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.AddEntityPipeline(new MySqlEntityPipeline
{
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), new TargetUrlExtractor
{
Region = new BaseSelector {
Type = SelectorType.XPath, Expression = "//span[@class=\"p-num\"]"
},
Patterns = new List <string> {
@"&page=[0-9]+&"
}
});
return(context);
}
protected override EntitySpider GetEntitySpider()
{
EntitySpider context = new EntitySpider(new Site
{
//HttpProxyPool = new HttpProxyPool(new KuaidailiProxySupplier("快代理API"))
})
{
UserId = "DotnetSpider",
TaskGroup = "JdSkuSampleSpider"
};
context.SetThreadNum(1);
context.SetIdentity("JD_sku_store_test_" + DateTime.Now.ToString("yyyy_MM_dd_hhmmss"));
context.AddEntityPipeline(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), new TargetUrlExtractor
{
Region = new BaseSelector {
Type = SelectorType.XPath, Expression = "//span[@class=\"p-num\"]"
},
Patterns = new List <string> {
@"&page=[0-9]+&"
}
});
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.AddEntityPipeline(
new MySqlEntityPipeline("Database='test';Data Source=localhost;User ID=root;Password=1qazZAQ!;Port=3306"));
context.AddStartUrl("http://www.cas.cn/kx/kpwz/index.shtml");
context.AddStartUrl("http://www.cas.cn/kx/kpwz/index_1.shtml");
context.AddEntityType(typeof(ArticleSummary), new TargetUrlExtractor
{
Patterns = new List <string> {
@"index_[0-9]+.shtml", "index.shtml"
}
});
context.AddEntityType(typeof(Article), new TargetUrlExtractor
{
Patterns = new List <string> {
@"t[0-9]+_[0-9]+.shtml"
}
});
return(context);
}
protected override EntitySpider GetEntitySpider()
{
EntitySpider context = new EntitySpider(new Site
{
//HttpProxyPool = new HttpProxyPool(new KuaidailiProxySupplier("http://dev.kuaidaili.com/api/getproxy/?orderid=917184806038194&num=999&b_pcchrome=1&b_pcie=1&b_pcff=1&protocol=1&method=2&an_tr=1&an_an=1&an_ha=1&sep=1"))
})
{
UserId = "86Research",
TaskGroup = "JdSkuSampleSpider"
};
context.SetThreadNum(1);
context.SetIdentity("JD_sku_store_test_" + DateTime.Now.ToString("yyyy_MM_dd"));
context.AddEntityPipeline(
new MySqlEntityPipeline("Database='test';Data Source=redis;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), new TargetUrlExtractor
{
Region = new BaseSelector {
Type = SelectorType.XPath, Expression = "//span[@class=\"p-num\"]"
},
Patterns = new List <string> {
@"&page=[0-9]+&"
}
});
return(context);
}
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);
}
public void MultiEntitiesInitPipelines()
{
EntitySpider context = new EntitySpider(new Site());
context.SetThreadNum(1);
context.SetIdentity("test-MultiEntitiesInitPipelines");
context.AddEntityPipeline(new MySqlEntityPipeline("Database='test';Data Source=localhost;User ID=root;Password=1qazZAQ!;Port=3306"));
context.AddEntityPipeline(new MySqlFileEntityPipeline());
context.AddEntityPipeline(new ConsoleEntityPipeline());
context.AddEntityPipeline(new JsonFileEntityPipeline());
#if !NET_CORE
//context.AddEntityPipeline(new MongoDbEntityPipeline("mongo"));
#endif
context.AddStartUrl("http://a.com");
context.AddEntityType(typeof(Entity13));
context.AddEntityType(typeof(Entity12));
context.Run("running-test");
var entityPipelines = context.EntityPipelines;
#if NET_CORE
Assert.Equal(4, entityPipelines.Count);
#else
Assert.Equal(4, entityPipelines.Count);
//Assert.Equal(5, entityPipelines.Count);
#endif
var pipeline1 = (MySqlEntityPipeline)entityPipelines[0];
Assert.Equal("Database='test';Data Source=localhost;User ID=root;Password=1qazZAQ!;Port=3306", pipeline1.ConnectString);
Assert.Equal("MySqlFileEntityPipeline", entityPipelines[1].GetType().Name);
