protected override EntitySpider GetEntitySpider()
{
EntitySpider context = new EntitySpider(new Site())
{
UserId = "DotnetSpider",
TaskGroup = "HaoBrowser",
Identity = "HaoBrowser Hao360Spider Buble " + DateTime.Now.ToString("yyyy-MM-dd HHmmss"),
CachedSize = 1,
ThreadNum = 1,
SkipWhenResultIsEmpty = true,
Downloader = new HttpClientDownloader
{
DownloadCompleteHandlers = new IDownloadCompleteHandler[]
{
new SubContentHandler {
Start = "sales[\"hotsite_yixing\"] = [",
End = "}}",
StartOffset = 27,
EndOffset = 0
},
new ReplaceContentHandler {
NewValue = "/",
OldValue = "\\/",
},
}
}
};
context.SetScheduler(new Extension.Scheduler.RedisScheduler("127.0.0.1:6379,serviceName=Scheduler.NET,keepAlive=8,allowAdmin=True,connectTimeout=10000,password=6GS9F2QTkP36GggE0c3XwVwI,abortConnect=True,connectRetry=20"));
context.AddPipeline(new MySqlEntityPipeline("Database='testhao';Data Source= localhost;User ID=root;Password=root@123456;Port=3306"));
context.AddStartUrl("https://hao.360.cn/");
context.AddEntityType(typeof(UpdateHao360Info));
return(context);
}
protected override EntitySpider GetEntitySpider()
{
EntitySpider context = new EntitySpider(new Site());
context.SetSite(new Site());
context.SetThreadNum(2);
context.ThreadNum = 1;
context.RetryWhenResultIsEmpty = false;
context.Deep = 100;
context.EmptySleepTime = 5000;
context.SetEmptySleepTime(5000);
context.ExitWhenComplete = true;
context.CachedSize = 1;
context.SetDownloader(new HttpClientDownloader());
context.SetScheduler(new QueueDuplicateRemovedScheduler());
context.SkipWhenResultIsEmpty = true;
context.SpawnUrl = true;
context.AddPipeline(new CollectEntityPipeline());
context.AddStartUrl("http://www.cas.cn/kx/kpwz/index.shtml");
context.AddEntityType(typeof(ArticleSummary));
Name = "qidian";
Batch = DateTime.Now.ToString("yyyy_MM_dd_HHmmss");
return(context);
}
public void TestCorrectRedisSetting()
{
EntitySpider spider = new EntitySpider(new Site());
spider.AddPipeline(new ConsoleEntityPipeline());
spider.AddEntityType(typeof(TestEntity));
spider.Run("running-test");
}
public void MultiEntitiesInitPipelines()
{
EntitySpider context = new EntitySpider(new Site());
context.SetIdentity(Guid.NewGuid().ToString("N"));
context.SetThreadNum(1);
context.AddPipeline(new MySqlEntityPipeline("Database='test';Data Source=localhost;User ID=root;Password=1qazZAQ!;Port=3306"));
context.AddPipeline(new MySqlFileEntityPipeline());
context.AddPipeline(new ConsoleEntityPipeline());
context.AddPipeline(new JsonFileEntityPipeline());
context.AddStartUrl("http://baidu.com");
context.AddEntityType(typeof(Entity13));
context.AddEntityType(typeof(Entity12));
context.Run("running-test");
var entityPipelines = context.Pipelines;
Assert.AreEqual(4, entityPipelines.Count);
var pipeline1 = (MySqlEntityPipeline)entityPipelines[0];
Assert.AreEqual("Database='test';Data Source=localhost;User ID=root;Password=1qazZAQ!;Port=3306", pipeline1.ConnectString);
Assert.AreEqual("MySqlFileEntityPipeline", entityPipelines[1].GetType().Name);
Assert.AreEqual("ConsoleEntityPipeline", entityPipelines[2].GetType().Name);
