wechet-anayze-main.py

"""
Create on 2020-11-21
@author: muxiaohe
"""
import imageio
import jieba
import pandas as pd
import time
import matplotlib.pyplot as plt
from pyecharts.charts import Bar
from pyecharts import options as opts
from pyecharts.charts import Pie
import wordcloud

# 数据预处理整体部分，如不关系具体预处理过程，直接使用此部分代码即可
def data_pretreatment(message_file_path, contact_file_path):
    message = pd.read_csv(message_file_path, sep=',', encoding="utf-8", low_memory=False)
    message = message[['type', 'isSend', 'createTime', 'talker', 'content']]
    message.loc[message.type != 1, 'content'] = 0

    contact = pd.read_csv(contact_file_path, sep=',', encoding="utf-8", low_memory=False)
    contact = contact[['username', 'alias', 'conRemark', 'nickname']]
    contact = contact.drop(contact[pd.isna(contact.conRemark)].index)
    contact = contact[['username', 'conRemark']]
    return message, contact

# 获取常用联系人聊天次数
def get_chat_nums(message, contact):
    """
    :param message_path: 预处理完成后的message表存储路径
    :param contact_path: 预处理完成后的contact表存储路径
    :return:
    """
    # message_path = r'D:\wechet-anayze\pre-message-2.txt'
    # contact_path = r'D:\wechet-anayze\pre-recontact.csv'

    # 提取出联系人列表中用户名和备注名称
    contact = contact[['username', 'conRemark']]
    # 将用户名提取出来
    username = contact['username'].tolist()
    print(type(username))
    # 将用户名及备注名提取为一个字典
    contact_dict = dict(zip(contact['username'], contact['conRemark']))
    # 联系人及其聊天次数集合
    contact_sum_message = {}
    # 全部联系人聊天次数集合
    sum_message = 0
    # 联系人列表
    uname_list = []
    # 联系人列表对应的聊天次数列表
    chat_num_list = []
    # 遍历联系人列表，并逐一统计聊天次数
    for uname in username:
        # 根据微信id获取真实姓名,key为真实姓名
        key = contact_dict.get(uname)
        # 根据微信id统计聊天次数，value:聊天次数
        value = (message['talker'] == uname).sum()
        # 过滤聊天次数为0的联系人，只保留聊天次数不为0的联系人
        if value != 0:
            contact_sum_message[key] = value
            sum_message += value
            uname_list.append(key)
            # 这里需特别注意：value值也即聊天的次数格式是int64,但是pyecharts中如果传入的是int64时，最终渲染出的html文件中会数据会丢失，
            # 所以需转为int值（血泪教训）
            chat_num_list.append(int(value))

    # print(contact_sum_message)
    print("总聊天次数： ", sum_message)

    # 使用pyecharts绘制柱状图
    c = (
        Bar(init_opts=opts.InitOpts(width="1600px", height="600px", page_title="聊天次数统计"))
        .add_xaxis(uname_list)
        .add_yaxis(series_name="聊天次数", y_axis=chat_num_list, color='#FF6666')
        .set_global_opts(
            # 标题配置
            title_opts=opts.TitleOpts(title="聊天次数统计"),
            # X轴区域缩放配置项,可使用list同时配置多个配置项
            datazoom_opts=[opts.DataZoomOpts(range_start=20, range_end=40), opts.DataZoomOpts(type_="inside")],
            # 区域选择组件
            brush_opts=opts.BrushOpts(),
            # X坐标轴旋转
            xaxis_opts=opts.AxisOpts(axislabel_opts=opts.LabelOpts(rotate=-15)),
            # 工具箱组件
            toolbox_opts=opts.ToolboxOpts(),
            # 图例配置
            legend_opts=opts.LegendOpts(is_show=False),
        )
        .set_series_opts(
            label_opts=opts.LabelOpts(is_show=False),
            # 配置最大值最小值刻度线
            markline_opts=opts.MarkLineOpts(
                data=[
                    opts.MarkLineItem(type_="min", name="最小值"),
                    opts.MarkLineItem(type_="max", name="最大值"),
                    opts.MarkLineItem(type_="average", name="平均值"),
                ]
            ),
        )
        .render("chat_num_count.html")
    )

# 获取各个消息聊天记录数量，并使用pycharts绘图
def get_message_type_frequency(message, wxid):
    """
    :param message: message表内容
    :param wxid: 待查询人的微信id
    :return:
    """
     # 进行数据筛选，选择message表中与所需微信id一致的数据
    message = message[message['talker'] == wxid]
    # 根据消息类型统计每种类型的频次（索引为数字编码）
    chat_type_count = message['type'].groupby(message['type']).size()
    # 消息类型对应关系
    message_type = {'1': '文本内容', "3": "图片及视频", "34": "语音消息", "42": "名片信息", "43": "图片及视频",
                    "47": "表情包", "48": "定位信息", "49": "小程序链接", "10000": "消息撤回提醒", "1048625": "网络照片",
                    "16777265": "链接信息", "419430449": "微信转账", "436207665": "红包", "469762097": "红包",
                    "-1879048186": "位置共享"}
    # 集合对象，功能与chat_type_count相同，存储（聊天类型：频次）信息（索引为对应中文类型）
    chat_type_count_dict = {}
    # 根据消息类型代码
    for key in chat_type_count.index:
        if str(key) in message_type.keys():
            # print(message_type.get(str(key)))
            chat_type_count_dict[message_type.get(str(key))] = chat_type_count[key]
        else:
            chat_type_count_dict[key] = chat_type_count[key]
    # print("结果集类型: ", type(chat_type_count_dict))
    # print(chat_type_count_dict)

    x_data = []
    y_data = []
    for key in chat_type_count_dict:
        temp = [str(key), chat_type_count_dict.get(key)]
        x_data.append(str(key))
        y_data.append(int(chat_type_count_dict.get(key)))

    a1 = []
    for z in zip(x_data, y_data):
        a1.append(z)

    pie = Pie(init_opts=opts.InitOpts(width="1600px", height="600px", page_title="消息类型统计"))
    pie.add(
        "",
        data_pair=a1,
        center=["35%", "60%"],
    )
    pie.set_global_opts(
        title_opts=opts.TitleOpts(title="Pie-调整位置"),
        legend_opts=opts.LegendOpts(pos_left="15%"),
    )
    pie.set_series_opts(label_opts=opts.LabelOpts(formatter="{b}:{c}"))
    pie.render("message_type_count.html")


