zikai/cyzg
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

1

Browse Source
This commit is contained in:
王子恺
2025-08-20 16:25:51 +08:00
commit c451fdf0e5
11 changed files with 1632 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
.gitignore vendored Normal file
View File

@@ -0,0 +1,3 @@
/.idea/*
/data.db
/.old/*

195
backend.py Normal file
View File

@@ -0,0 +1,195 @@
import os
import random
import sqlite3
import time
import zipfile
class Backend:
def __init__(self, db="data.db"):
self.db_path = db
# 先解压
if not os.path.exists(self.db_path) and os.path.exists(self.db_path + ".zip"):
with zipfile.ZipFile(self.db_path + ".zip", 'r') as zip_ref:
zip_ref.extractall(".")
self.time = time.time()
self.title_filter = self.get_source_type()
self.global_filter = []
def get_question(self):
conn = sqlite3.connect(self.db_path)
cursor = conn.cursor()
# 构造 IN 子句的占位符
placeholders = ','.join('?' for _ in self.title_filter)
# 总权重查询
total_query = f"""
SELECT SUM(count)
FROM questions
WHERE count > 0 AND title IN ({placeholders})
"""
#print(total_query)
#print(self.title_filter)
cursor.execute(total_query, self.title_filter)
total_weights = cursor.fetchone()[0]
if total_weights is None or total_weights == 0:
conn.close()
return None
# 生成随机数 [0, total_weights)
random_num = random.uniform(0, total_weights)
# 主查询：加权随机选择
query = f"""
SELECT *
FROM (
SELECT *,
SUM(count) OVER (ORDER BY id ROWS UNBOUNDED PRECEDING) AS cum_weight
FROM questions
WHERE count > 0 AND title IN ({placeholders})
)
WHERE ? < cum_weight
ORDER BY cum_weight
LIMIT 1;
"""
params = tuple(self.title_filter) + (random_num,)
cursor.execute(query, params)
row = cursor.fetchone()
print(total_weights, row)
conn.close()
self.time = time.time()
return list(row) if row else None
def update(self, id, state):
conn = sqlite3.connect(self.db_path)
cursor = conn.cursor()
# 作对-1 错了+2
if state != 0:
query = """
UPDATE questions
SET count = count + 2
WHERE id = ?
and count > 0"""
else:
query = """
UPDATE questions
SET count = count - 1
WHERE id = ?
"""
cursor.execute(query, (id,))
# 记录用时
query = "INSERT INTO answers_history (id, time_used, state) VALUES (?, ?, ?)"
cursor.execute(query, (id, time.time() - self.time, state))
conn.commit()
conn.close()
def reset_time(self):
self.time = time.time()
def get_acc(self, top_n=0):
conn = sqlite3.connect(self.db_path)
cursor = conn.cursor()
try:
if top_n > 0:
query = """
SELECT CAST(SUM(CASE WHEN state = 0 THEN 1 ELSE 0 END) AS FLOAT) / COUNT(*) AS accuracy
FROM (SELECT state
FROM answers_history
ORDER BY time DESC
LIMIT ?)
"""
params = (top_n,)
else:
query = """
SELECT CAST(SUM(CASE WHEN state = 0 THEN 1 ELSE 0 END) AS FLOAT) / COUNT(*) AS accuracy FROM answers_history
"""
params = ()
cursor.execute(query, params)
result = cursor.fetchone()[0]
return result * 100 if result is not None else 0.0
except sqlite3.Error as e:
print(f"数据库错误：{e}")
return 0.0
finally:
conn.close()
def get_avg_time(self, top_n=0):
conn = sqlite3.connect(self.db_path)
cursor = conn.cursor()
try:
if top_n > 0:
query = """
SELECT CAST(SUM(time_used) AS FLOAT) / COUNT(*) AS accuracy
FROM (SELECT time_used
FROM answers_history
where time_used is not null
ORDER BY time DESC
LIMIT ?)
"""
params = (top_n,)
else:
query = """
SELECT CAST(SUM(time_used) AS FLOAT) / COUNT(*) AS accuracy FROM answers_history where time_used is not null
"""
params = ()
cursor.execute(query, params)
result = cursor.fetchone()[0]
return result if result is not None else 0.0
except sqlite3.Error as e:
print(f"数据库错误：{e}")
return 0.0
finally:
conn.close()
def get_statistics(self):
avg_all = self.get_avg_time()
avg_50 = self.get_avg_time(50)
avg_100 = self.get_avg_time(100)
result = [
["正确率", f"{self.get_acc():.1f}/60%"],
[],
["最近50题正确率", f"{self.get_acc(50):.1f}/60%"],
["最近100题正确率", f"{self.get_acc(120):.1f}/60%"],
["平均耗时", f"{avg_all:.1f}/72s"],
["预计做完用时", f"{avg_all * 1.667:.1f}/120min"],
["50平均耗时", f"{avg_50:.1f}/72s"],
["预计做完用时", f"{avg_50 * 1.667:.1f}/120min"],
["120平均耗时", f"{avg_100:.1f}/72s"],
["预计做完用时", f"{avg_100 * 1.667:.1f}/120min"],
]
return result
def get_source_type(self):
conn = sqlite3.connect(self.db_path)
cursor = conn.cursor()
query = "SELECT DISTINCT title FROM questions"
cursor.execute(query)
result = cursor.fetchall()
new_list = [item[0] for item in result]
return new_list
def set_config(self,a,b):
self.title_filter = a
self.global_filter = b

4
components/__init__.py Normal file
View File

@@ -0,0 +1,4 @@
from .constants import *
from .page0 import Page0
from .page1 import Page1
from .page2 import Page2

