nlpertools 1.0.4__py3-none-any.whl → 1.0.6.dev0__py3-none-any.whl

nlpertools/other.py

@@ -1,188 +1,364 @@
-#!/usr/bin/python3.8
-# -*- coding: utf-8 -*-
-# @Author  : youshu.Ji
-import string
-import time
-from concurrent.futures import ThreadPoolExecutor
-from functools import wraps
-import requests
-from tqdm import tqdm
-
-try:
-    from elasticsearch import Elasticsearch
-    import pyquery as pq
-    from ltp import LTP
-except:
-    pass
-
-CHINESE_PUNCTUATION = list('，。；：‘’“”！？《》「」【】<>（）、')
-ENGLISH_PUNCTUATION = list(',.;:\'"!?<>()')
-# other ----------------------------------------------------------------------
-# 统计词频
-def calc_word_count(list_word, mode, path='tempcount.txt', sort_id=1, is_reverse=True):
-    word_count = {}
-    for key in list_word:
-        if key not in word_count:
-            word_count[key] = 1
-        else:
-            word_count[key] += 1
-    word_dict_sort = sorted(word_count.items(), key=lambda x: x[sort_id], reverse=is_reverse)
-    if mode == 'w':
-        for key in word_dict_sort:
-            writetxt_a(str(key[0]) + '\t' + str(key[1]) + '\n', path)
-    elif mode == 'p':
-        for key in word_dict_sort:
-            print(str(key[0]) + '\t' + str(key[1]))
-    elif mode == 'u':
-        return word_dict_sort
-
-# 字典去重
-def dupl_dict(dict_list, key):
-    new_dict_list, value_set = [], []
-    print('去重中...')
-    for i in tqdm(dict_list):
-        if i[key] not in value_set:
-            new_dict_list.append(i)
-            value_set.append(i[key])
-    return new_dict_list
-
-
-def del_special_char(sentence):
-    special_chars = ['\ufeff', '\xa0', '\u3000', '\xa0', '\ue627']
-    for i in special_chars:
-        sentence = sentence.replace(i, '')
-    return sentence
-
-
-def en_pun_2_zh_pun(sentence):
-    # TODO 因为引号的问题，所以我没有写
-    for i in ENGLISH_PUNCTUATION:
-        pass
-
-
-def spider(url):
-    """
-
-    :param url:
-    :return:
-    """
-    if 'baijiahao' in url:
-        content = requests.get(url)
-        # print(content.text)
-        html = pq.PyQuery(content.text)
-        title = html('.index-module_articleTitle_28fPT').text()
-        res = html('.index-module_articleWrap_2Zphx').text().rstrip('举报/反馈')
-        return '{}\n{}'.format(title, res)
-
-
-def eda(sentence):
-    url = 'http://x.x.x.x:x/eda'
-    json_data = dict({"sentence": sentence})
-    res = requests.post(url, json=json_data)
-    return res.json()['eda']
-
-
-def find_language(text):
-    #  TODO 替换为开源包
-    letters = list(string.ascii_letters)
-    if len(text) > 50:
-        passage = text[:50]
-        len_passage = 50
-    else:
-        len_passage = len(text)
-    count = 0
-    for c in passage:
-        if c in letters:
-            count += 1
-    if count / len_passage > 0.5:
-        return "en"
-    else:
-        return "not en"
-
-
-def print_prf():
-    from sklearn.metrics import precision_recall_fscore_support
-    y_true = [0, 1, 2, 1, 1, 2, 3, 1, 1, 1]
-    y_pred = [0, 1, 2, 1, 1, 2, 3, 1, 1, 1]
-    p, r, f, s = precision_recall_fscore_support(y_true=y_true, y_pred=y_pred)
-    print("p\t{}".format(p))
-    print("r\t{}".format(r))
-    print("f\t{}".format(f))
-    print("s\t{}".format(s))
-    """
-    result = precision_recall_fscore_support(y_true=y_true, y_pred=y_pred, labels=label)
-
-    for i in range(len(label)):
-        res = []
-        for k in result:
-            res.append('%.5f' % k[i])