# 从message文件中，根据createTime信息获取年份，月份，日期，星期，yday，小时，分钟信息，并将之存储到新的文件中
def get_time_file(message, save_path):
    message = message[['talker', 'createTime', 'isSend']]
    year = []
    month = []
    day = []
    hour = []
    minute = []
    yday = []
    wday = []
    content = []

    for sec_time in message['createTime']:
        sec_time = sec_time / 1000
        struct_time = time.localtime(sec_time)

        year.append(struct_time.tm_year)
        month.append(struct_time.tm_mon)
        day.append(struct_time.tm_mday)
        hour.append(struct_time.tm_hour)
        minute.append(struct_time.tm_min)
        yday.append(struct_time.tm_yday)
        wday.append(struct_time.tm_wday)

    message['year'] = year
    message['month'] = month
    message['day'] = day
    message['wday'] = wday
    message['yday'] = yday
    message['hour'] = hour
    message['minute'] = minute
    message['content'] = content
    # 写入到指定位置
    message.to_csv(save_path, encoding='utf_8_sig', header=True, index=False)

# 根据聊天记录生成词云
def get_wordcloud(message, stopword_path, wxid, image_path):
    """
    :param message: message表信息
    :param stopword_path: 停用词文件存储位置
    :param wxid: 待查询人微信id
    :return:
    """
    # 数据筛选，选择对应微信id的信息
    message = message[message['talker'] == wxid]
    # 提取聊天内容信息
    content = message['content']
    print(content.shape)
    print(content)
    # 中文字型存储路径
    font_path = r'C:\Windows\Fonts\MSYH.TTC'
    # 是否选用分词
    wordcut_flag = True
    # 词云图片输出路径
    image_out_name = 'word-heart.png'

    # struct_time = time.localtime()
    # image_out_name = str(struct_time.tm_year) + str(struct_time.tm_mon) + str(struct_time.tm_mday) + "-" + str(
    #     struct_time.tm_hour) + str(struct_time.tm_min) + "-" + image_out_name

    # 读取停用词表
    stopwords = [line.strip() for line in open(stopword_path, encoding='utf-8').readlines()]

    if image_out_name is None:
        image_out_name = 'word-heart.png'
    if wordcut_flag:
        print("进行中文分词")
        outstr = ""
        # text = ",".join(content)
        text = " ".join('%s' % id for id in content)
        text_list = jieba.lcut(text, cut_all=False)
        for word in text_list:
            if word not in stopwords:
                if word != '\t' and '\n':
                    outstr += word
                    outstr += " "

        # 如果想存储分词后结果，可取消下方注释
        # savepath = r'D:/wechet-anayze/textlist.txt'
        # fp = open(savepath, 'w', encoding='utf8', errors='ignore')
        # fp.write(outstr)
        # fp.close()

        text = outstr
    else:
        print("不进行中文分词")
        text = " ".join(content)
    # 词云形状图片位置
    mk = imageio.imread(image_path)

    # 构建并配置词云对象w，注意要加scale参数，提高清晰度
    w = wordcloud.WordCloud(width=1000,
                            height=700,
                            background_color='white',
                            font_path=font_path,
                            mask=mk,
                            scale=2,
                            stopwords=None,
                            contour_width=1,
                            contour_color='red')
    # 将string变量传入w的generate()方法，给词云输入文字
    w.generate(text)
    # 展示图片
    # 根据原始背景图片的色调进行上色
    image_colors = wordcloud.ImageColorGenerator(mk)
    plt.imshow(w.recolor(color_func=image_colors))
    # 根据原始黑白色调进行上色
    # plt.imshow(wc.recolor(color_func=grey_color_func, random_state=3), interpolation='bilinear') #生成黑白词云图
    # 根据函数原始设置进行上色
    # plt.imshow(wc)

    # 隐藏图像坐标轴
    plt.axis("off")
    plt.show()

    # 将词云图片导出到当前文件夹
    w.to_file(image_out_name)

if __name__ == '__main__':
    # 原始message表路径
    message_file_path = r''
    # 原始contact表路径
    contact_file_path = r''
    # 修改此处代码，改为聊天对象微信id
    wxid = ""
    # 数据预处理
    message, contact = data_pretreatment(message_file_path, contact_file_path)

    # 停用词表存储路径
    stopword_path = r'stopword.txt'
    # 词云形状图片存储路径，如想更换词云形状，更换为其他图形即可
    image_path = "heart.png"
    get_wordcloud(message, stopword_path, wxid, image_path)
    # 获取聊天次数并绘制柱状图，结果以html形式存储
    get_chat_nums(message, contact)
    # 针对某个聊天对象统计其聊天类型,并绘图，结果以html形式存储
    get_message_type_frequency(message, wxid=wxid)

    # # 如有需要可将微信消息，按照年份，月份，日期，星期，yday，小时，分钟信息，并将之存储到文件中，可结合excel透视表分析出更多有意思的内容
    # save_path = ""
    # get_time_file(message)