feat(CourseInformer): 添加课程通知系统
- 实现了读取教务系统导出的xls课程安排以及通过Napcat实例发送对应通知的功能 - 添加了通知配置文件和测试脚本 - 创建了项目结构和必要的配置文件
This commit is contained in:
3
.gitignore
vendored
Normal file
3
.gitignore
vendored
Normal file
@@ -0,0 +1,3 @@
|
||||
/resources/教学安排表.xls
|
||||
/CLAUDE.md
|
||||
/.idea/misc.xml
|
||||
9
.idea/.gitignore
generated
vendored
Normal file
9
.idea/.gitignore
generated
vendored
Normal file
@@ -0,0 +1,9 @@
|
||||
# 默认忽略的文件
|
||||
/shelf/
|
||||
/workspace.xml
|
||||
# 基于编辑器的 HTTP 客户端请求
|
||||
/httpRequests/
|
||||
# Datasource local storage ignored files
|
||||
/dataSources/
|
||||
/dataSources.local.xml
|
||||
/.name
|
||||
10
.idea/CourseInformer.iml
generated
Normal file
10
.idea/CourseInformer.iml
generated
Normal file
@@ -0,0 +1,10 @@
|
||||
<?xml version="1.0" encoding="UTF-8"?>
|
||||
<module type="PYTHON_MODULE" version="4">
|
||||
<component name="NewModuleRootManager">
|
||||
<content url="file://$MODULE_DIR$">
|
||||
<excludeFolder url="file://$MODULE_DIR$/.venv" />
|
||||
</content>
|
||||
<orderEntry type="jdk" jdkName="Python 3.13 (ClassInfoPublisher)" jdkType="Python SDK" />
|
||||
<orderEntry type="sourceFolder" forTests="false" />
|
||||
</component>
|
||||
</module>
|
||||
6
.idea/inspectionProfiles/profiles_settings.xml
generated
Normal file
6
.idea/inspectionProfiles/profiles_settings.xml
generated
Normal file
@@ -0,0 +1,6 @@
|
||||
<component name="InspectionProjectProfileManager">
|
||||
<settings>
|
||||
<option name="USE_PROJECT_PROFILE" value="false" />
|
||||
<version value="1.0" />
|
||||
</settings>
|
||||
</component>
|
||||
8
.idea/modules.xml
generated
Normal file
8
.idea/modules.xml
generated
Normal file
@@ -0,0 +1,8 @@
|
||||
<?xml version="1.0" encoding="UTF-8"?>
|
||||
<project version="4">
|
||||
<component name="ProjectModuleManager">
|
||||
<modules>
|
||||
<module fileurl="file://$PROJECT_DIR$/.idea/CourseInformer.iml" filepath="$PROJECT_DIR$/.idea/CourseInformer.iml" />
|
||||
</modules>
|
||||
</component>
|
||||
</project>
|
||||
6
.idea/vcs.xml
generated
Normal file
6
.idea/vcs.xml
generated
Normal file
@@ -0,0 +1,6 @@
|
||||
<?xml version="1.0" encoding="UTF-8"?>
|
||||
<project version="4">
|
||||
<component name="VcsDirectoryMappings">
|
||||
<mapping directory="$PROJECT_DIR$" vcs="Git" />
|
||||
</component>
|
||||
</project>
|
||||
161
informer.py
Normal file
161
informer.py
Normal file
@@ -0,0 +1,161 @@
|
||||
"""
|
||||
通知类定义:
|
||||
```
|
||||
InfoData:
|
||||
date: 当前日期(包括星期几)
|
||||
time: 当前时间
|
||||
course_time: 课程开始时间
|
||||
course_name: 课程名称
|
||||
course_teacher: 教师名称
|
||||
course_location: 上课地点
|
||||
course_time_interval: 距上课剩余时间,根据上文计算
|
||||
```
|
||||
通知格式参考InfoData内容
|
||||
|
||||
通知形式:参考下方bash脚本,将ip换为192.168.12.1,uid、路径均不变。抽取为方法
|
||||
```
|
||||
#!/bin/bash
|
||||
msg=$1
|
||||
json=$(jq -n --arg msg "$msg" --argjson uid 2998813882 '{user_id: $uid, message: $msg}')
|
||||
curl -s -o /dev/null -X POST http://127.0.0.1:3000/send_private_msg \
|
||||
