Manually labeled record pairs

annotator.py

import tkinter as tk
from tkinter.filedialog import askopenfilename
from tkinter import messagebox
from tempfile import gettempdir
from tkinter import ttk
import json
import pickle


class AnnotationGenerator(object):
    def __init__(self, path):
        with open(path) as f:
            self.annotation_data = json.load(f)

    def annotation_generator(self):
        for pair in self.annotation_data['pairs']:
            records = [pair['identifiers'][ids]["record"] for ids in pair['identifiers']]
            rows_container = []
            for field in pair['fields']:
                field_name = field["a"]['name']
                field_designators = [k for k in field.keys() if k!="similarity"]
                values = [val for val in [field[k]["value"] for k in field_designators]]
                if len(set(values)) == 1:
                    continue
                # TODO
                # форматтер значень -- дата як дата, None як ""
                rows_container.append([field_name] + values)
            yield {"rows": rows_container, "keys": records}


class Application(tk.Frame):
    def __init__(self, master=None):
        super().__init__(master)
        self.master = master
        self.master.resizable(width=False, height=False)
        self.create_widgets()
        self.json_path = None
        self.pack(fill='both', expand=True)
        self.current_key_pair = None
        self.id_store = []

    def open_file(self):
        path = askopenfilename(initialdir = gettempdir(),
                               filetypes=(("JSON", "*.json"), ("All Files", "*.*")),
                               title="Choose a file.")
        self.dataset.delete(0, tk.END)
        self.dataset.insert(0, path)
        if not path:
            return
        self.json_path = path
        ag = AnnotationGenerator(self.json_path)
        self.generator = ag.annotation_generator()


    def create_widgets(self):
        self.dataset = tk.Entry(self)
        self.dataset.grid(row=0, column=0, columnspan=3)
        self.terminal_scrollbar = tk.Scrollbar(root)
        self.file_open = tk.Button(self, command=self.open_file,
                                   text="...").grid(row=0, column=3)
        self.process = tk.Button(self, text="Start",
                                 command=self.open_compare).grid(row=2, column=2)
        self.quit = tk.Button(
            self, text="Quit",command=self.save_and_quit).grid(
            row=2, column=1,pady=5, padx=5)

    def save_exchange(self):
        if self.id_store:
            with open('codes_exchange.pickle', 'wb') as f:
                pickle.dump(self.id_store, f)


    def save_and_quit(self):
        self.save_exchange()
        self.master.destroy()


    def open_compare(self):
        self.compare = tk.Toplevel()
        self.compare.geometry("650x350")
        self.compare.grid_rowconfigure(0, weight=1)
        self.compare.grid_columnconfigure(0, weight=1)
        cols = ('Атрибут', 'Значення1', 'Значення2')
        # Treeview as table
        self.listBox = ttk.Treeview(self.compare, columns=cols, show='headings')
        self.listBox.configure(yscrollcommand=self.terminal_scrollbar.set)
        #self.listBox.columnconfigure(0, weight=1)
        #self.listBox.rowconfigure(0, weight=1)
        self.listBox.grid(row=0, column=0, columnspan=4,
                          pady=5, padx=2, sticky=tk.NSEW)
        # column headings
        for col in cols:
            self.listBox.heading(col, text=col)
        match = tk.Button(self.compare, text="MATCH", width=15, bg="green",
                          activebackground="green", command=self.click_match)
        distinct = tk.Button(self.compare, text="DISTINCT", width=15, bg="red",
                             activebackground="red", command=self.click_distinct)
        save_data = tk.Button(self.compare, text="SAVE", width=15,
                              command=self.save_exchange)
        save_data.grid(row=4, column=0, pady=5, sticky=tk.W)
        match.grid(row=4, column=1, pady=5, sticky=tk.W)
        distinct.grid(row=4, column=2, pady=5)
        self.show()


    def _clear_treeview(self):
        self.listBox.delete(*self.listBox.get_children())

    def click_distinct(self):
        # print(self.current_key_pair, "loss")
        self.show()

    def click_match(self):
        # print(self.current_key_pair, "stored")
        self.id_store.append(self.current_key_pair)
        self.show()

    def show(self):
        try:
            # {"rows": rows_container, "keys": records}
            table_data = next(self.generator)
            self.current_key_pair = table_data["keys"]
        except StopIteration:
            self.save_exchange()
            self.compare.destroy()
            while tk.messagebox.showinfo(
                    title="The END", message="Нарешті все") != "ok":
                return True
            self.master.destroy()
            return
        # childrens = self.listBox.get_children()
        self._clear_treeview()
        for (index, row) in enumerate(table_data["rows"], 1):
            tag = 'even' if (index % 2 == 0) else 'odd'
            self.listBox.insert("", "end", values=(row), tags=(tag,))
        # return {'status': 0, "keys": table_data['keys']}
        self.listBox.tag_configure('even', background='#E8E8E8')


root = tk.Tk()
app = Application(master=root)
app.mainloop()