-        print('{}: {} {} {}'.format(label[i], *res[:3]))
-    """
-
-
-def print_cpu():
-    import psutil
-    p = psutil.Process()
-    # pro_info = p.as_dict(attrs=['pid', 'name', 'username'])
-    print(psutil.cpu_count())
-
-
-@nlpertools.fn_timer
-def stress_test(func, ipts):
-    with ThreadPoolExecutor() as executor:
-        results = list(tqdm(executor.map(func, ipts), total=len(ipts)))
-    return results
-
-# 定义装饰器
-def fn_timer(function):
-    @wraps(function)
-    def function_timer(*args, **kwargs):
-        t0 = time.time()
-        result = function(*args, **kwargs)
-        t1 = time.time()
-        print('[finished {func_name} in {time:.2f}s]'.format(func_name=function.__name__, time=t1 - t0))
-        return result
-
-    return function_timer
-
-
-def get_substring_loc(text, subtext):
-    import re
-    res = re.finditer(
-        subtext.replace('\\', '\\\\').replace('?', '\?').replace('(', '\(').replace(')', '\)').replace(']',
-                                                                                                       '\]').replace(
-            '[', '\[').replace('+', '\+'), text)
-    l, r = [i for i in res][0].regs[0]
-    return l, r
-
-
-def tf_idf(corpus, save_path):
-    from sklearn.feature_extraction.text import TfidfTransformer
-    from sklearn.feature_extraction.text import CountVectorizer
-    tfidfdict = {}
-    vectorizer = CountVectorizer()  # 该类会将文本中的词语转换为词频矩阵，矩阵元素a[i][j] 表示j词在i类文本下的词频
-    transformer = TfidfTransformer()  # 该类会统计每个词语的tf-idf权值
-    tfidf = transformer.fit_transform(
-        vectorizer.fit_transform(corpus))  # 第一个fit_transform是计算tf-idf，第二个fit_transform是将文本转为词频矩阵
-    word = vectorizer.get_feature_names()  # 获取词袋模型中的所有词语
-    weight = tfidf.toarray()  # 将tf-idf矩阵抽取出来，元素a[i][j]表示j词在i类文本中的tf-idf权重
-    for i in range(len(weight)):  # 打印每类文本的tf-idf词语权重，第一个for遍历所有文本，第二个for便利某一类文本下的词语权重
-        for j in range(len(word)):
-            getword = word[j]
-            getvalue = weight[i][j]
-            if getvalue != 0:  # 去掉值为0的项
-                if getword in tfidfdict:  # 更新全局TFIDF值
-                    tfidfdict[getword] += float(getvalue)
-                else:
-                    tfidfdict.update({getword: getvalue})
-    sorted_tfidf = sorted(tfidfdict.items(), key=lambda d: d[1], reverse=True)
-    to_write = ['{} {}'.format(i[0], i[1]) for i in sorted_tfidf]
-    writetxt_w_list(to_write, save_path, num_lf=1)
-
-# 常用函数参考
-# import tensorflow as tf
-#
-# gpu_options = tf.compat.v1.GPUOptions(allow_growth=True)
-# sess = tf.compat.v1.Session(config=tf.compat.v1.ConfigProto(gpu_options=gpu_options))
-# for gpu in tf.config.experimental.list_physical_devices('GPU'):
-#     tf.config.experimental.set_memory_growth()
+#!/usr/bin/python3.8
+# -*- coding: utf-8 -*-
+# @Author  : youshu.Ji
+import itertools
+import os
+import re
+import string
+from concurrent.futures import ThreadPoolExecutor
+from functools import reduce
+import math
+import datetime
+import psutil
+from .io.file import writetxt_w_list, writetxt_a
+# import numpy as np
+# import psutil
+# import pyquery as pq
+# import requests
+# import torch
+# from sklearn.feature_extraction.text import CountVectorizer, TfidfTransformer
+# from sklearn.metrics import precision_recall_fscore_support
+# from tqdm import tqdm
+# from win32evtlogutil import langid
+from .utils.package import *
+
+CHINESE_PUNCTUATION = list('，。；：‘’“”！？《》「」【】<>（）、')
+ENGLISH_PUNCTUATION = list(',.;:\'"!?<>()')