-H "Content-Type: application/json" \
|
||||
-d "$json"
|
||||
```
|
||||
|
||||
定义方法publish_info负责推送,该类只暴露publish_info方法,如果涉及其他方法以及字段,均需要以`_`开头进行私有化
|
||||
"""
|
||||
|
||||
from dataclasses import dataclass
|
||||
from datetime import datetime
|
||||
import json
|
||||
import requests
|
||||
import os
|
||||
|
||||
|
||||
@dataclass
|
||||
class InfoData:
|
||||
"""Data class for notification information"""
|
||||
date: str # 当前日期(包括星期几)
|
||||
time: str # 当前时间
|
||||
course_time: str # 课程开始时间
|
||||
course_name: str # 课程名称
|
||||
course_teacher: str # 教师名称
|
||||
course_location: str # 上课地点
|
||||
course_time_interval: str # 距上课剩余时间,根据上文计算
|
||||
|
||||
|
||||
def publish_info(info_data: InfoData) -> bool:
|
||||
"""
|
||||
Publish course information notification
|
||||
|
||||
Args:
|
||||
info_data: InfoData object containing notification information
|
||||
|
||||
Returns:
|
||||
bool: True if notification was sent successfully, False otherwise
|
||||
"""
|
||||
message = f"""==========课程提醒==========
|
||||
|
||||
课程: {info_data.course_name}
|
||||
地点: {info_data.course_location}
|
||||
教师: {info_data.course_teacher}
|
||||
课程开始时间: {info_data.course_time}
|
||||
距离上课: {info_data.course_time_interval}
|
||||
|
||||
当前日期: {info_data.date}
|
||||
当前时间: {info_data.time}"""
|
||||
try:
|
||||
# Format the message based on InfoData
|
||||
|
||||
# Send the notification using curl
|
||||
_send_notification(message)
|
||||
return True
|
||||
except Exception as e:
|
||||
print(f"Failed to publish info: {e}")
|
||||
return False
|
||||
|
||||
|
||||
def _load_config():
|
||||
"""
|
||||
Load notification configuration from JSON file (private method)
|
||||
|
||||
Returns:
|
||||
dict: Configuration dictionary
|
||||
"""
|
||||
config_path = os.path.join(os.path.dirname(__file__), 'notification_config.json')
|
||||
with open(config_path, 'r', encoding='utf-8') as f:
|
||||
return json.load(f)
|
||||
|
||||
|
||||
def _send_notification(message: str) -> None:
|
||||
"""
|
||||
Send notification via HTTP POST request (private method)
|
||||
|
||||
Args:
|
||||
message: The message to send
|
||||
"""
|
||||
# Load configuration
|
||||
config = _load_config()
|
||||
|
||||
# Extract server and user configuration
|
||||
host = config['server']['host']
|
||||
port = config['server']['port']
|
||||
endpoint = config['server']['endpoint']
|
||||
uid = config['user']['uid']
|
||||
|
||||
# Create URL
|
||||
url = f"http://{host}:{port}{endpoint}"
|
||||
|
||||
# Create JSON payload
|
||||
payload = {
|
||||
"user_id": uid,
|
||||
"message": message
|
||||
}
|
||||
|
||||
# Send POST request
|
||||
headers = {"Content-Type": "application/json"}
|
||||
response = requests.post(url, json=payload, headers=headers)
|
||||
|
||||
# Raise an exception if the request was unsuccessful
|
||||
response.raise_for_status()
|
||||
|
||||
|
||||
def _get_test_info() -> InfoData:
|
||||
"""
|
||||
Get current course information (private method)
|
||||