Assert.AreEqual("JsonFileEntityPipeline", entityPipelines[3].GetType().Name);
var pipelines = context.GetPipelines();
Assert.AreEqual(4, pipelines.Count);
IEntityPipeline pipeline = (IEntityPipeline)pipelines[0];
//entityPipelines = pipeline.GetEntityPipelines();
//Assert.AreEqual(4, entityPipelines.Count);
//pipeline1 = (MySqlEntityPipeline)entityPipelines[0];
//Assert.AreEqual("db", pipeline1.GetSchema().Database);
//Assert.AreEqual("table13", pipeline1.GetSchema().Name);
using (MySqlConnection conn = new MySqlConnection("Database='mysql';Data Source=localhost;User ID=root;Password=1qazZAQ!;Port=3306"))
{
conn.Execute($"DROP table db.table12");
conn.Execute($"DROP table db.table13");
}
}
protected override EntitySpider GetEntitySpider()
{
EntitySpider context = new EntitySpider(new Site());
context.SetIdentity("cnblogs homepage " + DateTime.Now.ToString("yyyy-MM-dd HHmmss"));
context.AddStartUrl("http://www.cnblogs.com");
context.AddPipeline(new ConsoleEntityPipeline());
context.AddEntityType(typeof(HomePage));
return(context);
}
protected override EntitySpider GetEntitySpider()
{
var entitySpider = new EntitySpider(new Site());
entitySpider.AddStartUrl("http://www.jd.com/allSort.aspx");
entitySpider.AddEntityType(typeof(Category));
entitySpider.AddPipeline(new MySqlEntityPipeline("Database='mysql';Data Source=localhost;User ID=root;Password=1qazZAQ!;Port=3306"));
var t = JsonConvert.SerializeObject(entitySpider);
return(entitySpider);
}
protected override EntitySpider GetEntitySpider()
{
var entitySpider = new EntitySpider(new Site())
{
Identity = "JdCategory Daliy Tracking " + DateTimeUtils.Day1OfThisWeek.ToString("yyyy-MM-dd")
};
entitySpider.AddStartUrl("http://www.jd.com/allSort.aspx");
entitySpider.AddEntityType(typeof(Category));
entitySpider.AddPipeline(new MySqlEntityPipeline("Database='mysql';Data Source=localhost;User ID=root;Password=1qazZAQ!;Port=3306"));
return(entitySpider);
}
protected override EntitySpider GetEntitySpider()
{
EntitySpider context = new EntitySpider(new Site());
context.SetIdentity("JD sku/store test " + DateTime.Now.ToString("yyyy-MM-dd HHmmss"));
context.AddPipeline(new MySqlEntityPipeline("Database='mysql';Data Source=localhost;User ID=root;Password=1qazZAQ!;Port=3306"));
context.AddStartUrl("http://list.jd.com/list.html?cat=9987,653,655&page=2&JL=6_0_0&ms=5#J_main", new Dictionary <string, object> {
{ "name", "手机" }, { "cat3", "655" }
});
context.AddEntityType(typeof(Product));
context.SetDownloader(new WebDriverDownloader(Browser.Chrome));
return(context);
}
protected override EntitySpider GetEntitySpider()
{
EntitySpider context = new EntitySpider(new Site());
context.SetIdentity("ShanxizhaoshengSpider " + DateTime.Now.ToString("yyyy-MM-dd HHmmss"));
context.SetSite(new Site
{
EncodingName = "GB2312"
});
context.AddPipeline(new SqlServerEntityPipeline("Data Source=.\\SQLEXPRESS;Initial Catalog=master;Integrated Security=True"));
context.AddStartUrl("http://www.sneac.com/pgjhcx/ypbkyxjg.jsp?a11709CountNo=2000");
context.AddEntityType(typeof(Item));
return(context);
}
public void ClearScheduler()
{
