MainGUI.py

#!/usr/bin/env python
# -*- coding:utf-8 -*-
# @FileName  :MainGUI.py
# @Time      :2020/2/17 11:34
# @Author    :supakito
import tkinter as tk
from tkinter import messagebox
from tkinter import filedialog
from tkinter import ttk
from tkinter import Toplevel
from tkinter import Label
from tkinter import Checkbutton
from tkinter import IntVar
from OriginalData import OriginalData
from analyzer.Analyzer import Columns
from analyzer.Analyzer import VolumeAnalyzer
from analyzer.Analyzer import OutpatientDescriptionAnalyzer
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
import config
from matplotlib.font_manager import FontProperties
import functools

class Main_GUI():
    def __init__(self):
        self.top = tk.Tk()
        self.top.title('Outpatient Volume Analysis Tools')
        self.top.geometry('500x300')
        # 原始的门诊信息
        self.__original_data = pd.DataFrame()
        # 根据门诊量统计出的门诊量信息
        self.__vol_data = pd.DataFrame()
        self.__init_menu()
        self.__init_table()
        # self.__init_config()
        self.__profile = config.Config.load()

    def run(self):
        self.top.mainloop()


    def __init_table(self):
        self.scrollbar = tk.Scrollbar(self.top, )
        self.scrollbar.pack(side=tk.RIGHT, fill=tk.Y)
        self.table = ttk.Treeview(self.top,
                                  yscrollcommand=self.scrollbar.set,
                                  show='headings')

    def __init_menu(self):
        menubar = tk.Menu(self.top)
        file_menu = tk.Menu(self.top)
        for item, func in zip(['打开CSV',
                               '打开序列化文件',
                               '另存为...',
                               ],
                              (self.__open_csv_files,
                               self.__open_dump_file,
                               self.__save_file)):
            file_menu.add_command(label=item, command=func)
        menubar.add_cascade(label="文件", menu=file_menu)

        process_menu = tk.Menu(self.top)
        for item, func in zip(['分组统计', '关键字分析'], [self.__statistic, self.__stat_word]):
            process_menu.add_command(label=item, command=func)
        menubar.add_cascade(label="数据处理", menu=process_menu)
        self.top['menu'] = menubar

        help_menu = tk.Menu(self.top)
        for item, func in zip(['配置'], [self.__init_config]):
            help_menu.add_command(label=item, command=func)
        menubar.add_cascade(label="help", menu=help_menu)
        self.top['menu'] = menubar

    def __init_stat_config_window(self):
        '''
        初始化门诊量统计的配置窗口
        :return:
        '''
        def stat_config_dropdown_list_selected(event):
            col1 = self._comvalue_1.get()
            col2 = self._comvalue_2.get()
            if col1 == Columns.Age:
                age_edit_box1.grid(column=2, row=0, )
            else:
                age_edit_box1.grid_forget()
            if col2 == Columns.Age:
                age_edit_box2.grid(column=2, row=1, )
            else:
                age_edit_box2.grid_forget()

        def confirm_button_click():
            col1 = self._comvalue_1.get()
            col2 = self._comvalue_2.get()
            if not col1:
                messagebox.showinfo('Error', '请选择一列进行统计')
                return
            if col1==col2==Columns.Age:
                messagebox.showinfo('Error', '两列不能都为AGE')
                return

            if col1 == Columns.Age:
                age_bins = age_edit_box1.get()
            elif col2 == Columns.Age:
                age_bins = age_edit_box2.get()
            else:
                age_bins = list()
            stat_conf_window.destroy()
            self.__plot(col1, col2, age_bins)


        stat_conf_window = Toplevel()
        stat_conf_window.title('分组统计配置')
        stat_conf_window.geometry('550x100')
        stat_conf_window.resizable(0, 0)
        selectable_items = [
            # Columns.Date,
            Columns.Year,
            Columns.Month,
            Columns.Day,
            Columns.Week,
            Columns.Is_holiday,
            Columns.Sex,
            Columns.Age,
            Columns.Len_pre_hld,
            Columns.Len_after_hld,
            Columns.Dist_pre_hld,
            Columns.Dist_after_hld
        ]

        lb1 = Label(stat_conf_window, text="分组统计参数1（x轴）：")
        lb1.grid(column=0, row=0, padx=8, pady=4)
        lb2 = Label(stat_conf_window, text="分组统计参数2（分组，可不选）：")
        lb2.grid(column=0, row=1, padx=8, pady=4)

        self._comvalue_1 = tk.StringVar()
        dropdownlist_1 = ttk.Combobox(stat_conf_window, textvariable=self._comvalue_1, state='readonly')  # 初始化
        dropdownlist_1["values"] = selectable_items
        dropdownlist_1.grid(column=1, row=0, padx=8, pady=4)
        dropdownlist_1.bind("<<ComboboxSelected>>", stat_config_dropdown_list_selected)

        self._comvalue_2 = tk.StringVar()
        dropdownlist_2 = ttk.Combobox(stat_conf_window, textvariable=self._comvalue_2, state='readonly')  # 初始化
        dropdownlist_2["values"] = selectable_items
        dropdownlist_2.grid(column=1, row=1, padx=8, pady=4)
        dropdownlist_2.bind("<<ComboboxSelected>>", stat_config_dropdown_list_selected)