|
||||
Returns:
|
||||
InfoData: Current course information
|
||||
"""
|
||||
# Get current date and time
|
||||
now = datetime.now()
|
||||
date_str = now.strftime("%Y-%m-%d")
|
||||
time_str = now.strftime("%H:%M:%S")
|
||||
|
||||
# For demonstration, we'll use sample data
|
||||
# In a real implementation, this would fetch from reader.py
|
||||
info_data = InfoData(
|
||||
date=date_str+" 星期一",
|
||||
time=time_str,
|
||||
course_time="08:00",
|
||||
course_name="示例课程",
|
||||
course_teacher="示例教师",
|
||||
course_location="示例地点",
|
||||
course_time_interval="30分钟"
|
||||
)
|
||||
|
||||
return info_data
|
||||
|
||||
|
||||
# Example usage (for testing purposes)
|
||||
if __name__ == "__main__":
|
||||
# Get current info
|
||||
info = _get_test_info()
|
||||
|
||||
# Publish the info
|
||||
success = publish_info(info)
|
||||
|
||||
if success:
|
||||
print("Notification sent successfully")
|
||||
else:
|
||||
print("Failed to send notification")
|
||||
326
main.py
Normal file
326
main.py
Normal file
@@ -0,0 +1,326 @@
|
||||
"""
|
||||
该类负责整合`reader.py`以及`informer.py`完成读取与通知流程。
|
||||
启动后先从reader读取完整课程信息,然后根据时间安排发送通知。
|
||||
时间安排包括:当前属于第几周、今日的所有课程都会在何时开始。
|
||||
其中课程的开始时间与持续时间需要根据reader中定义的CourseSchedule的timeslot: TimeRange字段来确定
|
||||
比如:from=1,to=2,即对应第一节课到第二节课(包括第二节)
|
||||
而时间详细安排如下:
|
||||
上午:
|
||||
08:00 - 08:45 第一节
|
||||
08:55 - 09:40 第二节
|
||||
09:55 - 10:40 第三节
|
||||
10:50 - 11:35 第四节
|
||||
11:45 - 12:30 第五节
|
||||
|
||||
下午:
|
||||
14:00 - 14:45 第六节
|
||||
14:55 - 15:40 第七节
|
||||
15:55 - 16:40 第八节
|
||||
16:50 - 17:35 第九节
|
||||
17:45 - 18:30 第十节
|
||||
|
||||
晚上:
|
||||
19:00 - 19:45 第十一节
|
||||
19:55 - 20:40 第十二节
|
||||
|
||||
周数安排如下:
|
||||
从指定日期(新学期将从2025-08-25开始,包括08-25当天),每七天为算作一周
|
||||
reader中定义的CourseSchedule类中包括字段week: TimeRange,若该字段的from=1,to=3,即对应第1,2,3周均存在该课程
|
||||
|
||||
通知逻辑按照如下规则:
|
||||
假设对应第x节存在课程的话:
|
||||
如果为第一节存在课程:
|
||||
在07:00发送通知
|
||||
若为第六节:
|
||||
在13:00发送通知
|
||||
若为第十一节:
|
||||
在18:30发送通知
|
||||
其他节课程提前十分钟发送通知即可
|
||||
|
||||
通知的发送时间严格按照上述定义,不得提前发送
|
||||
|
||||
通知将在合适时间发送,可使用时间轮算法完成周、日的课程安排,但需要先对reader返回的课程时间进行解析。该脚本将持续运行并检测安排,在合适的时间发送通知
|
||||
"""
|
||||
|
||||
import time
|
||||
import datetime
|
||||
from typing import List
|
||||
import reader
|
||||
import informer
|
||||
|
||||
|
||||
# 课程时间安排
|
||||
CLASS_SCHEDULE = {
|
||||
"上午": {
|
||||
1: "08:00",
|
||||
2: "08:55",
|
||||
3: "09:55",
|
||||
4: "10:50",
|
||||
5: "11:45"
|
||||
},
|
||||
"下午": {
|
||||
6: "14:00",
|
||||
7: "14:55",
|
||||
8: "15:55",
|
||||
9: "16:50",
|
||||
10: "17:45"
|
||||
},
|
||||
"晚上": {
|
||||
11: "19:00",
|
||||
12: "19:55"
|
||||
}
|
||||
}
|
||||
|
||||
# 通知发送时间
|
||||
NOTIFICATION_TIMES = {
|
||||
1: "07:00", # 第一节课在07:00发送通知
|
||||
6: "13:00", # 第六节课在13:00发送通知
|
||||
11: "18:30" # 第十一节课在18:30发送通知
|
||||
}
|
||||
|
||||
WEEKDAYS = {
|
||||
"Monday": "星期一",
|
||||
"Tuesday": "星期二",
|
||||
"Wednesday": "星期三",
|
||||
"Thursday": "星期四",
|
||||
"Friday": "星期五",
|
||||
"Saturday": "星期六",
