B_GroupingJob.cs 6.6 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154
  1. using System;
  2. using System.Collections.Generic;
  3. using System.Configuration;
  4. using System.Linq;
  5. using System.Text;
  6. using System.Threading.Tasks;
  7. using FCS.Crawler.ZCLotteryMatchs;
  8. using FCS.Interface;
  9. using FCS.Common;
  10. using FCS.Models.Entity;
  11. using HtmlAgilityPack;
  12. using System.Threading;
  13. namespace FCS.Crawler.Basketball
  14. {
  15. public class B_GroupingJob : CommonJob
  16. {
  17. private string eventId = "", url = "";
  18. public void Click()
  19. {
  20. GetNBA();
  21. }
  22. public void GetNBA()
  23. {
  24. var result = new List<B_Grouping>();
  25. var url = "http://lanqiu.zgzcw.com/1/{0}/saiAll.do".FormatMe(
  26. DateTime.Now.AddYears(-1).Year.ToString().Substring(2) + "-" + DateTime.Now.Year.ToString().Substring(2));
  27. eventId = new B_EventsJob().GetEventList().Where(p => p.Name == ConfigurationManager.AppSettings["NBAEventName"]).ToList()[0].Id;
  28. var doc = CommonHelper.GetHtml(url, "足彩");//加载主页面
  29. var season = GetNBASeason(doc.DocumentNode).ToList();
  30. foreach (var model in season)
  31. {
  32. taskList.Add(Task.Run(() =>
  33. {
  34. if (model.Remark != url)
  35. doc = CommonHelper.GetHtml(model.Remark, "足彩");//加载主页面
  36. var parentGroup = doc.DocumentNode.SelectSingleNode(".//div[@attr='saiAll']").SelectNodes("./span");
  37. foreach (var item in parentGroup)
  38. {
  39. var childId = item.Attributes["Id"].Value.Replace("Title", "") + "s";
  40. var childGroup_Div = doc.DocumentNode.SelectSingleNode(".//div[@id='{0}']".FormatMe(childId)).SelectSingleNode(".//div[@class='xq']");
  41. var id = CommonHelper.GetGuid().ToString();
  42. result.Add(GetModel(new B_Grouping
  43. {
  44. Id = id,
  45. Name = item.SelectNodes("./a")[0].InnerText.Trim(),
  46. Sort = parentGroup.IndexOf(item) + 1,
  47. Remark = model.Remark,
  48. Season = model.Season,
  49. Extended1 = childGroup_Div == null ? "-1" : string.Empty,
  50. }));
  51. if (childGroup_Div != null)
  52. {
  53. var childGroup = childGroup_Div.SelectNodes(".//a");
  54. if (childGroup != null)
  55. {
  56. foreach (var _item in childGroup)
  57. {
  58. result.Add(GetModel(new B_Grouping
  59. {
  60. Name = _item.InnerText.Trim(),
  61. Sort = childGroup.IndexOf(_item) + 1,
  62. ParentId = id,
  63. Remark = model.Remark,
  64. Season = model.Season,
  65. Extended1 = _item.Attributes["attr"].Value
  66. }));
  67. }
  68. }
  69. }
  70. }
  71. }));
  72. }
  73. Task.WaitAll(taskList.ToArray());
  74. var data = new List<B_Grouping>();
  75. var groupList = services.Query<B_Grouping>(" AND EventId='{0}'".FormatMe(eventId));
  76. //新增的父分组、子分组
  77. var parentList = (from a in result
  78. where !(from b in groupList where string.IsNullOrEmpty(b.ParentId) && a.EventId == b.EventId select b.Name.Trim()).Contains(a.Name.Trim()) && string.IsNullOrEmpty(a.ParentId)
  79. select a).ToList();
  80. parentList.ForEach(p =>
  81. {
  82. data.Add(p);
  83. var childList = result.Where(q => q.ParentId == p.Id).ToList();
  84. childList.ForEach(q =>
  85. {
  86. data.Add(q);
  87. });
  88. });
  89. //新增的子分组
  90. groupList.Where(p => string.IsNullOrEmpty(p.ParentId)).ToList().ForEach(q =>
  91. {
  92. var oldList = groupList.Where(s => s.ParentId.Trim() == q.Id && s.EventId == q.EventId).ToList();//数据库子节点的数据
  93. var newParentId = result.Where(s => s.Name.Trim() == q.Name.Trim() && s.EventId == q.EventId && string.IsNullOrEmpty(s.ParentId)).ToList();//爬取数据子节点数据
  94. var newList = result.Where(s => s.ParentId == (newParentId.Count > 0 ? newParentId[0].Id : "0")).ToList();
  95. if (newList.Count > 0 && oldList.Count != newList.Count)
  96. {
  97. var list = (from a in newList
  98. where !(from b in oldList select b.Name.Trim()).Contains(a.Name.Trim())
  99. select a).ToList();
  100. list.ForEach(d =>
  101. {
  102. d.ParentId = q.Id;
  103. data.Add(d);
  104. });
  105. }
  106. });
  107. services.SqlBulkCopyAdd(data);
  108. }
  109. private B_Grouping GetModel(B_Grouping model)
  110. {
  111. if (model.Id.IsEmpty())
  112. model.Id = CommonHelper.GetGuid().ToString();
  113. model.CreateDateTime = DateTime.Now;
  114. model.EventId = eventId;
  115. return model;
  116. }
  117. /// <summary>
  118. /// 得到NBA的期数
  119. /// </summary>
  120. /// <param name="doc"></param>
  121. /// <returns></returns>
  122. private IEnumerable<B_Grouping> GetNBASeason(HtmlNode doc)
  123. {
  124. // var list = new List< B_Grouping >()
  125. var oldSeason = (from a in services.Query<B_Grouping>(" AND EventId='{0}'".FormatMe(eventId))
  126. group a by a.Season into g
  127. select g.Key).ToList();
  128. var season = doc.SelectSingleNode(".//select[@id='selectSeason']").SelectNodes(".//option");
  129. foreach (var item in season)
  130. {
  131. if (!oldSeason.Contains(item.Attributes["value"].Value) || item.Attributes["value"].Value.Contains(DateTime.Now.Year.ToString()))
  132. yield return new B_Grouping
  133. {
  134. Remark = "http://lanqiu.zgzcw.com/1/" + item.Attributes["value"].Value + "/saiAll.do",
  135. Season = item.Attributes["value"].Value
  136. };
  137. }
  138. }
  139. }
  140. }