125 lines
5.2 KiB
Python
125 lines
5.2 KiB
Python
from django.db import models
|
|
from django.utils import timezone
|
|
import json
|
|
|
|
|
|
class Website(models.Model):
|
|
name = models.CharField(max_length=100, unique=True)
|
|
base_url = models.URLField()
|
|
description = models.TextField(blank=True, null=True)
|
|
article_list_url = models.URLField(blank=True, null=True)
|
|
article_selector = models.CharField(max_length=255, blank=True, null=True)
|
|
content_selector = models.CharField(max_length=255, blank=True, null=True)
|
|
enabled = models.BooleanField(default=True)
|
|
|
|
def __str__(self):
|
|
return self.name
|
|
|
|
|
|
class Article(models.Model):
|
|
website = models.ForeignKey(Website, on_delete=models.CASCADE)
|
|
title = models.CharField(max_length=300)
|
|
url = models.URLField(unique=True)
|
|
pub_date = models.DateTimeField(null=True, blank=True)
|
|
content = models.TextField() # html内容
|
|
created_at = models.DateTimeField(auto_now_add=True)
|
|
media_files = models.JSONField(default=list, blank=True)
|
|
|
|
def __str__(self):
|
|
return self.title
|
|
|
|
|
|
class CrawlTask(models.Model):
|
|
"""爬取任务模型"""
|
|
TASK_STATUS_CHOICES = [
|
|
('pending', '等待中'),
|
|
('running', '运行中'),
|
|
('completed', '已完成'),
|
|
('failed', '失败'),
|
|
('cancelled', '已取消'),
|
|
]
|
|
|
|
TASK_TYPE_CHOICES = [
|
|
('keyword', '关键词搜索'),
|
|
('historical', '历史文章'),
|
|
('full_site', '全站爬取'),
|
|
]
|
|
|
|
name = models.CharField(max_length=200, verbose_name="任务名称")
|
|
task_type = models.CharField(max_length=20, choices=TASK_TYPE_CHOICES, default='keyword', verbose_name="任务类型")
|
|
keyword = models.CharField(max_length=200, blank=True, null=True, verbose_name="搜索关键词")
|
|
websites = models.ManyToManyField(Website, blank=True, verbose_name="目标网站")
|
|
start_date = models.DateField(blank=True, null=True, verbose_name="开始日期")
|
|
end_date = models.DateField(blank=True, null=True, verbose_name="结束日期")
|
|
max_pages = models.IntegerField(default=10, verbose_name="最大页数")
|
|
max_articles = models.IntegerField(default=100, verbose_name="最大文章数")
|
|
|
|
status = models.CharField(max_length=20, choices=TASK_STATUS_CHOICES, default='pending', verbose_name="状态")
|
|
progress = models.IntegerField(default=0, verbose_name="进度百分比")
|
|
current_website = models.CharField(max_length=100, blank=True, null=True, verbose_name="当前网站")
|
|
current_action = models.CharField(max_length=200, blank=True, null=True, verbose_name="当前操作")
|
|
|
|
total_articles = models.IntegerField(default=0, verbose_name="总文章数")
|
|
success_count = models.IntegerField(default=0, verbose_name="成功数")
|
|
failed_count = models.IntegerField(default=0, verbose_name="失败数")
|
|
|
|
created_at = models.DateTimeField(auto_now_add=True, verbose_name="创建时间")
|
|
started_at = models.DateTimeField(blank=True, null=True, verbose_name="开始时间")
|
|
completed_at = models.DateTimeField(blank=True, null=True, verbose_name="完成时间")
|
|
|
|
error_message = models.TextField(blank=True, null=True, verbose_name="错误信息")
|
|
result_details = models.JSONField(default=dict, blank=True, verbose_name="结果详情")
|
|
|
|
created_by = models.CharField(max_length=100, blank=True, null=True, verbose_name="创建者")
|
|
|
|
class Meta:
|
|
verbose_name = "爬取任务"
|
|
verbose_name_plural = "爬取任务"
|
|
ordering = ['-created_at']
|
|
|
|
def __str__(self):
|
|
return f"{self.name} ({self.get_status_display()})"
|
|
|
|
def get_websites_display(self):
|
|
"""获取网站列表的显示文本"""
|
|
try:
|
|
websites = self.websites.all()
|
|
if not websites:
|
|
return "所有网站"
|
|
# 确保网站名称是字符串并可以被join处理
|
|
website_names = [str(w.name) for w in websites if w.name]
|
|
return ", ".join(website_names) if website_names else "所有网站"
|
|
except Exception:
|
|
# 如果出现任何异常,返回默认值
|
|
return "所有网站"
|
|
|
|
def get_duration(self):
|
|
"""获取任务执行时长"""
|
|
if not self.started_at:
|
|
return None
|
|
end_time = self.completed_at or timezone.now()
|
|
return end_time - self.started_at
|
|
|
|
def is_running(self):
|
|
"""判断任务是否正在运行"""
|
|
return self.status == 'running'
|
|
|
|
def can_cancel(self):
|
|
"""判断任务是否可以取消"""
|
|
return self.status in ['pending', 'running']
|
|
|
|
def get_progress_display(self):
|
|
"""获取进度显示文本"""
|
|
if self.status == 'pending':
|
|
return "等待开始"
|
|
elif self.status == 'running':
|
|
if self.current_website and self.current_action:
|
|
return f"正在处理 {self.current_website}: {self.current_action}"
|
|
return f"运行中 ({self.progress}%)"
|
|
elif self.status == 'completed':
|
|
return f"已完成 ({self.success_count}/{self.total_articles})"
|
|
elif self.status == 'failed':
|
|
return f"失败: {self.error_message[:50]}..." if self.error_message else "失败"
|
|
elif self.status == 'cancelled':
|
|
return "已取消"
|
|
return "未知状态" |