26
components/constants.py Normal file
View File

@@ -0,0 +1,26 @@
import tkinter.font as font
# 定义
BG_COLOR = "#f4f6f9" # 背景色
FG_COLOR = "#2c3e50" # 主文字色
ACCENT_COLOR = "#3498db" # 主按钮色
CORRECT_COLOR = "#d4edda" # 正确答案 淡绿色
MISS_COLOR = "#5bc04b" # 漏选 绿色
WRONG_COLOR = "#f8d7da" # 错误 红色
SELECTED_COLOR = "#d6eaf8" # 选中/悬浮 浅蓝色
FONT_FAMILY = "微软雅黑"
# 映射
indexMap = {6: "A. ", 7: "B. ", 8: "C. ", 9: "D. "}
# 全局
bottom_options = ["全部的题", "未做过的题", "错过的题", ]
instructions = f"""后面的选项仅在'{bottom_options[0]}'未被勾选时起效"""
# 样式
# 创建字体
class Fonts:
def __init__(self):
self.title = font.Font(family=FONT_FAMILY, size=18, weight="bold")
self.button = font.Font(family=FONT_FAMILY, size=14)
self.text = font.Font(family=FONT_FAMILY, size=13)

307
components/page0.py Normal file
View File

@@ -0,0 +1,307 @@
# 答题页
import tkinter as tk
import random
from functools import partial
from components import *
# 选项
class WrappingButton(tk.Frame):
def __init__(self, parent, text="", command=None,title_font=None,button_font=None,text_font=None, **kwargs):
tk.Frame.__init__(self, parent, **kwargs)
self.configure(bg=BG_COLOR, relief="raised", bd=1)
self.title_front = title_font
self.button_font = button_font
self.text_font = text_font
self.label = tk.Label(
self,
text=text,
justify="left",
anchor="w",
wraplength=800,
padx=15,
pady=12,
bg=BG_COLOR,
fg=FG_COLOR,
font=self.button_font
)
self.label.pack(fill="both", expand=True)
# 绑定点击事件
self.bind("<Button-1>", lambda e: command() if command else None)
self.label.bind("<Button-1>", lambda e: command() if command else None)
# 绑定鼠标悬停事件
self.bind("<Enter>", self._on_enter)
self.bind("<Leave>", self._on_leave)
self.label.bind("<Enter>", self._on_enter)
self.label.bind("<Leave>", self._on_leave)
# 当前状态
self.current_style = "Default"
self._state = "normal"
def _on_enter(self, event):
if self._state == "disabled":
return
if self.current_style == "Default":
self.label.configure(bg=SELECTED_COLOR)
self.configure(bg=SELECTED_COLOR)
def _on_leave(self, event):
if self._state == "disabled":
return
if self.current_style == "Default":
self.label.configure(bg=BG_COLOR)
self.configure(bg=BG_COLOR)
def config(self, **kwargs):
if "text" in kwargs:
self.label.config(text=kwargs["text"])
kwargs.pop("text")
if "style" in kwargs:
style = kwargs["style"].split(".")[0]
self.current_style = style
if style == "Default":
self.label.configure(bg=BG_COLOR, fg=FG_COLOR)
self.configure(bg=BG_COLOR)
elif style == "Selected":
self.label.configure(bg=SELECTED_COLOR, fg=FG_COLOR)
self.configure(bg=SELECTED_COLOR)
elif style == "Correct":
self.label.configure(bg=CORRECT_COLOR, fg="#155724")
self.configure(bg=CORRECT_COLOR)
elif style == "Miss":
self.label.configure(bg=MISS_COLOR, fg="#155724")
self.configure(bg=MISS_COLOR)
elif style == "Wrong":
self.label.configure(bg=WRONG_COLOR, fg="#721c24")
self.configure(bg=WRONG_COLOR)
kwargs.pop("style")
if "state" in kwargs:
self._state = kwargs["state"]
if kwargs["state"] == "disabled":
self.label.configure(fg="#999999")
else:
self.label.configure(fg=FG_COLOR)
kwargs.pop("state")
tk.Frame.config(self, **kwargs)
class Page0(tk.Frame):
# 定义
indexMap = {6: "A. ", 7: "B. ", 8: "C. ", 9: "D. "} # 映射
def __init__(self, parent,fonts,c_backend, **kwargs):
tk.Frame.__init__(self, parent, **kwargs)
# 全局变量
self.buttons = []
self.text1 = None
self.text2 = None
self.row = []
self.correct_answer = []
self.my_answer = [0, 0, 0, 0]
self.shuffle_index = [6, 7, 8, 9]
self.flag = False
self.button_disable = True
# 入参
self.c_backend = c_backend
# 题目
frame_top = tk.Frame(parent, bg=BG_COLOR)
frame_top.pack(pady=15, padx=20, fill='x')
self.text1 = tk.Text(
frame_top,
font=fonts.title,
wrap='word',
height=5,
bg="white",
fg=FG_COLOR,
relief="flat",
padx=15,
pady=15,
selectbackground=ACCENT_COLOR,
highlightbackground=BG_COLOR
)
self.text1.config(state="disabled")
self.text1.pack(fill='x', expand=True)
# 按钮
frame_buttons = tk.Frame(parent, bg=BG_COLOR)
frame_buttons.pack(pady=10, padx=20, fill='x')
for i in range(4):
btn = WrappingButton(
frame_buttons,
text=f"选项 {i + 1}",
command=partial(self.on_button_click, i),
title_font=fonts.title,
button_font=fonts.button,
text_font=fonts.text,
)
self.buttons.append(btn)
btn.pack(pady=6, fill='x')
# 确定按钮
button5 = tk.Button(
parent,
text="确定",
font=fonts.button,
bg=ACCENT_COLOR,
fg="white",
command=self.on_ok_click,
relief="flat",
padx=20,
pady=10,
borderwidth=0,
cursor="hand2"
)
button5.pack(pady=12)
def on_enter(e): button5.config(bg="#2980b9")
def on_leave(e): button5.config(bg=ACCENT_COLOR)
button5.bind("<Enter>", on_enter)
button5.bind("<Leave>", on_leave)
# 解析区
frame_bottom = tk.Frame(parent, bg=BG_COLOR)
frame_bottom.pack(pady=10, padx=20, fill='both', expand=True)
self.text2 = tk.Text(
frame_bottom,
font=fonts.text,
wrap='word',
bg="white",
fg="#555",
relief="flat",
padx=15,
pady=15,
spacing3=10,
selectbackground=ACCENT_COLOR
)
self.text2.config(state="disabled")
scroll = tk.Scrollbar(frame_bottom, orient="vertical", command=self.text2.yview)
self.text2.configure(yscrollcommand=scroll.set)
self.text2.pack(side="left", fill='both', expand=True)
scroll.pack(side="right", fill='y')
def _show_result(self, ):
if self.row == None or len(self.row) < 15:
return
elif self.button_disable:
return
else:
self.button_disable = True
self.flag = True
self.text2.config(state="normal")
self.text2.delete("1.0", "end")
self.text2.insert("end", self.row[14])
self.text2.config(state="disabled")
correct = 0
for i in range(4):
state = self.correct_answer[i] + 2 * self.my_answer[i] # 0:未选对, 1:漏选, 2:错选, 3:选对
full_text = self.indexMap.get(self.shuffle_index[i]) + self.row[self.shuffle_index[i]]
if state == 1: # 漏选（正确但未选）→ 绿色
self.buttons[i].config(text=full_text, style="Miss.TButton")
elif state == 2: # 选错 → 红色
self.buttons[i].config(text=full_text, style="Wrong.TButton")
correct = 1
elif state == 3: # 选对 → 淡绿色
self.buttons[i].config(text=full_text, style="Correct.TButton")
else: # 未选且错误 → 不变（保持原始或默认）
self.buttons[i].config(text=full_text, style="Default.TButton")
self.c_backend.update(self.row[0], correct)
self.winfo_toplevel().focus_set()
def init(self):
self.my_answer = [0, 0, 0, 0]
self.correct_answer = []
self.flag = False
self.button_disable = False
self.row = self.c_backend.get_question()
if self.row == None:
print("没有了")
return
# 判断题处理
elif self.row[6] == "" and self.row[7] == "":
self.row[6] = "正确"
self.row[7] = "错误"
self.buttons[2].config(state="disabled")
self.buttons[3].config(state="disabled")
else:
self.buttons[2].config(state="normal")
self.buttons[3].config(state="normal")
# 更新题目
self.text1.config(state="normal")
self.text1.delete("1.0", "end")
self.text1.insert("end", self.row[4] + "\n" + self.row[5])
self.text1.config(state="disabled")
# 清空解析
self.text2.config(state="normal")
self.text2.delete("1.0", "end")
self.text2.config(state="disabled")
# 混洗选项
self.shuffle_index = [6, 7, 8, 9]
random.shuffle(self.shuffle_index)
for i in range(4):
btn_text = self.row[self.shuffle_index[i]].strip()
self.buttons[i].config(
text=f" {btn_text}",
state="normal",
style="Default.TButton"
)
self.correct_answer.append(1 if self.row[self.shuffle_index[i] + 4] == 1 else 0)
self.winfo_toplevel().focus_set()
def on_button_click(self,button_number):
if self.button_disable:
return
if sum(self.correct_answer) > 1:
if self.my_answer[button_number] == 0:
self.my_answer[button_number] = 1
self.buttons[button_number].config(style="Selected.TButton")
else:
self.my_answer[button_number] = 0
self.buttons[button_number].config(style="Default.TButton")
else:
self.my_answer[button_number] = 1
self._show_result()
self.winfo_toplevel().focus_set()
def on_ok_click(self):
if self.flag:
self.init()
else:
self. _show_result()
self.winfo_toplevel().focus_set()
if __name__ == "__main__":
from backend import Backend
root = tk.Tk()
font = Fonts()
c_backend = Backend()
app = Page0(root,font,c_backend)
root.mainloop()

