flags=re.IGNORECASE) is not None:
return v
return None
def predict(self, text: str):
result = self.pattern_predict(text)
if result is not None:
return result
result = self.contains_advertise(text)
if result is not None:
return result
return None
class ContentPreprocess(object):
def __init__(self):
self.s_type = SentenceType()
self.punctuation_map = {
",": ",",
"。": ".",
"、": ",",
"?": "?",
":": ":",
"{": "{",
"}": "}",
"(": "(",
")": ")",
"【": "(",
"】": ")",
"「": "\"",
"」": "\"",
"『": "\"",
"』": "\"",
"《": "(",
"》": ")",
"”": "\"",
"“": "\"",
"‘": "\'",
"’": "\'",
"=": "",
"": "",
" ": "",
" ": "",
"\t": "",
"\n": "",
"\r": "",
"\v": "",
"\f": "",
}
self.string_map = {
"^_^": "",
"◆": "",
"☆": "",
"...": "…",
"()": "",
"{}": "",
"『』": "",
"《》": "",
"&lt;": "<",
"&gt;": ">",
"&amp;": " ",
"&nbsp;": " ",
"quot;": "\"",
"nbsp;": " ",
"amp;": "",
"&#10008": "✘",
"龙腾小说网ltxs520.com": "",
"龙腾小说吧www.ltxsba.net": "",
"龙腾小说网 ltxs520.com": "",
"龙腾小说 ltxs520.com": "",
"龙腾小说ltxs520.com": "",
"(..)免费": "",
"\"\"": "\"\n\"",
}
self.pattern_map = {
r"</(?:.*?)>": "",
r"⌒{4,}": "###",
r"#{4,}": "###"
}
self.queue = list()
self.queue_length = 0
self.max_seq_length = 128
self.output_list = list()
def char_b2q(self, uchar):
"""单个字符 半角转全角"""
inside_code = ord(uchar)
if inside_code < 0x0020 or inside_code > 0x7e:
return uchar
if inside_code == 0x0020:
inside_code = 0x3000
else:
inside_code += 0xfee0
return chr(inside_code)
def char_q2b(self, uchar):
"""单个字符 全角转半角"""
inside_code = ord(uchar)
if inside_code == 0x3000:
inside_code = 0x0020
else:
inside_code -= 0xfee0
if inside_code < 0x0020 or inside_code > 0x7e:
return uchar
return chr(inside_code)
def q2b(self, text: str):
"""全角转半角"""
result = ""
for c in text:
c = self.char_q2b(c)
result += c
return result
def remove_space(self, text: str):
text = text.replace(" ", "")
return text
def replace_punctuation(self, text: str):
text_ = text
for k, v in self.punctuation_map.items():
text_ = text_.replace(k, v)
if text_ != text:
text_ = self.replace_punctuation(text_)
return text_
def replace_string(self, text: str):
for k, v in self.string_map.items():
text = text.replace(k, v)
return text
def replace_by_pattern(self, text: str):
for k, v in self.pattern_map.items():
text = re.sub(k, v, text)
return text
def process(self, text: str):
text = self.q2b(text)
text = self.replace_punctuation(text)
text = self.replace_by_pattern(text)
text = self.replace_string(text)
self.in_queue(text)
return text
def in_queue(self, text: str):
if self.s_type.predict(text) is not None:
return
if len(text) == 0:
return
text_splits = text.split(".")
num_splits = len(text_splits)
for idx, text_split in enumerate(text_splits):
char_num = len(text_split)
if self.queue_length + char_num > self.max_seq_length:
text_ = "".join(self.queue)
self.output_list.append(text_)
self.queue = list()
self.queue_length = 0
if idx == 0 and len(self.queue) != 0:
self.queue.append("\n")
if num_splits == 1:
self.queue.append(text_split)
self.queue_length += char_num
else:
if idx + 1 != num_splits:
self.queue.append("{}.".format(text_split))
self.queue_length += char_num
else:
self.queue.append(text_split)
self.queue_length += char_num
return
def main():
args = get_args()
data_dir = Path(args.data_dir)
filename_list = data_dir.glob("*/*.txt")
s_type = SentenceType()
c_preprocess = ContentPreprocess()
result = list()
for filename in filename_list:
name = filename.stem
splits = name.split("_")
idx = splits[-1]
novel_name = "_".join(splits[:-1])
to_filename = Path(filename.as_posix().replace("/data2/", "/data3/"))
to_filename.parent.mkdir(parents=True, exist_ok=True)
with open(filename.as_posix(), "r", encoding="utf-8") as fin, \
open(to_filename.as_posix(), "w", encoding="utf-8") as fout:
for row in fin:
row = str(row).strip()
c_preprocess.process(row)
for row_ in c_preprocess.output_list:
# print(row_)
# print(len(row_))
# print("-" * 150)
# result.append(row_)
fout.write("{}\n".format(row_))
c_preprocess.output_list = list()
return
if __name__ == '__main__':
main()