|
||||
"Sunday": "星期日"
|
||||
}
|
||||
|
||||
|
||||
def _get_current_week(start_date: datetime.date = datetime.date(2025, 8, 25)) -> int:
|
||||
"""
|
||||
计算当前是第几周
|
||||
|
||||
Args:
|
||||
start_date: 学期开始日期,默认为2025-08-25
|
||||
|
||||
Returns:
|
||||
int: 当前周数
|
||||
"""
|
||||
today = datetime.date.today()
|
||||
delta = today - start_date
|
||||
week = delta.days // 7 + 1
|
||||
return max(week, 1) # 确保至少为第1周
|
||||
|
||||
|
||||
def _get_notification_time(period: str, timeslot_from: int) -> str:
|
||||
"""
|
||||
获取课程通知发送时间
|
||||
|
||||
Args:
|
||||
period: 课程时段(上午/下午/晚上)
|
||||
timeslot_from: 课程开始节次
|
||||
|
||||
Returns:
|
||||
str: 通知发送时间(HH:MM格式)
|
||||
"""
|
||||
# 特殊节次的通知时间
|
||||
if timeslot_from in NOTIFICATION_TIMES:
|
||||
return NOTIFICATION_TIMES[timeslot_from]
|
||||
|
||||
# 其他节次提前十分钟发送通知
|
||||
class_time = CLASS_SCHEDULE.get(period, {}).get(timeslot_from, "08:00")
|
||||
hour, minute = map(int, class_time.split(":"))
|
||||
|
||||
# 提前10分钟
|
||||
total_minutes = hour * 60 + minute - 10
|
||||
if total_minutes < 0:
|
||||
total_minutes += 24 * 60 # 处理跨天情况
|
||||
|
||||
notify_hour = total_minutes // 60
|
||||
notify_minute = total_minutes % 60
|
||||
|
||||
return f"{notify_hour:02d}:{notify_minute:02d}"
|
||||
|
||||
|
||||
def _calculate_time_interval(class_time: str) -> str:
|
||||
"""
|
||||
计算距离上课的时间间隔
|
||||
|
||||
Args:
|
||||
class_time: 课程开始时间(HH:MM格式)
|
||||
|
||||
Returns:
|
||||
str: 时间间隔描述
|
||||
"""
|
||||
now = datetime.datetime.now()
|
||||
class_hour, class_minute = map(int, class_time.split(":"))
|
||||
class_datetime = now.replace(hour=class_hour, minute=class_minute, second=0, microsecond=0)
|
||||
|
||||
if class_datetime < now:
|
||||
class_datetime += datetime.timedelta(days=1) # 如果是今天已过的时间,计算明天的
|
||||
|
||||
delta = class_datetime - now
|
||||
hours = delta.seconds // 3600
|
||||
minutes = (delta.seconds % 3600) // 60
|
||||
|
||||
if hours > 0:
|
||||
return f"{hours}小时{minutes}分钟"
|
||||
else:
|
||||
return f"{minutes}分钟"
|
||||
|
||||
|
||||
def _should_send_notification(course, current_week: int, current_day: str) -> bool:
|
||||
"""
|
||||
判断是否应该发送课程通知
|
||||
|
||||
Args:
|
||||
course: CourseSchedule对象
|
||||
current_week: 当前周数
|
||||
current_day: 当前星期几
|
||||
|
||||
Returns:
|
||||
bool: 是否应该发送通知
|
||||
"""
|
||||
# 检查当前周数是否在课程周数范围内
|
||||
if not (course.week.from_value <= current_week <= course.week.to_value):
|
||||
return False
|
||||
|
||||
# 检查当前星期是否匹配
|
||||
if course.day != current_day:
|
||||
return False
|
||||
|
||||
return True
|
||||
|
||||
|
||||
def _send_course_notification(course, period: str) -> None:
|
||||
"""
|
||||
发送课程通知
|
||||
|
||||
Args:
|
||||
course: CourseSchedule对象
|
||||
period: 课程时段
|
||||
"""
|
||||
# 获取课程开始时间
|
||||
class_time = CLASS_SCHEDULE.get(period, {}).get(course.timeslot.from_value, "08:00")
|
||||
|
||||
# 计算距离上课的时间
|
||||
time_interval = _calculate_time_interval(class_time)
|
||||
|
||||
# 创建通知数据
|
||||