EntitySpider spider = new EntitySpider(new Site());
spider.Identity = Guid.NewGuid().ToString("N");
spider.SetScheduler(new RedisScheduler("127.0.0.1:6379,serviceName=Scheduler.NET,keepAlive=8,allowAdmin=True,connectTimeout=10000,password=6GS9F2QTkP36GggE0c3XwVwI,abortConnect=True,connectRetry=20"));
spider.AddStartUrl("https://baidu.com");
spider.AddPipeline(new ConsoleEntityPipeline());
spider.AddEntityType(typeof(TestEntity));
spider.Run();
var confiruation = new ConfigurationOptions()
{
ServiceName = "DotnetSpider",
ConnectTimeout = 65530,
KeepAlive = 8,
ConnectRetry = 3,
ResponseTimeout = 3000,
Password = "******",
AllowAdmin = true
};
confiruation.EndPoints.Add(new DnsEndPoint("127.0.0.1", 6379));
var redis = ConnectionMultiplexer.Connect(confiruation);
var db = redis.GetDatabase(0);
var md5 = Encrypt.Md5Encrypt(spider.Identity);
var itemKey = "item-" + md5;
var setKey = "set-" + md5;
var queueKey = "queue-" + md5;
var errorCountKey = "error-record" + md5;
var successCountKey = "success-record" + md5;
//queue
Assert.AreEqual(0, db.ListLength(queueKey));
//set
Assert.AreEqual(0, db.SetLength(setKey));
//item
Assert.AreEqual(0, db.HashLength(itemKey));
//error-count
Assert.AreEqual(false, db.StringGet(errorCountKey).HasValue);
//success-count
Assert.AreEqual(false, db.StringGet(successCountKey).HasValue);
}
protected override EntitySpider GetEntitySpider()
{
EntitySpider context = new EntitySpider(new Site
{
//HttpProxyPool = new HttpProxyPool(new KuaidailiProxySupplier("快代理API"))
});
context.SetThreadNum(1);
context.SetIdentity("JD_sku_store_test_" + DateTime.Now.ToString("yyyy_MM_dd_hhmmss"));
// dowload html by http client
context.SetDownloader(new HttpClientDownloader());
// save data to mysql.
context.AddPipeline(new MySqlEntityPipeline("Database='test';Data Source=localhost;User ID=root;Password=1qazZAQ!;Port=3306"));
context.AddStartUrl("http://list.jd.com/list.html?cat=9987,653,655&page=2&JL=6_0_0&ms=5#J_main", new Dictionary <string, object> {
{ "name", "手机" }, { "cat3", "655" }
});
context.AddEntityType(typeof(Product));
return(context);
}
protected override EntitySpider GetEntitySpider()
{
Name = "JD Shop details " + DateTimeUtils.RunIdOfMonday;
var context = new EntitySpider(new Site())
{
TaskGroup = "JD SKU Weekly",
CachedSize = 1,
ThreadNum = 8,
Scheduler = new RedisScheduler("127.0.0.1:6379,serviceName=Scheduler.NET,keepAlive=8,allowAdmin=True,connectTimeout=10000,password=6GS9F2QTkP36GggE0c3XwVwI,abortConnect=True,connectRetry=20"),
Downloader = new HttpClientDownloader
{
DownloadCompleteHandlers = new IDownloadCompleteHandler[]
{
new SubContentHandler
{
Start = "json(",
End = ");",
StartOffset = 5,
EndOffset = 0
}
}
},
PrepareStartUrls = new PrepareStartUrls[]
{
new BaseDbPrepareStartUrls()
{
Source = DataSource.MySql,
ConnectString = "Database='test';Data Source= localhost;User ID=root;Password=1qazZAQ!;Port=3306",
QueryString = $"SELECT * FROM jd.sku_v2_{DateTimeUtils.RunIdOfMonday} WHERE shopname is null or shopid is null order by sku",
Columns = new [] { new DataColumn {
Name = "sku"
} },
FormateStrings = new List <string> {