105
components/page1.py Normal file
View File

@@ -0,0 +1,105 @@
import tkinter as tk
from tkinter import ttk
from components import *
class Page1:
def __init__(self, root, top_options, bottom_options):
self.root = root
self.top_options = top_options
self.bottom_options = bottom_options
# ========== 顶部复选框区域 ==========
self.top_frame = tk.Frame(self.root, bg=BG_COLOR, pady=25)
self.top_frame.pack(side=tk.TOP, fill=tk.X)
self.top_vars = []
for option in top_options:
var = tk.BooleanVar(value=True)
chk = ttk.Checkbutton(self.top_frame, text=option, variable=var, style="Custom.TCheckbutton")
chk.pack(side=tk.LEFT, padx=20)
self.top_vars.append(var)
# ========== 分割线 ==========
separator = tk.Frame(self.root, height=1, bg="black")
separator.pack(side=tk.TOP, fill=tk.X, padx=20)
# ========== 底部复选框区域 ==========
self.bottom_frame = tk.Frame(self.root, bg=BG_COLOR, pady=25)
self.bottom_frame.pack(side=tk.TOP, fill=tk.X)
self.bottom_vars = []
for option in bottom_options:
var = tk.BooleanVar(value=True)
chk = ttk.Checkbutton(self.bottom_frame, text=option, variable=var, style="Custom.TCheckbutton")
chk.pack(side=tk.LEFT, padx=20)
self.bottom_vars.append(var)
# ========== 说明文本区域 ==========
text_frame = tk.Frame(self.root, bg=BG_COLOR)
text_frame.pack(side=tk.TOP, padx=40, pady=10, fill=tk.X)
text_widget = tk.Text(
text_frame,
height=6,
width=60,
wrap=tk.WORD,
bg=BG_COLOR,
fg=FG_COLOR,
relief=tk.FLAT,
borderwidth=0,
font=(FONT_FAMILY, 12),
padx=10,
pady=10
)
text_widget.pack(expand=True, fill=tk.X)
text_widget.insert(tk.END, instructions)
text_widget.config(state=tk.DISABLED)
def setup_styles(self):
"""设置 ttk 样式"""
def top_selected(self):
"""返回第一行中被选中的项"""
return [opt for opt, var in zip(self.top_options, self.top_vars) if var.get()]
def bottom_selected(self):
"""返回第二行中被选中的项"""
return [opt for opt, var in zip(self.bottom_options, self.bottom_vars) if var.get()]
if __name__ == "__main__":
top_options = ["1", "2", "3", "4"]
bottom_options = ["全部的题", "未做过的题", "错过的题", ]
root = tk.Tk()
app = Page1(root, top_options, bottom_options)
# 按钮区域
button_frame = tk.Frame(root)
button_frame.pack(side=tk.TOP, pady=20)
# 打印上一行选中项的按钮
btn_print_top = tk.Button(
button_frame,
text="1",
command=lambda: print(app.top_selected()),
font=("Helvetica", 14),
bg="lightblue",
width=20
)
btn_print_top.pack(side=tk.LEFT, padx=10)
# 打印下一行选中项的按钮
btn_print_bottom = tk.Button(
button_frame,
text="2",
command=lambda: print(app.bottom_selected()),
font=("Helvetica", 14),
bg="lightgreen",
width=20
)
btn_print_bottom.pack(side=tk.LEFT, padx=10)
root.mainloop()