info_data = informer.InfoData(
|
||||
date=f"{datetime.date.today()} {course.day}",
|
||||
time=datetime.datetime.now().strftime("%H:%M:%S"),
|
||||
course_time=class_time,
|
||||
course_name=course.course_name,
|
||||
course_teacher=course.instructor,
|
||||
course_location=course.location,
|
||||
course_time_interval=time_interval
|
||||
)
|
||||
|
||||
# 发送通知
|
||||
informer.publish_info(info_data)
|
||||
|
||||
|
||||
def main():
|
||||
"""主函数:读取课程信息并根据时间安排发送通知"""
|
||||
print("课程通知系统启动...")
|
||||
|
||||
try:
|
||||
# 读取课程信息
|
||||
student_info, courses = reader.read_course_schedule()
|
||||
print(f"已读取 {len(courses)} 门课程信息")
|
||||
|
||||
# 持续运行并检测课程安排
|
||||
while True:
|
||||
# 获取当前周数和星期几
|
||||
current_week = _get_current_week()
|
||||
current_day = datetime.date.today().strftime("%A")
|
||||
# 转换为中文星期几
|
||||
current_day = WEEKDAYS.get(current_day, current_day)
|
||||
|
||||
# 获取当前时间
|
||||
now = datetime.datetime.now()
|
||||
current_time = now.strftime("%H:%M")
|
||||
|
||||
# 检查今天的课程
|
||||
for course in courses:
|
||||
# 判断是否应该发送通知
|
||||
if _should_send_notification(course, current_week, current_day):
|
||||
# 获取通知发送时间
|
||||
notify_time = _get_notification_time(course.period, course.timeslot.from_value)
|
||||
|
||||
# 如果当前时间匹配通知时间,则发送通知
|
||||
if current_time == notify_time:
|
||||
print(f"发送课程通知: {course.course_name}")
|
||||
_send_course_notification(course, course.period)
|
||||
time.sleep(60*40)
|
||||
|
||||
# 每30s检查一次
|
||||
time.sleep(30)
|
||||
|
||||
except KeyboardInterrupt:
|
||||
print("\n课程通知系统已停止")
|
||||
except Exception as e:
|
||||
print(f"课程通知系统出错: {e}")
|
||||
|
||||
|
||||
def _test_day_notifications(courses, current_week: int, current_day: str, day_name: str):
|
||||
"""测试指定日期的通知逻辑(私有方法)"""
|
||||
print(f"\n=== 测试{day_name} ===")
|
||||
# 测试不同时段的通知,从06:55开始到20:00结束,每分钟递增
|
||||
test_times = []
|
||||
start_hour, start_minute = 6, 55
|
||||
end_hour, end_minute = 20, 0
|
||||
|
||||
current_hour, current_minute = start_hour, start_minute
|
||||
while current_hour < end_hour or (current_hour == end_hour and current_minute <= end_minute):
|
||||
test_times.append(f"{current_hour:02d}:{current_minute:02d}")
|
||||
# 递增分钟
|
||||
current_minute += 1
|
||||
if current_minute >= 60:
|
||||
current_minute = 0
|
||||
current_hour += 1
|
||||
|
||||
for test_time in test_times:
|
||||
# 检查课程
|
||||
for course in courses:
|
||||
# 判断是否应该发送通知
|
||||
if _should_send_notification(course, current_week, current_day):
|
||||
# 获取通知发送时间
|
||||
notify_time = _get_notification_time(course.period, course.timeslot.from_value)
|
||||
|
||||
# 如果当前时间匹配通知时间,则发送通知
|
||||
if test_time == notify_time:
|
||||
print(f"\n达到模拟时间: {test_time}")
|
||||
print(f" [通知] {course.course_name} - {course.instructor} - {course.location}")
|
||||
# 不实际发送通知,只模拟
|
||||
# _send_course_notification(course, course.period)
|
||||
|
||||
|
||||
def test_notification_logic():
|
||||
"""测试方法:模拟第一周第一天、第二天、第三天的课程安排通知逻辑"""
|
||||
print("开始测试通知逻辑...")