        age_edit_box1 = tk.Entry(stat_conf_window)
        age_edit_box1.insert(0, '10, 20, 30, 40, 50, 60, 70, 80, 90')
        age_edit_box2 = tk.Entry(stat_conf_window)
        age_edit_box2.insert(0, '10, 20, 30, 40, 50, 60, 70, 80, 90')

        confirm_button = tk.Button(stat_conf_window, text="开始绘图", command=confirm_button_click)
        confirm_button.grid(column=1, row=2, padx=8, pady=4)

    def __init_stat_word_window(self):
        '''
        对词进行分析
        :return:
        '''
        def confirm_button_click():
            cols = edit_box1.get().split()
            poses = edit_box2.get().split()
            self.__profile.nlp_fields = cols
            self.__profile.poses = poses
            config.Config.store(self.__profile)
            f = cb_freq_v.get()
            tf = cb_tfidf_v.get()
            tr = cb_textrank_v.get()
            if f+tf+tr == 0:
                messagebox.showinfo('提示', '必须选择一种权重计算方式')
                return
            else:
                results = list()
                if f:
                    result = OutpatientDescriptionAnalyzer.get_freq_rank(data = self.__original_data,
                                                                         focused_fields = self.__profile.nlp_fields,
                                                                         poses = self.__profile.poses,
                                                                         topK = self.__profile.topK)
                    df = pd.DataFrame([dict(result)]).T
                    df.columns = ['freq']
                    results.append(df)
                if tf:
                    result = OutpatientDescriptionAnalyzer.get_tf_idf_rank(data = self.__original_data,
                                                                           focused_fields = self.__profile.nlp_fields,
                                                                           poses = self.__profile.poses,
                                                                           topK = self.__profile.topK)
                    df = pd.DataFrame([dict(result)]).T
                    df.columns = ['tfidf']
                    results.append(df)
                if tr:
                    result = OutpatientDescriptionAnalyzer.get_textrank_rank(data = self.__original_data,
                                                                             focused_fields = self.__profile.nlp_fields,
                                                                             poses = self.__profile.poses,
                                                                             topK = self.__profile.topK)
                    df = pd.DataFrame([dict(result)]).T
                    df.columns = ['textrank']
                    results.append(df)

                df_r = pd.concat(results, axis=1)
                file_path = filedialog.asksaveasfilename(title=u'保存文件',
                                                 filetypes=[("csv", ".csv")],
                                                 defaultextension='.csv')
                df_r.to_csv(file_path, encoding='utf-8')
                stat_word_conf_window.destroy()

        stat_word_conf_window = Toplevel()
        stat_word_conf_window.title('关键字统计')
        stat_word_conf_window.geometry('450x250')
        stat_word_conf_window.resizable(0, 0)
        Label(stat_word_conf_window, text="需要统计关键字的字段：").grid(column=0, row=0, padx=8, pady=4)

        edit_box1 = tk.Entry(stat_word_conf_window)
        edit_box1.insert(0, self.__profile.nlp_fields)
        edit_box1.grid(column=2, row=0, )

        Label(stat_word_conf_window, text="需要统计的词性：").grid(column=0, row=1, padx=8, pady=4)
        edit_box2 = tk.Entry(stat_word_conf_window)
        edit_box2.insert(0, self.__profile.poses)
        edit_box2.grid(column=2, row=1, )

        Label(stat_word_conf_window, text="需要统计的词性：").grid(column=0, row=1, padx=8, pady=4)
        edit_box2 = tk.Entry(stat_word_conf_window)
        edit_box2.insert(0, self.__profile.poses)
        edit_box2.grid(column=2, row=1, )

        cb_freq_v = IntVar()
        cb_freq = Checkbutton(stat_word_conf_window, text="Frequency",
                              variable=cb_freq_v, onvalue = 1, offvalue = 0)
        cb_freq.select()
        cb_freq.grid(column=0, row=3, padx=8, pady=4)

        cb_tfidf_v = IntVar()
        cb_tfidf = Checkbutton(stat_word_conf_window, text="TF-IDF",
                               variable=cb_tfidf_v, onvalue = 1, offvalue = 0)
        cb_tfidf.select()
        cb_tfidf.grid(column=0, row=4, padx=8, pady=4)

        cb_textrank_v = IntVar()
        cb_textrank = Checkbutton(stat_word_conf_window, text="TextRank",
                                  variable=cb_textrank_v, onvalue = 1, offvalue = 0)
        cb_textrank.select()
        cb_textrank.grid(column=0, row=5, padx=8, pady=4)


        confirm_button = tk.Button(stat_word_conf_window, text="确定", command=confirm_button_click)
        confirm_button.grid(column=1, row=6, padx=8, pady=4)

    def __init_config(self):
        '''
        Menu中的配置信息
        :return:
        '''
        def confirm_button_click():
            focused_fields = edit_box1.get()
            is_filter = True if cb_isfilter_v.get() else False
            self.__profile.focused_fields = focused_fields
            self.__profile.is_filter = is_filter
            config.Config.store(self.__profile)
            config_window.destroy()

        config_window = Toplevel()
        config_window.title('系统配置')
        config_window.geometry('450x250')
        config_window.resizable(0, 0)