98
components/page2.py Normal file
View File

@@ -0,0 +1,98 @@
# 统计数据页
import tkinter as tk
from tkinter import ttk
class Page2:
def __init__(self, root):
self.root = root
# 创建主框架
main_frame = ttk.Frame(root, padding="20")
main_frame.grid(row=0, column=0, sticky=(tk.W, tk.E, tk.N, tk.S))
main_frame.columnconfigure(0, weight=1)
main_frame.rowconfigure(1, weight=1)
# 标题
title_label = ttk.Label(main_frame, text="当前数据统计", font=("Arial", 16, "bold"))
title_label.grid(row=0, column=0, columnspan=4, pady=(0, 20))
# === 关键指标区域 ===
self.kpi_data = []
# KPI 容器
self.kpi_container = ttk.Frame(main_frame)
self.kpi_container.grid(row=1, column=0, sticky=(tk.W, tk.E, tk.N, tk.S))
for i in range(4):
self.kpi_container.columnconfigure(i, weight=1)
self.value_labels = []
self.create_dashboard_units_4col()
def create_kpi_unit(self, parent, name, value, row, col):
"""创建单个单元"""
unit_frame = ttk.Frame(parent, relief="solid", borderwidth=1, padding=10)
unit_frame.grid(row=row, column=col, padx=8, pady=8, sticky="nsew")
# 设置行权重，使内容垂直居中
parent.rowconfigure(row, weight=1)
if name and value:
name_label = ttk.Label(unit_frame, text=f"{name}:", font=("Arial", 11))
name_label.pack(anchor="w")
value_label = ttk.Label(unit_frame, text=value, font=("Arial", 11, "bold"))
value_label.pack(anchor="e")
return value_label
else:
# 如果 name 或 value 为空，显示占位符或空白
ttk.Label(unit_frame, text="—", foreground="gray").pack()
return None
def create_dashboard_units_4col(self):
"""创建4列布局的单元"""
rows = (len(self.kpi_data) + 3) // 4 # 向上取整
for r in range(rows):
self.kpi_container.rowconfigure(r, weight=1)
for idx, item in enumerate(self.kpi_data):
row_idx = idx // 4
col_idx = idx % 4
if item and len(item) >= 2:
name, value = item[0], item[1]
else:
name, value = None, None
value_label = self.create_kpi_unit(self.kpi_container, name, value, row_idx, col_idx)
self.value_labels.clear()
self.value_labels.append(value_label)
def update_data(self, new_values):
self.kpi_data = new_values
self.create_dashboard_units_4col()
if __name__ == "__main__":
import backend
import random
# 实例化后端
c_backend = backend.Backend()
# 创建 GUI
root = tk.Tk()
root.title("统计面板")
root.geometry("800x600")
app = Page2(root)
def on_refresh():
s = c_backend.get_statistics()
for v in s:
if v is not None and len(v) >= 2:
v[1] = str(random.randint(1,100))
app.update_data(s)
# 刷新按钮
refresh_btn = ttk.Button(root, text="刷新数据", command=on_refresh)
refresh_btn.grid(row=2, column=0, pady=10)
root.mainloop()