+OTHER_PUNCTUATION = list('!@#$%^&*')
+
+
+def seed_everything():
+    import torch
+    # seed everything
+    seed = 7777777
+    np.random.seed(seed)
+    torch.manual_seed(seed)  # CPU随机种子确定
+    torch.cuda.manual_seed(seed)
+    torch.cuda.manual_seed_all(seed)
+
+
+def sent_email(mail_user, mail_pass, receiver, title, content, attach_path=None):
+    import smtplib
+    from email.mime.multipart import MIMEMultipart
+    from email.mime.text import MIMEText
+    from email.mime.application import MIMEApplication
+
+    mail_host = 'smtp.qq.com'
+    mail_user = mail_user
+    mail_pass = mail_pass
+    sender = mail_user
+
+    message = MIMEMultipart()
+    message.attach(MIMEText(content, 'plain', 'utf-8'))
+    if attach_path:
+        attachment = MIMEApplication(open(attach_path, 'rb').read())
+        attachment["Content-Type"] = 'application/octet-stream'
+        attachment.add_header('Content-Dispositon', 'attachment',
+                              filename=('utf-8', '', attach_path))  # 注意：此处basename要转换为gbk编码，否则中文会有乱码。
+        message.attach(attachment)
+    message['Subject'] = title
+    message['From'] = sender
+    message['To'] = receiver
+
+    try:
+        smtp_obj = smtplib.SMTP()
+        smtp_obj.connect(mail_host, 25)
+        smtp_obj.login(mail_user, mail_pass)
+        smtp_obj.sendmail(sender, receiver, message.as_string())
+        smtp_obj.quit()
+        print('send email success')
+    except smtplib.SMTPException as e:
+        print('send failed', e)
+
+
+def convert_np_to_py(obj):
+    if isinstance(obj, dict):
+        return {k: convert_np_to_py(v) for k, v in obj.items()}
+    elif isinstance(obj, list):
+        return [convert_np_to_py(v) for v in obj]
+    elif isinstance(obj, np.float64) or isinstance(obj, np.float32):
+        return float(obj)
+    else:
+        return obj
+
+
+def git_push():
+    """
+    针对国内提交github经常失败，自动提交
+    """
+    num = -1
+    while 1:
+        num += 1
+        print("retry num: {}".format(num))
+        info = os.system("git push --set-upstream origin main")
+        print(str(info))
+        if not str(info).startswith("fatal"):
+            print("scucess")
+            break
+
+
+def snake_to_camel(s: str) -> str:
+    """
+    author: u
+    将 snake case 转换到 camel case.
+    :param s: snake case variable
+    :return:
+    """
+    return s.title().replace("_", "")
+
+
+def camel_to_snake(s: str) -> str:
+    """
+    将 camel case 转换到 snake case.
+    :param s: camel case variable
+    :return:
+    """
+    return reduce(lambda x, y: x + ('_' if y.isupper() else '') + y, s).lower()
+
+
+# other ----------------------------------------------------------------------
+# 统计词频
+def calc_word_count(list_word, mode, path='tempcount.txt', sort_id=1, is_reverse=True):
+    word_count = {}
+    for key in list_word:
+        if key not in word_count:
+            word_count[key] = 1
+        else:
+            word_count[key] += 1
+    word_dict_sort = sorted(word_count.items(), key=lambda x: x[sort_id], reverse=is_reverse)
+    if mode == 'w':
+        for key in word_dict_sort:
+            writetxt_a(str(key[0]) + '\t' + str(key[1]) + '\n', path)
+    elif mode == 'p':
+        for key in word_dict_sort:
+            print(str(key[0]) + '\t' + str(key[1]))
+    elif mode == 'u':
+        return word_dict_sort
+
+
+# 字典去重
+def dupl_dict(dict_list, key):
+    new_dict_list, value_set = [], []
+    print('去重中...')