"http://chat1.jd.com/api/checkChat?my=list&pidList={0}&callback=json"
}
}
}
};
context.AddPipeline(new MySqlEntityPipeline("Database='taobao';Data Source=localhost ;User ID=root;Password=1qazZAQ!;Port=4306"));
context.AddEntityType(typeof(ProductUpdater));
return(context);
}
protected override EntitySpider GetEntitySpider()
{
EntitySpider context = new EntitySpider(new Site
{
EncodingName = "UTF-8"
});
context.AddPipeline(
new MySqlEntityPipeline("Database='test';Data Source=localhost;User ID=root;Password=1qazZAQ!;Port=3306"));
var word = "可乐|雪碧";
context.AddStartUrl(string.Format("http://news.baidu.com/ns?word={0}&tn=news&from=news&cl=2&pn=0&rn=20&ct=1", word), new Dictionary <string, dynamic> {
{ "Keyword", word }
});
context.AddEntityType(typeof(BaiduSearchEntry));
return(context);
}
public void SqlServerDataTypeTests()
{
using (var conn = new SqlConnection("Server=.\\SQLEXPRESS;Database=test;Trusted_Connection=True;MultipleActiveResultSets=true"))
{
EntitySpider context = new EntitySpider(new Site());
context.SetIdentity(Guid.NewGuid().ToString("N"));
context.SetThreadNum(1);
context.AddPipeline(new SqlServerEntityPipeline("Server=.\\SQLEXPRESS;Database=test;Trusted_Connection=True;MultipleActiveResultSets=true"));
context.AddStartUrl("http://baidu.com");
context.AddEntityType(typeof(Entity15));
context.Run("running-test");
var columns = conn.Query <ColumnInfo>("USE [test];select b.name Name,c.name+'(' + cast(c.length as varchar)+')' [Type] from sysobjects a,syscolumns b,systypes c where a.id=b.id and a.name='table15' and a.xtype='U'and b.xtype=c.xtype").ToList();;
Assert.AreEqual(11, columns.Count);
Assert.AreEqual("Int", columns[0].Name);
Assert.AreEqual("Time", columns[1].Name);
Assert.AreEqual("CDate", columns[2].Name);
Assert.AreEqual("Float", columns[3].Name);
Assert.AreEqual("Double", columns[4].Name);
Assert.AreEqual("BigInt", columns[5].Name);
Assert.AreEqual("__Id", columns[6].Name);
Assert.AreEqual("String", columns[7].Name);
Assert.AreEqual("String1", columns[8].Name);
Assert.AreEqual("int(4)", columns[0].Type);
Assert.AreEqual("datetime(8)", columns[1].Type);
Assert.AreEqual("datetime(8)", columns[2].Type);
Assert.AreEqual("float(8)", columns[3].Type);
Assert.AreEqual("float(8)", columns[4].Type);
Assert.AreEqual("bigint(8)", columns[5].Type);
Assert.AreEqual("bigint(8)", columns[6].Type);
Assert.AreEqual("nvarchar(8000)", columns[7].Type);
Assert.AreEqual("nvarchar(8000)", columns[8].Type);
conn.Execute("USE [test]; drop table [test].dbo.[table15]");
}
}
public void MySqlDataTypeTests()
{
using (MySqlConnection conn = new MySqlConnection("Database='mysql';Data Source=localhost;User ID=root;Password=1qazZAQ!;Port=3306"))
{
EntitySpider context = new EntitySpider(new Site());
context.SetIdentity(Guid.NewGuid().ToString("N"));
context.SetThreadNum(1);
context.AddPipeline(new MySqlEntityPipeline("Database='test';Data Source=localhost;User ID=root;Password=1qazZAQ!;Port=3306"));
context.AddStartUrl("http://baidu.com");
context.AddEntityType(typeof(Entity15));
context.Run("running-test");