388
crawler/getLeCheng_chapter.ipynb Normal file
View File

@@ -0,0 +1,388 @@
{
"cells": [
{
"cell_type": "code",
"id": "initial_id",
"metadata": {
"collapsed": true,
"ExecuteTime": {
"end_time": "2025-08-19T01:03:22.758571Z",
"start_time": "2025-08-19T01:03:22.753008Z"
}
},
"source": [
"import time\n",
"\n",
"from selenium import webdriver\n",
"from selenium.webdriver.edge.service import Service\n",
"from selenium.webdriver.common.by import By\n",
"from selenium.webdriver.support.ui import WebDriverWait\n",
"from selenium.webdriver.support import expected_conditions as EC\n",
"from selenium.webdriver.edge.options import Options"
],
"outputs": [],
"execution_count": 49
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T01:03:23.224371Z",
"start_time": "2025-08-19T01:03:23.220216Z"
}
},
"cell_type": "code",
"source": [
"from bs4 import BeautifulSoup\n",
"import sqlite3"
],
"id": "59b26d9f105eae85",
"outputs": [],
"execution_count": 50
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T01:03:25.179818Z",
"start_time": "2025-08-19T01:03:25.173558Z"
}
},
"cell_type": "code",
"source": "db_path = '../data.db'",
"id": "37a70656848ceced",
"outputs": [],
"execution_count": 51
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T01:03:25.713012Z",
"start_time": "2025-08-19T01:03:25.704775Z"
}
},
"cell_type": "code",
"source": [
"conn = sqlite3.connect(db_path)\n",
"conn.execute('''CREATE TABLE \"questions\"\n",
"(\n",
" id INTEGER\n",
" constraint questions_pk\n",
" primary key autoincrement,\n",
" title TEXT,\n",
" chapter TEXT,\n",
" q_num text,\n",
" q_type text,\n",
" question TEXT not null,\n",
" a TEXT not null,\n",
" b TEXT not null,\n",
" c TEXT not null,\n",
" d TEXT not null,\n",
" a_result BLOB default false,\n",
" b_result BLOB default false,\n",
" c_result BLOB default false,\n",
" d_result BLOB default false,\n",
" explanation TEXT,\n",
" count integer default 3 not null\n",
")''')\n",
"\n",
"conn.execute('''CREATE TABLE \"answers_history\"\n",
"(\n",
" id INTEGER not null\n",
" constraint answers_history__questions_id_fk\n",
" references questions,\n",
" time_used INTEGER,\n",
" state INTEGER,\n",
" time text default CURRENT_TIMESTAMP\n",
")''')\n",
"\n",
"conn.execute('''CREATE TABLE url\n",
" (\n",
" id INTEGER not null,\n",
" url TEXT\n",
" )''')\n",
"\n",
"\n",
"conn.commit()\n"
],
"id": "d70a270099e8b056",
"outputs": [],
"execution_count": 52
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T01:03:27.430817Z",
"start_time": "2025-08-19T01:03:27.423603Z"
}
},
"cell_type": "code",
"source": [
"edge_options = Options()\n",
"#edge_options.add_argument(\"--headless\") # 可选：无界面模式\n",
"edge_options.add_argument(\"--disable-gpu\")\n",
"edge_options.add_argument(\"--no-sandbox\")\n",
"edge_options.add_argument(\"--disable-extensions\")\n",
"edge_options.add_argument(\"--disable-plugins\")\n",
"edge_options.add_argument(\"--disable-popup-blocking\")\n",
"edge_options.add_argument(\"--disable-infobars\")\n",
"edge_options.add_argument(\"--disable-notifications\")\n",
"edge_options.add_argument(\"--no-first-run\")\n",
"edge_options.add_argument(\"--no-default-browser-check\")\n",
"\n",
"user_data_dir = r\"D:\\code\\edge\"\n",
"edge_options.add_argument(f\"--user-data-dir={user_data_dir}\")\n",
"# 指定配置文件（可选，默认是 Default）\n",
"edge_options.add_argument(\"--profile-directory=Default\")"
],
"id": "e4a35062c4549f44",
"outputs": [],
"execution_count": 53
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T01:03:30.978615Z",
"start_time": "2025-08-19T01:03:28.414779Z"
}
},
"cell_type": "code",
"source": [
"# 指定 EdgeDriver 路径（可选，若已配置环境变量可省略）\n",
"service = Service(executable_path=r\"D:\\app\\edgeDriver\\msedgedriver.exe\")\n",
"# 创建 Edge 浏览器实例\n",
"driver = webdriver.Edge(service=service, options=edge_options)"
],
"id": "9b48ddaca80598aa",
"outputs": [],
"execution_count": 54
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T01:00:42.177993Z",
"start_time": "2025-08-19T01:00:42.171173Z"
}
},
"cell_type": "code",
"source": [
"def get_web(url):\n",
" driver.get(url)\n",
"\n",
" # 等待页面渲染完成（例如等待 body 加载）\n",
" wait = WebDriverWait(driver, 720)\n",
" wait.until(EC.presence_of_element_located((By.TAG_NAME, \"body\")))\n",
" time.sleep(3)\n",
"\n",
" #进入背题模式\n",
" clickable_element = wait.until(EC.element_to_be_clickable((By.CSS_SELECTOR, \".setting-type.iconfont.icon-setting\")))\n",
" clickable_element.click()\n",
" wait.until(\n",
" EC.element_to_be_clickable((By.CSS_SELECTOR, \".question-setting-button.ant-btn.ant-btn-default\"))).click()\n",
"\n",
" # 获取渲染后的 HTML\n",
" rendered_html = driver.page_source\n",
" return rendered_html"
],
"id": "2b02063fec8abbdd",
"outputs": [],
"execution_count": 43
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T01:00:42.212116Z",
"start_time": "2025-08-19T01:00:42.206610Z"
}
},
"cell_type": "code",
"source": [
"def list_get(lst, index, default=\"\"):\n",
" try:\n",
" return lst[index]\n",
" except IndexError:\n",
" return default"
],
"id": "de9650bb0e005d4a",
"outputs": [],
"execution_count": 44
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T01:00:42.247110Z",
"start_time": "2025-08-19T01:00:42.237114Z"
}
},
"cell_type": "code",
"source": [
"def write2db(rendered_html, info):\n",
" # 解析web并登记\n",
" soup = BeautifulSoup(rendered_html, 'html.parser')\n",
" full_left = soup.find_all('div', class_='pull-left')\n",
" for questions in full_left:\n",
" for question in questions.children:\n",
" title_info = [] #num,type,question\n",
" answers_info = []\n",
" answers_correct_info = []\n",
" explain_info = \"\"\n",
"\n",
" # 标题信息\n",
" for title in question.find_all('div', class_='p-stem'):\n",
" for element in title.children:\n",
" title_info.append(element.text)\n",
"\n",
" # 题目信息\n",
" for answer in question.find_all('div', class_='answer-ul'):\n",
" for element in answer.find_all(\"div\", recursive=False):\n",
" # 答案\n",
" if \"answer\" in element.get(\"class\"):\n",
" answers_correct_info.append(True)\n",
" else:\n",
" answers_correct_info.append(False)\n",
"\n",
" # 问题\n",
" text_elements = element.select(\"div > div > div > div > p\")\n",
" for text_element in text_elements:\n",
" if text_element.text is not None and text_element.text != \"\":\n",
" answers_info.append(text_element.text)\n",
"\n",
" # 解析\n",
" for explain in question.find_all('div', class_='practise-answer-text'):\n",
" explain_info += str(explain.get_text(strip=True))\n",
"\n",
" cursor = conn.execute(\n",
" \"INSERT INTO questions (title, chapter, q_num, q_type, question, a, b, c, d, a_result, b_result, c_result, d_result, explanation) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)\",\n",
" (\n",
" info[1],\n",
" info[2],\n",
" list_get(title_info,0),\n",
" list_get(title_info,1),\n",
" list_get(title_info,2),\n",
" list_get(answers_info,0),\n",
" list_get(answers_info,1),\n",
" list_get(answers_info,2),\n",
" list_get(answers_info,3),\n",
" list_get(answers_correct_info,0,False),\n",
" list_get(answers_correct_info,1,False),\n",
" list_get(answers_correct_info,2,False),\n",
" list_get(answers_correct_info,3,False),\n",
" explain_info,\n",
" )\n",
" )\n",
" inserted_id = cursor.lastrowid\n",
" conn.execute(\n",
" \"INSERT INTO url (id, url) VALUES (?, ?)\",\n",
" (inserted_id, info[0], )\n",
" )\n",
" conn.commit()"
],
"id": "c28a23cbd84f6ea0",
"outputs": [],
"execution_count": 45
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T01:00:42.273569Z",
"start_time": "2025-08-19T01:00:42.265569Z"
}
},
"cell_type": "code",
"source": [
"bg_infos = [\n",
" [\"期货乐橙章节练习\", 1, 1414, 1],\n",
" [\"期货乐橙章节练习\", 2, 1419, 3],\n",
" [\"期货乐橙章节练习\", 3, 1448, 2],\n",
" [\"期货乐橙章节练习\", 4, 1485, 2],\n",
" [\"期货乐橙章节练习\", 5, 1523, 2],\n",
" [\"期货乐橙章节练习\", 6, 1543, 2],\n",
"]"
],
"id": "f8ed3be15b2a69a7",
"outputs": [],
"execution_count": 46
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T01:00:42.311569Z",
"start_time": "2025-08-19T01:00:42.302568Z"
}
},
"cell_type": "code",
"source": [
"def main():\n",
" try:\n",
" for bg_info in bg_infos:\n",
" for index in range(1, bg_info[3]+1):\n",
" url = f\"https://www.bestlec.com/practise/practise?title=%E9%A1%BA%E5%BA%8F%E7%BB%83%E4%B9%A0&qBankId=39&qBankTitle=%E3%80%90%E6%9C%9F%E8%B4%A7%E6%B3%95%E8%A7%84%E3%80%91%E7%AB%A0%E8%8A%82%E7%BB%83%E4%B9%A0&chapterId={bg_info[2]}&practise=1&type=practise&selectSec={index}\"\n",
" rendered_html = get_web(url)\n",
" write2db(rendered_html, [url, bg_info[0], bg_info[1]])\n",
" except Exception as e:\n",
" print(\"error: \" + e)\n",
" finally:\n",
" try:\n",
" conn.close()\n",
" except Exception as e:\n",
" print(\"db:\", e)\n",
"\n",
" try:\n",
" driver.quit()\n",
" except Exception as e:\n",
" print(\"brother:\", e)"
],
"id": "fcfc560b46c29aaa",
"outputs": [],
"execution_count": 47
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T01:01:51.670165Z",
"start_time": "2025-08-19T01:00:42.337618Z"
}
},
"cell_type": "code",
"source": [
"if __name__ == '__main__':\n",
" main()"
],
"id": "811c9d3647c46f8b",
"outputs": [],
"execution_count": 48
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T01:01:51.740128Z",
"start_time": "2025-08-19T01:01:51.737199Z"
}
},
"cell_type": "code",
"source": "",
"id": "5224515d66fe0b",
"outputs": [],
"execution_count": null
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 2
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython2",
"version": "2.7.6"
}
},
"nbformat": 4,
"nbformat_minor": 5
}