|
||||
|
||||
try:
|
||||
# 读取课程信息
|
||||
student_info, courses = reader.read_course_schedule()
|
||||
print(f"已读取 {len(courses)} 门课程信息")
|
||||
|
||||
# 模拟第一周
|
||||
current_week = 1
|
||||
print(f"模拟第 {current_week} 周")
|
||||
|
||||
# 测试第一天(星期一)
|
||||
_test_day_notifications(courses, current_week, "星期一", "第一天(星期一)")
|
||||
|
||||
# 测试第二天(星期二)
|
||||
_test_day_notifications(courses, current_week, "星期二", "第二天(星期二)")
|
||||
|
||||
# 测试第三天(星期三)
|
||||
_test_day_notifications(courses, current_week, "星期三", "第三天(星期三)")
|
||||
|
||||
print("\n测试完成")
|
||||
|
||||
except Exception as e:
|
||||
print(f"测试过程中出错: {e}")
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
# 运行测试方法
|
||||
test_notification_logic()
|
||||
|
||||
# 如果要运行主方法,取消下面的注释
|
||||
# main()
|
||||
10
notification_config.json
Normal file
10
notification_config.json
Normal file
@@ -0,0 +1,10 @@
|
||||
{
|
||||
"server": {
|
||||
"host": "192.168.12.1",
|
||||
"port": 3000,
|
||||
"endpoint": "/send_private_msg"
|
||||
},
|
||||
"user": {
|
||||
"uid": 2998813882
|
||||
}
|
||||
}
|
||||
247
reader.py
Normal file
247
reader.py
Normal file
@@ -0,0 +1,247 @@
|
||||
"""
|
||||
Module for reading course schedule information from HTML file
|
||||
"""
|
||||
|
||||
from dataclasses import dataclass
|
||||
from typing import List
|
||||
import os
|
||||
import re
|
||||
from bs4 import BeautifulSoup
|
||||
|
||||
|
||||
@dataclass
|
||||
class TimeRange:
|
||||
"""Data class representing a time range with from and to values"""
|
||||
from_value: int
|
||||
to_value: int
|
||||
|
||||
|
||||
@dataclass
|
||||
class CourseSchedule:
|
||||
"""Data class representing a course schedule entry"""
|
||||
course_name: str
|
||||
instructor: str
|
||||
week: TimeRange # e.g., "1-8", "11-18", etc.
|
||||
timeslot: TimeRange # e.g., "[1-2]", "[6-7]", "[8-10]", etc.
|
||||
location: str # e.g., "[龙]二号楼2301"
|
||||
day: str # e.g., "星期一", "星期二", etc.