var columns = conn.Query <ColumnInfo>("SELECT COLUMN_NAME as `Name`, COLUMN_TYPE as `Type` FROM information_schema.columns WHERE table_name='table15' AND table_schema = 'test';").ToList();;
Assert.AreEqual(9, columns.Count);
Assert.AreEqual("Int", columns[0].Name);
Assert.AreEqual("BigInt", columns[1].Name);
Assert.AreEqual("String", columns[2].Name);
Assert.AreEqual("Time", columns[3].Name);
Assert.AreEqual("Float", columns[4].Name);
Assert.AreEqual("Double", columns[5].Name);
Assert.AreEqual("String1", columns[6].Name);
Assert.AreEqual("cdate", columns[7].Name);
Assert.AreEqual("__id", columns[8].Name);
Assert.AreEqual("int(11)", columns[0].Type);
Assert.AreEqual("bigint(20)", columns[1].Type);
Assert.AreEqual("text", columns[2].Type);
Assert.AreEqual("timestamp", columns[3].Type);
Assert.AreEqual("float", columns[4].Type);
Assert.AreEqual("double", columns[5].Type);
Assert.AreEqual("varchar(100)", columns[6].Type);
Assert.AreEqual("timestamp", columns[7].Type);
Assert.AreEqual("bigint(20)", columns[8].Type);
conn.Execute("drop table `test`.`table15`");
}
}
protected override EntitySpider GetEntitySpider()
{
EntitySpider context = new EntitySpider(new Site());
context.SetThreadNum(1);
context.SetIdentity("JD_sku_store_test_" + DateTime.Now.ToString("yyyy_MM_dd_HHmmss"));
context.AddPipeline(new MySqlEntityPipeline(null)
{
UpdateConnectString = new DbUpdateConnectString
{
ConnectString = "Database='test';Data Source=localhost;User ID=root;Password=1qazZAQ!;Port=3306",
QueryString = "SELECT value from `dotnetspider`.`settings` where `type`='ConnectString' and `key`='MySql01' LIMIT 1"
}
});
context.AddStartUrl("http://list.jd.com/list.html?cat=9987,653,655&page=2&JL=6_0_0&ms=5#J_main",
new Dictionary <string, object> {
{ "name", "手机" }, { "cat3", "655" }
});
context.AddEntityType(typeof(JdSkuSampleSpider.Product));
return(context);
}
protected override EntitySpider GetEntitySpider()
{
EntitySpider context = new EntitySpider(new Site());
context.SetSite(new Site
{
CookiesStringPart = "sid=dea284fc36c24e8cbcd447343d7b8a4e; sn=DD962248; ctid=000000; ctnm=%E5%8F%A4%E9%95%87%E7%81%AF%E9%A5%B0%E6%89%B9%E5%8F%91; ctpv=%E5%B9%BF%E4%B8%9C; JSESSIONID=acbBqFfOD4I63d9PziDvv; DDENG=c4fc08ae2e3ba3efeddbc667c2f45e615a85e80009169501dc244a03e87908aa61146548b97ed9c7dc07af23bfd80bff5008f8c8867a9165d4bd2732aca0db7dedae2e042d3968fcad1150f36be242e8a32a3f59db2a0b39216a59f1628508c5799644532a9d99925f9841b3c13a1f97; userId=10003379; previousUser=%E5%A4%95%E7%8E%89; Hm_lvt_9e33f153f28be198970d205d90a24f28=1466146335; Hm_lpvt_9e33f153f28be198970d205d90a24f28=1466146392; Hm_lvt_54b4cb498afd05463ab4611b38a6f289=1466146335; Hm_lpvt_54b4cb498afd05463ab4611b38a6f289=1466146392; CNZZDATA1256982382=395301521-1466143554-%7C1466143554",
Headers = new Dictionary <string, string>
{
{ "Cache-Control", "max-age=0" },
{ "Upgrade-Insecure-Requests", "1" }
},
UserAgent = "Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/50.0.2661.102 Safari/537.36",
Accept = "text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8"
});
context.AddPipeline(new SqlServerEntityPipeline("Server=.\\SQLEXPRESS;Database=test;Trusted_Connection=True;MultipleActiveResultSets=true"));