415
crawler/getLeCheng_exam.ipynb Normal file
View File

@@ -0,0 +1,415 @@
{
"cells": [
{
"cell_type": "code",
"id": "initial_id",
"metadata": {
"collapsed": true,
"ExecuteTime": {
"end_time": "2025-08-19T02:05:36.929261Z",
"start_time": "2025-08-19T02:05:36.924419Z"
}
},
"source": [
"import time\n",
"\n",
"from selenium import webdriver\n",
"from selenium.webdriver.edge.service import Service\n",
"from selenium.webdriver.common.by import By\n",
"from selenium.webdriver.support.ui import WebDriverWait\n",
"from selenium.webdriver.support import expected_conditions as EC\n",
"from selenium.webdriver.edge.options import Options"
],
"outputs": [],
"execution_count": 86
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T02:05:37.477084Z",
"start_time": "2025-08-19T02:05:37.467743Z"
}
},
"cell_type": "code",
"source": [
"from bs4 import BeautifulSoup\n",
"import sqlite3"
],
"id": "59b26d9f105eae85",
"outputs": [],
"execution_count": 87
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T02:05:38.051715Z",
"start_time": "2025-08-19T02:05:38.047420Z"
}
},
"cell_type": "code",
"source": "db_path = '../data.db'",
"id": "37a70656848ceced",
"outputs": [],
"execution_count": 88
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T02:05:38.522754Z",
"start_time": "2025-08-19T02:05:38.515572Z"
}
},
"cell_type": "code",
"source": [
"conn = sqlite3.connect(db_path)\n",
"conn.execute('''CREATE TABLE \"questions\"\n",
"(\n",
" id INTEGER\n",
" constraint questions_pk\n",
" primary key autoincrement,\n",
" title TEXT,\n",
" chapter TEXT,\n",
" q_num text,\n",
" q_type text,\n",
" question TEXT not null,\n",
" a TEXT not null,\n",
" b TEXT not null,\n",
" c TEXT not null,\n",
" d TEXT not null,\n",
" a_result BLOB default false,\n",
" b_result BLOB default false,\n",
" c_result BLOB default false,\n",
" d_result BLOB default false,\n",
" explanation TEXT,\n",
" count integer default 3 not null\n",
")''')\n",
"\n",
"conn.execute('''CREATE TABLE \"answers_history\"\n",
"(\n",
" id INTEGER not null\n",
" constraint answers_history__questions_id_fk\n",
" references questions,\n",
" time_used INTEGER,\n",
" state INTEGER,\n",
" time text default CURRENT_TIMESTAMP\n",
")''')\n",
"\n",
"conn.execute('''CREATE TABLE url\n",
" (\n",
" id INTEGER not null,\n",
" url TEXT\n",
" )''')\n",
"\n",
"\n",
"conn.commit()\n"
],
"id": "d70a270099e8b056",
"outputs": [],
"execution_count": 89
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T02:05:39.660112Z",
"start_time": "2025-08-19T02:05:39.654188Z"
}
},
"cell_type": "code",
"source": [
"edge_options = Options()\n",
"#edge_options.add_argument(\"--headless\") # 可选：无界面模式\n",
"edge_options.add_argument(\"--disable-gpu\")\n",
"edge_options.add_argument(\"--no-sandbox\")\n",
"edge_options.add_argument(\"--disable-extensions\")\n",
"edge_options.add_argument(\"--disable-plugins\")\n",
"edge_options.add_argument(\"--disable-popup-blocking\")\n",
"edge_options.add_argument(\"--disable-infobars\")\n",
"edge_options.add_argument(\"--disable-notifications\")\n",
"edge_options.add_argument(\"--no-first-run\")\n",
"edge_options.add_argument(\"--no-default-browser-check\")\n",
"\n",
"user_data_dir = r\"D:\\code\\edge\"\n",
"edge_options.add_argument(f\"--user-data-dir={user_data_dir}\")\n",
"# 指定配置文件（可选，默认是 Default）\n",
"edge_options.add_argument(\"--profile-directory=Default\")"
],
"id": "e4a35062c4549f44",
"outputs": [],
"execution_count": 90
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T02:05:42.646764Z",
"start_time": "2025-08-19T02:05:40.137721Z"
}
},
"cell_type": "code",
"source": [
"# 指定 EdgeDriver 路径（可选，若已配置环境变量可省略）\n",
"service = Service(executable_path=r\"D:\\app\\edgeDriver\\msedgedriver.exe\")\n",
"# 创建 Edge 浏览器实例\n",
"driver = webdriver.Edge(service=service, options=edge_options)"
],
"id": "9b48ddaca80598aa",
"outputs": [],
"execution_count": 91
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T01:58:29.792348Z",
"start_time": "2025-08-19T01:58:29.786681Z"
}
},
"cell_type": "code",
"source": [
"def get_web(url):\n",
" driver.get(url)\n",
"\n",
" # 等待页面渲染完成（例如等待 body 加载）\n",
" wait = WebDriverWait(driver, 720)\n",
" wait.until(EC.presence_of_element_located((By.TAG_NAME, \"body\")))\n",
" time.sleep(3)\n",
"\n",
" #进入背题模式\n",
" # clickable_element = wait.until(EC.element_to_be_clickable((By.CSS_SELECTOR, \".setting-type.iconfont.icon-setting\")))\n",
" # clickable_element.click()\n",
" # wait.until(\n",
" # EC.element_to_be_clickable((By.CSS_SELECTOR, \".question-setting-button.ant-btn.ant-btn-default\"))).click()\n",
"\n",
" # 获取渲染后的 HTML\n",
" rendered_html = driver.page_source\n",
" return rendered_html"
],
"id": "2b02063fec8abbdd",
"outputs": [],
"execution_count": 80
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T01:58:30.323362Z",
"start_time": "2025-08-19T01:58:30.317695Z"
}
},
"cell_type": "code",
"source": [
"def list_get(lst, index, default=\"\"):\n",
" try:\n",
" return lst[index]\n",
" except IndexError:\n",
" return default"
],
"id": "de9650bb0e005d4a",