|
||||
period: str # e.g., "上午", "下午", "晚上"
|
||||
|
||||
|
||||
@dataclass
|
||||
class StudentInfo:
|
||||
"""Data class representing student information"""
|
||||
student_id: str
|
||||
student_name: str
|
||||
class_name: str
|
||||
total_credits: float
|
||||
|
||||
|
||||
def read_course_schedule(file_path: str = "resources/教学安排表.xls") -> tuple[StudentInfo, List[CourseSchedule]]:
|
||||
"""
|
||||
Read _course schedule information from HTML file
|
||||
|
||||
Args:
|
||||
file_path: Path to the HTML file containing _course schedule
|
||||
|
||||
Returns:
|
||||
Tuple of (StudentInfo, List of CourseSchedule objects)
|
||||
"""
|
||||
if not os.path.exists(file_path):
|
||||
raise FileNotFoundError(f"Course schedule file not found: {file_path}")
|
||||
|
||||
# Read the HTML file with GBK encoding
|
||||
with open(file_path, 'r', encoding='GBK') as file:
|
||||
content = file.read()
|
||||
|
||||
# Parse HTML content
|
||||
soup = BeautifulSoup(content, 'html.parser')
|
||||
|
||||
# Extract student information
|
||||
student_id = "Unknown"
|
||||
student_name = "Unknown"
|
||||
class_name = "Unknown"
|
||||
total_credits = 0.0
|
||||
|
||||
# Get student info from hidden inputs
|
||||
xh_input = soup.find('input', {'id': 'xh'})
|
||||
if xh_input and xh_input.get('value'):
|
||||
student_id = xh_input['value']
|
||||
|
||||
# Get student name and class from the first table
|
||||
tables = soup.find_all('table')
|
||||
if tables:
|
||||
first_table = tables[0]
|
||||
table_rows = first_table.find_all('tr')
|
||||
for row in table_rows:
|
||||
cells = row.find_all('td')
|
||||
for cell in cells:
|
||||
text = cell.get_text().strip()
|
||||
if text.startswith("姓名:"):
|
||||
student_name = text.replace("姓名:", "")
|
||||
elif text.startswith("所在班级:"):
|
||||
class_name = text.replace("所在班级:", "")
|
||||
|
||||
_student_info = StudentInfo(
|
||||
student_id=student_id,
|
||||
student_name=student_name,
|
||||
class_name=class_name,
|
||||
total_credits=total_credits
|
||||
)
|
||||
|
||||
# Extract _course schedule information
|
||||
_courses = []
|
||||
|
||||
# Get day names from the header row
|
||||
days = ["星期一", "星期二", "星期三", "星期四", "星期五"] # Default days
|
||||
header_row = soup.find('tr', class_='H')
|
||||
if header_row:
|
||||
day_cells = header_row.find_all('td', class_='td0')
|
||||
days = []
|
||||
for cell in day_cells:
|
||||
day_text = cell.get_text().strip()
|
||||
if day_text and "星期" in day_text: # Only add actual day names
|
||||
days.append(day_text)
|
||||
|
||||
# Get _course information from the schedule table
|
||||
course_divs = soup.find_all('div', class_='div1')
|
||||
for div in course_divs:
|
||||
# Extract _course information from each div
|
||||
xkinfo = div.find('span', class_='xkinfo')
|
||||
if xkinfo:
|
||||
# Each div may contain multiple _courses
|
||||
course_blocks = xkinfo.find_all('div', style=lambda x: x and 'padding-bottom:5px' in x)
|
||||
for block in course_blocks:
|
||||
# Get the raw text
|
||||
raw_text = block.get_text()
|
||||
|
||||
# Initialize variables
|
||||
course_name = "Unknown Course"
|
||||
instructor = "Unknown"
|
||||
location = "Unknown"
|
||||
week_range = TimeRange(1, 1)
|
||||
timeslot_range = TimeRange(1, 1)
|
||||
|
||||
# Parse the concatenated text based on the format:
|
||||
# course_name + instructor + weeks[timeslot] + location
|
||||
# e.g., "企业资源计划(ERP)黄伟 1-8[3-4][龙]一号楼1307"
|
||||
|
||||
# Look for the pattern: numbers followed by brackets (timeslot)
|
||||
time_pattern = r'([0-9\-]+)(\[[0-9\-]+\])'
|
||||
time_match = re.search(time_pattern, raw_text)
|
||||
|
||||
if time_match:
|
||||
# Extract time information
|
||||
weeks_str = time_match.group(1)
|
||||
timeslot_str = time_match.group(2)
|
||||
|
||||
# Parse week range
|
||||
week_range = TimeRange(1, 1)
|
||||
week_range = time_parser(weeks_str,week_range)