+    for i in tqdm(dict_list):
+        if i[key] not in value_set:
+            new_dict_list.append(i)
+            value_set.append(i[key])
+    return new_dict_list
+
+
+def multi_thread_run(_task, data):
+    with ThreadPoolExecutor() as executor:
+        result = list(tqdm(executor.map(_task, data), total=len(data)))
+    return result
+
+
+def del_special_char(sentence):
+    special_chars = ['\ufeff', '\xa0', '\u3000', '\xa0', '\ue627']
+    for i in special_chars:
+        sentence = sentence.replace(i, '')
+    return sentence
+
+
+def en_pun_2_zh_pun(sentence):
+    # TODO 因为引号的问题，所以我没有写
+    for i in ENGLISH_PUNCTUATION:
+        pass
+
+
+def spider(url):
+    """
+
+    :param url:
+    :return:
+    """
+    if 'baijiahao' in url:
+        content = requests.get(url)
+        # print(content.text)
+        html = pq.PyQuery(content.text)
+        title = html('.index-module_articleTitle_28fPT').text()
+        res = html('.index-module_articleWrap_2Zphx').text().rstrip('举报/反馈')
+        return '{}\n{}'.format(title, res)
+
+
+def eda(sentence):
+    url = 'https://x.x.x.x:x/eda'
+    json_data = dict({"sentence": sentence})
+    res = requests.post(url, json=json_data)
+    return res.json()['eda']
+
+
+def find_language(text):
+    #  TODO 替换为开源包
+    letters = list(string.ascii_letters)
+    if len(text) > 50:
+        passage = text[:50]
+        len_passage = 50
+    else:
+        len_passage = len(text)
+    count = 0
+    for c in passage:
+        if c in letters:
+            count += 1
+    if count / len_passage > 0.5:
+        return "en"
+    else:
+        return "not en"
+
+
+def print_prf(y_true, y_pred, label=None):
+    # y_true = [0, 1, 2, 1, 1, 2, 3, 1, 1, 1]
+    # y_pred = [0, 1, 2, 1, 1, 2, 3, 1, 1, 1]
+    # p, r, f, s = precision_recall_fscore_support(y_true=y_true, y_pred=y_pred)
+    # print("p\t{}".format(p))
+    # print("r\t{}".format(r))
+    # print("f\t{}".format(f))
+    # print("s\t{}".format(s))
+    result = precision_recall_fscore_support(y_true=y_true, y_pred=y_pred, labels=label)
+
+    for i in range(len(label)):
+        res = []
+        for k in result:
+            res.append('%.5f' % k[i])
+        print('{}: {} {} {}'.format(label[i], *res[:3]))
+
+
+def print_cpu():
+    p = psutil.Process()
+    # pro_info = p.as_dict(attrs=['pid', 'name', 'username'])
+    print(psutil.cpu_count())
+
+
+def stress_test(func, ipts):
+    with ThreadPoolExecutor() as executor:
+        results = list(tqdm(executor.map(func, ipts), total=len(ipts)))
+    return results
+
+
+def get_substring_loc(text, subtext):
+    res = re.finditer(
+        subtext.replace('\\', '\\\\').replace('?', '\?').replace('(', '\(').replace(')', '\)').replace(']',
+                                                                                                       '\]').replace(
+            '[', '\[').replace('+', '\+'), text)
+    l, r = [i for i in res][0].regs[0]
+    return l, r
+
+
+def squeeze_list(high_dim_list):
+    return list(itertools.chain.from_iterable(high_dim_list))
+
+
+def unsqueeze_list(flatten_list, each_element_len):
+    two_dim_list = [flatten_list[i * each_element_len:(i + 1) * each_element_len] for i in
+                    range(len(flatten_list) // each_element_len)]
+    return two_dim_list
+
+
+def auto_close():
+    """
+    针对企业微信15分钟会显示离开的机制，假装自己还在上班
+    """
+    import pyautogui as pg
+    import time
+    import os
+    cmd = 'schtasks /create /tn shut /tr "shutdown -s -f" /sc once /st 23:30'
+    os.system(cmd)
+    while 1:
+        pg.moveTo(970, 17, 2)
+        pg.click()
+        time.sleep(840)
+
+
+def tf_idf(corpus, save_path):
+    tfidfdict = {}
+    vectorizer = CountVectorizer()  # 该类会将文本中的词语转换为词频矩阵，矩阵元素a[i][j] 表示j词在i类文本下的词频
+    transformer = TfidfTransformer()  # 该类会统计每个词语的tf-idf权值
+    tfidf = transformer.fit_transform(
+        vectorizer.fit_transform(corpus))  # 第一个fit_transform是计算tf-idf，第二个fit_transform是将文本转为词频矩阵
+    word = vectorizer.get_feature_names()  # 获取词袋模型中的所有词语
+    weight = tfidf.toarray()  # 将tf-idf矩阵抽取出来，元素a[i][j]表示j词在i类文本中的tf-idf权重
+    for i in range(len(weight)):  # 打印每类文本的tf-idf词语权重，第一个for遍历所有文本，第二个for便利某一类文本下的词语权重
+        for j in range(len(word)):
+            getword = word[j]
+            getvalue = weight[i][j]
+            if getvalue != 0:  # 去掉值为0的项
+                if getword in tfidfdict:  # 更新全局TFIDF值
+                    tfidfdict[getword] += float(getvalue)
+                else:
+                    tfidfdict.update({getword: getvalue})
+    sorted_tfidf = sorted(tfidfdict.items(), key=lambda d: d[1], reverse=True)
+    to_write = ['{} {}'.format(i[0], i[1]) for i in sorted_tfidf]
+    writetxt_w_list(to_write, save_path, num_lf=1)
+
+
+class GaussDecay(object):
+    """
+    当前只实现了时间的，全部使用默认值
+    """
+
+    def __init__(self, origin='2022-08-02', scale='90d', offset='5d', decay=0.5, task="time"):
+        self.origin = origin
+        self.task = task
+        self.scale, self.offset = self.translate(scale, offset)
+        self.decay = decay
+        self.time_coefficient = 0.6
+        self.related_coefficient = 0.4
+
+    def translate(self, scale, offset):
+        """
+        将领域的输入转化为标准
+        :return:
+        """
+        if self.task == "time":
+            scale = 180
+            offset = 5
+        else:
+            scale = 180
+            offset = 5
+        return scale, offset
+
+    @staticmethod
+    def translated_minus(field_value):
+        origin = datetime.datetime.now()
+        field_value = datetime.datetime.strptime(field_value, '%Y-%m-%d %H:%M:%S')
+        return (origin - field_value).days
+
+    def calc_exp(self):
+        pass
+
+    def calc_liner(self):
+        pass
+
+    def calc_gauss(self, raw_score, field_value):
+        """
+        $$S(doc)=exp(-\frac{max(0,|fieldvalues_{doc}-origin|-offset)^2}{2σ^2})$$ -
+        $$σ^2=-scale^2/(2·ln(decay))$$
+        :param raw_score:
+        :param field_value:
+        :return:
+        """
+        numerator = max(0, (abs(self.translated_minus(field_value)) - self.offset)) ** 2
+        sigma_square = -1 * self.scale ** 2 / (2 * math.log(self.decay, math.e))
+        denominator = 2 * sigma_square
+        s = math.exp(-1 * numerator / denominator)
+        return round(self.time_coefficient * s + self.related_coefficient * raw_score, 7)
+
+
+if __name__ == '__main__':
+    gauss_decay = GaussDecay()
+    res = gauss_decay.calc_gauss(raw_score=1, field_value="2021-05-29 14:31:13")
+    print(res)
+    # res = gauss_decay.calc_gauss(raw_score=1, field_value="2022-05-29 14:31:13")
+    # print(res)
+    # res = gauss_decay.calc_gauss(raw_score=1, field_value="2022-05-29 14:31:13")
+    # print(res)
+    # res = gauss_decay.calc_gauss(raw_score=1, field_value="2022-05-29 14:31:13")
+    # print(res)
+
+# 常用函数参考
+# import tensorflow as tf
+#
+# gpu_options = tf.compat.v1.GPUOptions(allow_growth=True)
+# sess = tf.compat.v1.Session(config=tf.compat.v1.ConfigProto(gpu_options=gpu_options))
+# for gpu in tf.config.experimental.list_physical_devices('GPU'):
+#     tf.config.experimental.set_memory_growth()