"outputs": [],
"execution_count": 81
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T01:58:30.884914Z",
"start_time": "2025-08-19T01:58:30.871279Z"
}
},
"cell_type": "code",
"source": [
"def write2db(rendered_html, info):\n",
" # 解析web并登记\n",
" soup = BeautifulSoup(rendered_html, 'html.parser')\n",
" full_left = soup.find_all('div', class_='pull-left')\n",
"\n",
" for questions in full_left:\n",
" for question in questions.children:\n",
" title_info = [] #num,type,question\n",
" answers_info = []\n",
" answers_correct_info = [False,False,False,False,]\n",
" explain_info = \"\"\n",
"\n",
" # 标题信息\n",
" for title in question.find_all('div', class_='p-stem'):\n",
" for element in title.children:\n",
" title_info.append(element.text)\n",
"\n",
"\n",
" # 题目信息\n",
" for answer in question.find_all('div', class_='answerClass'):\n",
" # 答案\n",
"\n",
"\n",
" # 带选项\n",
" text_elements = answer.select(\"div > div > p\")\n",
" for text_element in text_elements:\n",
" if text_element.text is not None and text_element.text != \"\":\n",
" answers_info.append(text_element.text)\n",
"\n",
"\n",
" # 解析\n",
" explains = question.find_all('div', class_='practise-answer-text')\n",
"\n",
" try:\n",
" for i in explains[0].text:\n",
" match i:\n",
" case \"A\":\n",
" answers_correct_info[0] = True\n",
" case \"B\":\n",
" answers_correct_info[1] = True\n",
" case \"C\":\n",
" answers_correct_info[2] = True\n",
" case \"D\":\n",
" answers_correct_info[3] = True\n",
" case _ :\n",
" print(i)\n",
" pass\n",
" except Exception as e:\n",
" # e.with_traceback()\n",
" # print(title_info)\n",
" pass\n",
"\n",
" if len(explains[0]) == 0:\n",
" print(\"0 answers found!!\", title_info)\n",
"\n",
" for explain in explains:\n",
" explain_info += str(explain.get_text(strip=True))\n",
"\n",
"\n",
" cursor = conn.execute(\n",
" \"INSERT INTO questions (title, chapter, q_num, q_type, question, a, b, c, d, a_result, b_result, c_result, d_result, explanation) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)\",\n",
" (\n",
" info[1],\n",
" info[2],\n",
" list_get(title_info,0),\n",
" list_get(title_info,1),\n",
" list_get(title_info,2),\n",
" list_get(answers_info,0),\n",
" list_get(answers_info,1),\n",
" list_get(answers_info,2),\n",
" list_get(answers_info,3),\n",
" list_get(answers_correct_info,0,False),\n",
" list_get(answers_correct_info,1,False),\n",
" list_get(answers_correct_info,2,False),\n",
" list_get(answers_correct_info,3,False),\n",
" explain_info,\n",
" )\n",
" )\n",
" inserted_id = cursor.lastrowid\n",
" conn.execute(\n",
" \"INSERT INTO url (id, url) VALUES (?, ?)\",\n",
" (inserted_id, info[0], )\n",
" )\n",
" conn.commit()"
],
"id": "c28a23cbd84f6ea0",
"outputs": [],
"execution_count": 82
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T01:58:31.409810Z",
"start_time": "2025-08-19T01:58:31.404300Z"
}
},
"cell_type": "code",
"source": [
"ttt = \"期货乐橙假题\"\n",
"bg_infos = [\n",
" [ttt, 0, 1385, 1],\n",
" [ttt, 0, 1392, 1],\n",
" [ttt, 0, 1393, 1],\n",
" [ttt, 0, 1394, 1],\n",
" [ttt, 0, 1395, 1],\n",
"]"
],
"id": "f8ed3be15b2a69a7",
"outputs": [],
"execution_count": 83
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T01:58:32.106767Z",
"start_time": "2025-08-19T01:58:32.099456Z"
}
},
"cell_type": "code",
"source": [
"def main():\n",
" try:\n",
" for bg_info in bg_infos:\n",
" for index in range(1, bg_info[3]+1):\n",
" #url = f\"https://www.bestlec.com/practise/practise?title=%E9%A1%BA%E5%BA%8F%E7%BB%83%E4%B9%A0&qBankId=39&qBankTitle=%E3%80%90%E6%9C%9F%E8%B4%A7%E6%B3%95%E8%A7%84%E3%80%91%E7%AB%A0%E8%8A%82%E7%BB%83%E4%B9%A0&chapterId={bg_info[2]}&practise=1&type=practise&selectSec={index}\"\n",
"\n",
" url = f\"https://www.bestlec.com/practise/practise?id={bg_info[2]}&qBankTitle=%E3%80%90%E6%9C%9F%E8%B4%A7%E6%B3%95%E8%A7%84%E3%80%91%E5%8E%86%E5%B9%B4%E7%9C%9F%E9%A2%98&title=%E7%9C%9F%E9%A2%98%E8%80%83%E8%AF%95&type=test&testType=search_paper\"\n",
"\n",
" rendered_html = get_web(url)\n",
" write2db(rendered_html, [url, bg_info[0], bg_info[1]])\n",
" except Exception as e:\n",
" print(\"error: \" + e)\n",
" finally:\n",
" try:\n",
" conn.close()\n",
" except Exception as e:\n",
" print(\"db:\", e)\n",
"\n",
" try:\n",
" driver.quit()\n",
" except Exception as e:\n",
" print(\"brother:\", e)"
],
"id": "fcfc560b46c29aaa",
"outputs": [],
"execution_count": 84
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T01:59:02.604021Z",
"start_time": "2025-08-19T01:58:32.772829Z"
}
},
"cell_type": "code",
"source": [
"if __name__ == '__main__':\n",
" main()"
],
"id": "811c9d3647c46f8b",
"outputs": [],
"execution_count": 85
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-08-19T01:59:03.138625Z",
"start_time": "2025-08-19T01:59:03.135518Z"
}
},
"cell_type": "code",
"source": "",
"id": "5224515d66fe0b",
"outputs": [],
"execution_count": null
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 2
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython2",
"version": "2.7.6"
}
},
"nbformat": 4,
"nbformat_minor": 5
}