|
||||
|
||||
# Parse timeslot range (remove brackets)
|
||||
timeslot_range = TimeRange(1, 1)
|
||||
timeslot_clean = timeslot_str.strip('[]')
|
||||
timeslot_range = time_parser(timeslot_clean, timeslot_range)
|
||||
|
||||
# Extract location (everything after the time info)
|
||||
time_end = time_match.end()
|
||||
if time_end < len(raw_text):
|
||||
location = raw_text[time_end:].strip()
|
||||
|
||||
# Extract the part before time info
|
||||
time_start = time_match.start()
|
||||
before_time = raw_text[:time_start].strip()
|
||||
|
||||
# Split the part before time to get _course name and instructor
|
||||
# Look for Chinese characters to identify the instructor
|
||||
instructor_pattern = r'([\u4e00-\u9fff]+)$'
|
||||
instructor_match = re.search(instructor_pattern, before_time)
|
||||
if instructor_match:
|
||||
instructor = instructor_match.group(1)
|
||||
# Course name is everything before the instructor
|
||||
instructor_start = instructor_match.start()
|
||||
course_name = before_time[:instructor_start].strip()
|
||||
else:
|
||||
# If we can't find instructor, use the whole part as _course name
|
||||
course_name = before_time
|
||||
else:
|
||||
# Fallback: try to extract at least the _course name
|
||||
# Assume the first part is the _course name
|
||||
parts = raw_text.split()
|
||||
if parts:
|
||||
course_name = parts[0]
|
||||
|
||||
# Determine day and period based on div id
|
||||
day = "Unknown"
|
||||
period = "Unknown"
|
||||
div_id = div.get('id', '')
|
||||
if div_id.startswith('k') and len(div_id) >= 3:
|
||||
# Extract day from div id (k11, k21, etc.)
|
||||
# First digit after 'k' represents the day (1=Monday, 2=Tuesday, etc.)
|
||||
try:
|
||||
day_index = int(div_id[1]) - 1
|
||||
if 0 <= day_index < len(days):
|
||||
day = days[day_index]
|
||||
except (ValueError, IndexError):
|
||||
pass
|
||||
|
||||
# Extract period from div id (last digit represents the period)
|
||||
# 1,2 = 上午, 3,4 = 下午, 5 = 晚上
|
||||
try:
|
||||
period_index = int(div_id[2])
|
||||
if period_index in [1, 2]:
|
||||
period = "上午"
|
||||
elif period_index in [3, 4]:
|
||||
period = "下午"
|
||||
elif period_index == 5:
|
||||
period = "晚上"
|
||||
except (ValueError, IndexError):
|
||||
pass
|
||||
|
||||
_course = CourseSchedule(
|
||||
course_name=course_name,
|
||||
instructor=instructor,
|
||||
week=week_range,
|
||||
timeslot=timeslot_range,
|
||||
location=location,
|
||||
day=day,
|
||||
period=period
|
||||
)
|
||||
_courses.append(_course)
|
||||
|
||||
return _student_info, _courses
|
||||
|
||||
|
||||
def time_parser(timeslot_clean, timeslot_range):
|
||||
if '-' in timeslot_clean:
|
||||
timeslot_parts = timeslot_clean.split('-')
|
||||
if len(timeslot_parts) == 2:
|
||||
try:
|
||||
timeslot_range = TimeRange(int(timeslot_parts[0]), int(timeslot_parts[1]))
|
||||
except ValueError:
|
||||
pass
|
||||
else:
|
||||
try:
|
||||
timeslot_value = int(timeslot_clean)
|
||||
timeslot_range = TimeRange(timeslot_value, timeslot_value)
|
||||
except ValueError:
|
||||
pass
|
||||
return timeslot_range
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
# Test the function
|
||||
try:
|
||||
student_info, courses = read_course_schedule()
|
||||
print(f"Student: {student_info.student_name} ({student_info.student_id})")
|
||||
print(f"Class: {student_info.class_name}")
|
||||
print("\nCourses:")
|
||||
for course in courses: # Print first 5 courses
|
||||
print(f"- {course.course_name} by {course.instructor}")
|
||||
print(f" Time: {course.week.from_value}-{course.week.to_value} [{course.timeslot.from_value}-{course.timeslot.to_value}]")
|
||||
print(f" Location: {course.location}")
|
||||
print(f" Day: {course.day}, Period: {course.period}")
|
||||
print()
|
||||
except Exception as e:
|
||||
print(f"Error reading course schedule: {e}")
|
||||
Reference in New Issue
Block a user