        Label(config_window, text="需要加载的列：").grid(column=0, row=0, padx=8, pady=4)
        edit_box1 = tk.Entry(config_window)
        edit_box1.insert(0, self.__profile.focused_fields)
        edit_box1.grid(column=1, row=0, padx=8, pady=4)

        cb_isfilter_v = IntVar()
        cb_isfilter = Checkbutton(config_window, text="是否过滤看病、取药的数据",
                               variable=cb_isfilter_v, onvalue=1, offvalue=0)
        cb_isfilter.select()
        cb_isfilter.grid(column=0, row=2, padx=8, pady=4)

        confirm_button = tk.Button(config_window, text="确定", command=confirm_button_click)
        confirm_button.grid(column=1, row=3, padx=8, pady=4)

    def __statistic(self):
        if len(self.__original_data) == 0:
            messagebox.showinfo('提示信息', '尚未读取数据，请加载原始的门诊量数据')
            return
        self.__init_stat_config_window()


    def __stat_word(self):
        if len(self.__original_data) == 0:
            messagebox.showinfo('提示信息', '尚未读取数据，请加载原始的门诊量数据')
            return
        self.__init_stat_word_window()

    def __save_file(self):
        if len(self.__original_data) == 0:
            messagebox.showinfo('','尚未加载数据')
            return
        file_path = filedialog.asksaveasfilename(title=u'保存文件',
                                                 filetypes=[('dump','dmp'), ("csv", ".csv")],
                                                 defaultextension='.dmp')
        if file_path.endswith('.csv'):
            OriginalData.save_csv_file(self.__original_data, file_path)
        elif file_path.endswith('.dmp'):
            OriginalData.save_dump_file(self.__original_data, file_path)
        else:
            return

    def __open_csv_files(self):
        '''
        读取多个csv文件
        :return:
        '''
        files = filedialog.askopenfilenames(title='读取数据文件',
                                            filetypes=[('csv', '*.csv')])
        if files and len(files) > 0:
            self.__original_data = OriginalData.read_csv_files(files,
                                                               cols = self.__profile.focused_fields,
                                                               is_filter = self.__profile.is_filter)
            self.__show_data(self.__original_data)

    def __open_dump_file(self):
        '''
        读取序列化的文件
        :return:
        '''
        file = filedialog.askopenfilename(title='读取序列化的数据文件',
                                          filetypes=[('dump',('*.dump', '*.dmp'))])
        if file:
            self.__original_data = OriginalData.read_dump_file(file,
                                                               cols=self.__profile.focused_fields,
                                                               is_filter = self.__profile.is_filter)
            self.__show_data(self.__original_data)

    def __show_data(self,
                    data:pd.DataFrame,
                    col_length = [90, 90, 70, 180, 180, 40]):
        '''
        展示加载的数据
        :param data:需要显示的数据
        :param col_length: 每一列的长度
        :return:
        '''
        self.table['columns'] = data.columns.to_list()
        for c, w in zip(data.columns, col_length):
            self.table.column(c, width=w, anchor='center')
            self.table.heading(c, text=c)

        for i in self.table.get_children():
            # 清空原有的数据内容
            self.table.delete(i)
        for index, row in data.iterrows():
            self.table.insert('', 'end', values=row.to_list())
            if index>100:
                break

        self.scrollbar.config(command=self.table.yview)
        self.table.pack(side=tk.TOP, expand=tk.YES, fill=tk.BOTH)

    def __plot(self, xname, group, agebins=[]):
        class_cols = [Columns.Date]
        if xname != Columns.Age:
            class_cols.append(xname)
        else:
            age_cuts = pd.cut(self.__original_data['AGE'],
                              bins=[int(a) for a in agebins.split(',')], right=False)
            class_cols.append(age_cuts)

        if group != Columns.Age:
            class_cols.append(group)
        else:
            age_cuts = pd.cut(self.__original_data['AGE'],
                              bins=[int(a) for a in agebins.split(',')], right=False)
            class_cols.append(age_cuts)

        vol = VolumeAnalyzer.get_outpatient_volume(self.__original_data, class_cols)
        myfont = FontProperties(fname=r'C:\Windows\Fonts\simhei.ttf', size=14)
        sns.set(font=myfont.get_name(), style='white')

        g = sns.boxplot(x=xname,
                        y=Columns.OutpatientVol,
                        hue=group,
                        data=vol,
                        )
        plt.xticks(rotation=45)
        plt.show()

    def __config(self):
        pass

def data_not_empt(data):
    def my_decorator(func):
        @functools.wraps(func)
        def wrapper(*args, **kwargs):
            if len(data) == 0:
                messagebox.showinfo('提示信息', '尚未读取数据，请加载原始的门诊量数据')
                return
            func(*args, **kwargs)
        return wrapper
    return my_decorator