BIN
data.db.zip Normal file
View File

Binary file not shown.

91
main.py Normal file
View File

@@ -0,0 +1,91 @@
import tkinter as tk
from tkinter import ttk
from backend import Backend
from components import *
# 后端
c_backend = Backend()
# 创建主窗口
root = tk.Tk()
root.title("zrx")
root.geometry("1000x900")
root.configure(bg=BG_COLOR)
# 字体文件
fonts = Fonts()
notebook = ttk.Notebook(root)
notebook.pack(fill='both', expand=True, padx=10, pady=10)
style = ttk.Style()
style.theme_use("winnative")
# 自定义复选框样式
style.configure(
"Custom.TCheckbutton",
font=(FONT_FAMILY, 14),
background=BG_COLOR,
foreground=FG_COLOR,
padding=8
)
style.map(
"Custom.TCheckbutton",
background=[("active", BG_COLOR)],
foreground=[("active", ACCENT_COLOR)]
)
# 页面0
frame0 = ttk.Frame(notebook, padding=20)
notebook.add(frame0, text="主页")
page0 = Page0(frame0,fonts,c_backend)
page0.init()
# 页面1
frame1 = ttk.Frame(notebook, padding=20)
notebook.add(frame1, text="设置")
top_options = c_backend.get_source_type()
page1 = Page1(frame1,top_options,bottom_options)
# 页面2
frame2 = ttk.Frame(notebook, padding=20)
notebook.add(frame2, text="统计")
page2 = Page2(frame2)
# 事件
def page0_event():
c_backend.reset_time()
c_backend.set_config(
page1.top_selected(),
page1.bottom_selected()
)
def page1_event():
pass
def page2_event():
r = c_backend.get_statistics()
print(r)
page2.update_data(r)
tab_handlers = [
page0_event,
page1_event,
page2_event,
]
def on_space_pressed(event):
page0.init()
def on_tab_changed(event):
try:
tab_handlers[notebook.index("current")]()
except Exception as e:
print(e)
# 绑定事件
root.bind('<space>', on_space_pressed)
notebook.bind("<<NotebookTabChanged>>", on_tab_changed)
# 运行主循环
root.mainloop()