context.AddStartUrl("http://www.ddeng.com/product/982227");
context.AddEntityType(typeof(Corp));
return(context);
}
protected override EntitySpider GetEntitySpider()
{
Site site = new Site();
using (var reader = new StreamReader(File.OpenRead("taobaokeyword.txt")))
{
string keyword;
while (!string.IsNullOrEmpty(keyword = reader.ReadLine()))
{
site.AddStartUrl("https://" + $"s.taobao.com/search?q={keyword}&imgfile=&js=1&stats_click=search_radio_all%3A1&ie=utf8&sort=sale-desc&s=0&tab={1}&fs=1&filter_tianmao=tmall", new Dictionary <string, object>
{
{ "keyword", keyword }
});
}
}
var context = new EntitySpider(site)
{
ThreadNum = 5,
SkipWhenResultIsEmpty = true,
Scheduler = new RedisScheduler("127.0.0.1:6379,serviceName = DotnetSpider,keepAlive = 8,allowAdmin = True,connectTimeout = 10000,password = 6GS9F2QTkP36GggE0c3XwVwI,abortConnect = True,connectRetry = 20"),
Downloader = new HttpClientDownloader
{
DownloadCompleteHandlers = new IDownloadCompleteHandler[]
{
new SubContentHandler
{
StartOffset = 16,
EndOffset = 22,
Start = "g_page_config = {",
End = "g_srp_loadCss();"
},
new IncrementTargetUrlsCreator("&s=0", null, 44)
}
}
};
context.AddPipeline(new MySqlEntityPipeline("Database = 'mysql'; Data Source = localhost; User ID = root; Password = 1qazZAQ!; Port = 3306"));
context.AddEntityType(typeof(Item), new MyDataHanlder());
return(context);
}
protected override EntitySpider GetEntitySpider()
{
EntitySpider context = new EntitySpider(new Site())
{
Downloader = new HttpClientDownloader
{
DownloadCompleteHandlers = new IDownloadCompleteHandler[]
{
new IncrementTargetUrlsCreator("index_1.shtml")
}
},
};
context.SetThreadNum(10);
context.SetIdentity("qidian_" + DateTime.Now.ToString("yyyy_MM_dd_HHmmss"));
context.AddPipeline(
new MySqlEntityPipeline("Database='test';Data Source=localhost;User ID=root;Password=1qazZAQ!;Port=3306"));
context.AddStartUrl("http://oa.jlu.edu.cn/. ");
context.AddEntityType(typeof(ArticleSummary));
context.AddEntityType(typeof(Article));
return(context);
}
protected override EntitySpider GetEntitySpider()
{
var site = new Site()
{
UserAgent = "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/58.0.3029.110 Safari/537.36",
Accept = "text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8",
Headers = new Dictionary <string, string>
{
{ "Accept-Encoding", "gzip, deflate, sdch" },
{ "Upgrade-Insecure-Requests", "1" },
{ "Accept-Language", "en,en-US;q=0.8" },
{ "Cache-Control", "ax-age=0" },
}
};
site.AddStartUrl("http://chat1.jd.com/api/checkChat?my=list&pidList=3355984&callback=json");
site.AddStartUrl("http://chat1.jd.com/api/checkChat?my=list&pidList=3682523&callback=json");
var context = new EntitySpider(site)
{
Downloader = new HttpClientDownloader
{
DownloadCompleteHandlers = new IDownloadCompleteHandler[]
{
new SubContentHandler
{
Start = "json(",
End = ");",
StartOffset = 5,
EndOffset = 2
}
}
}
};
context.AddPipeline(new MySqlEntityPipeline("Database='mysql';Data Source=localhost ;User ID=root;Password=1qazZAQ!;Port=3306"));
context.AddEntityType(typeof(ProductUpdater));
return(context);
}