Assert.Equal("ConsoleEntityPipeline", entityPipelines[2].GetType().Name);
Assert.Equal("JsonFileEntityPipeline", entityPipelines[3].GetType().Name);
#if !NET_CORE
//Assert.Equal("MongoDbEntityPipeline", entityPipelines[4].GetType().Name);
//var pipeline2 = (MySqlEntityPipeline)entityPipelines[4];
//Assert.Equal("mongo", pipeline2.ConnectString);
#endif
var pipelines = context.GetPipelines();
Assert.Equal(1, pipelines.Count);
EntityPipeline pipeline = (EntityPipeline)pipelines[0];
entityPipelines = pipeline.GetEntityPipelines();
Assert.Equal(4, entityPipelines.Count);
pipeline1 = (MySqlEntityPipeline)entityPipelines[0];
Assert.Equal("db", pipeline1.GetSchema().Database);
Assert.Equal("table", pipeline1.GetSchema().TableName);
#if !NET_CORE
//var pipeline2 = (MongoDbEntityPipeline)entityPipelines[4];
//Assert.Equal("db", pipeline2.GetSchema().Database);
//Assert.Equal("table", pipeline2.GetSchema().TableName);
#endif
using (MySqlConnection conn = new MySqlConnection("Database='mysql';Data Source=localhost;User ID=root;Password=1qazZAQ!;Port=3306"))
{
conn.Execute($"DROP table db.table");
}
}
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);
}
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"));
context.AddEntityPipeline(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);
}
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");
}
}
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.SetIdentity("JD sku/store test " + DateTime.Now.ToString("yyyy-MM-dd HHmmss"));
context.AddEntityPipeline(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), new TargetUrlExtractor
{
Region = new Selector {
Type = SelectorType.XPath, Expression = "//span[@class=\"p-num\"]"
},
Patterns = new List <string> {
@"&page=[0-9]+&"
}
});
context.SetDownloader(new WebDriverDownloader(Browser.Chrome));
return(context);
}
protected override EntitySpider GetEntitySpider()
{
//Connecting string
const string connstr = "Data Source=localhost;Initial Catalog=test;User ID=sa;Password=1234";
EntitySpider context = new EntitySpider(new Site {
})
{
UserId = "DotnetSpider",
TaskGroup = "RuthSpider"
};
context.SetThreadNum(1);
context.SetIdentity("RuthSpider " + DateTime.Now.ToString("yyyy_MM_dd_hhmmss"));
context.AddEntityPipeline(new MySqlEntityPipeline("Database='mysql';Data Source=localhost;User ID=root;Password=1qazZAQ!;Port=3306"));
context.AddStartUrl("http://www.tsia.org.tw/member_list.php?page=1");
//添加公司列表頁面Entity
context.AddEntityType(typeof(CompanySummary), new TargetUrlExtractor
{
Patterns = new List <string> {
@"member_list.php\?page=\d+"
}
});
//添加公司詳情頁面Entity
context.AddEntityType(typeof(Company), new TargetUrlExtractor
{
Patterns = new List <string> {
@"member_info.php\?ID=\d+"
}
});
//Config Redis
context.SetScheduler(new RedisScheduler
{
Host = "localhost",
Password = "",
Port = 6379
});
return(context);
}
protected override EntitySpider GetEntitySpider()
{
EntitySpider context = new EntitySpider(new Site());
context.SetIdentity("ddeng.com " + DateTime.Now.ToString("yyyy-MM-dd HHmmss"));
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 MySqlEntityPipeline("Database='test';Data Source=localhost;User ID=root;Password=1qazZAQ!;Port=3306"));
context.AddStartUrl("http://www.ddeng.com/product/982227");
context.AddEntityType(typeof(Corp));
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.SetIdentity("qidian_" + DateTime.Now.ToString("yyyy_MM_dd_HHmmss"));
context.AddPipeline(new CollectEntityPipeline());
context.AddStartUrl("http://www.cas.cn/kx/kpwz/index.shtml");
context.AddEntityType(typeof(ArticleSummary));
return(context);
}