From 5019e46d088765ea8df54861874c30a9dbc79c90 Mon Sep 17 00:00:00 2001 From: eson <474420502@qq.com> Date: Sun, 9 Dec 2018 21:39:44 +0800 Subject: [PATCH] add proxy --- person.go | 12 ++++++++---- task.go | 28 ++++++++++++++-------------- task_test.go | 2 +- test.curl | 2 +- test.yaml | 2 +- toutiao.curl | 1 + 6 files changed, 26 insertions(+), 21 deletions(-) create mode 100644 toutiao.curl diff --git a/person.go b/person.go index aef14b0..8984fb9 100644 --- a/person.go +++ b/person.go @@ -97,9 +97,10 @@ func splitTasks(conf *Config) []ITask { initTask(conf, task, curl) for _, cnode := range proxies.GetLoopValues() { proxy := cnode.GetValue().(string) + log.Println(proxy) task.AddProxies(proxy) } - + log.Println(task.GetProxies()) tasks = append(tasks, task) case 1: @@ -118,7 +119,7 @@ func splitTasks(conf *Config) []ITask { initTask(conf, ptask, ncurl) ptask.AddProxies(proxy) - tasks = append(tasks, task) + tasks = append(tasks, ptask) } } } @@ -166,16 +167,19 @@ func ExecuteOnPlan(task ITask, result chan string) { urlname := task.GetCurl().Name for { + interval := time.Second * 2 if task.TimeUp() { rtask = task.Execute(data) // 事件 在这里变化 + ntime := task.NextTime() + interval = ntime.Sub(time.Now()) if rtask == nil { // log.Println("rtask is nil") result <- fmt.Sprintf("rtask is nil, the first task = %s, name = %s", taskname, urlname) return } + task = rtask + } - interval := task.NextTime().Sub(time.Now()) time.Sleep(interval) - task = rtask } } diff --git a/task.go b/task.go index f89f315..011b1cf 100644 --- a/task.go +++ b/task.go @@ -44,10 +44,7 @@ func (task *Task) GetCurl() *curl2info.CURL { // AddProxies 添加代理集合 func (task *Task) AddProxies(proxy string) { - for _, cnode := range task.proxies.GetLoopValues() { - proxy := cnode.GetValue().(string) - task.proxies.Append(proxy) - } + task.proxies.Append(proxy) } // GetProxies 获取代理的字符串 @@ -78,23 +75,25 @@ func (task *Task) NextTime() time.Time { // Workflow 根据persistent 是否返回现有session(or 新建 session),创建Workflow的信息, 便于设置或者更改参数, Session持久化 func (task *Task) Workflow(persistent bool) *requests.Workflow { - task.workflow = task.curl.CreateWorkflow(task.Session()) if persistent { + task.workflow = task.curl.CreateWorkflow(task.Session()) return task.workflow } - return task.curl.CreateWorkflow(nil) + proxies := task.GetProxies() + ses := task.curl.CreateSession() + if proxies.Size() > 0 { + proxy := proxies.Cursor().GetValue().(string) + ses.SetConfig(requests.CProxy, proxy) + proxies.MoveNext() + } + + return task.curl.CreateWorkflow(ses) } // Request 根据curl信息执行,没持久化 func (task *Task) Request() (*requests.Response, error) { - proxies := task.GetProxies() - ses := task.curl.CreateSession() - if proxies.Size() > 0 { - ses.SetConfig(requests.CProxy, proxies.Cursor().GetValue().(string)) - proxies.MoveNext() - } - return task.curl.CreateWorkflow(ses).Execute() + return task.Workflow(false).Execute() } // Session 获取Session的信息 只保留session的数据和url参数. @@ -103,7 +102,8 @@ func (task *Task) Session() *requests.Session { task.session = task.curl.CreateSession() } if task.proxies.Size() > 0 { - task.session.SetConfig(requests.CProxy, task.proxies.Cursor().GetValue().(string)) + proxy := task.proxies.Cursor().GetValue().(string) + task.session.SetConfig(requests.CProxy, proxy) task.proxies.MoveNext() } diff --git a/task_test.go b/task_test.go index 6a5ce06..8c7edf2 100644 --- a/task_test.go +++ b/task_test.go @@ -29,7 +29,7 @@ type Toutiao struct { func (tt *Toutiao) Execute(data interface{}) ITask { resp, err := tt.Request() - log.Println(resp, err) + log.Println(resp.Content()[0:20], err) return nil } diff --git a/test.curl b/test.curl index aeabce7..52b6ac1 100644 --- a/test.curl +++ b/test.curl @@ -1,2 +1,2 @@ -curl --name "tt-information" 'http://is.snssdk.com/2/article/information/v24/?latitude=22.831367&longitude=113.511515&group_id=6565653745026204168&item_id=6565653745026204168&aggr_type=1&context=1&from_category=news_game&article_page=0&iid=34903754482&device_id=41148471494&ac=wifi&channel=oppo-cpa&aid=13&app_name=news_article&version_code=676&version_name=6.7.6&device_platform=android&ab_version=304489%2C261579%2C373245%2C360501%2C374617%2C366851%2C356335%2C345191%2C271178%2C357704%2C326524%2C326532%2C292723%2C366036%2C323233%2C371779%2C346557%2C351090%2C319958%2C372620%2C362184%2C214069%2C31643%2C333971%2C366873%2C374962%2C372618%2C280449%2C281298%2C366489%2C325619%2C373770%2C357402%2C361073%2C362402%2C290191%2C370014%2C353484%2C375739%2C373725%2C295827%2C353305%2C375426%2C374426%2C239095%2C360541%2C344347%2C170988%2C371590%2C368831%2C368827%2C368775%2C374117%2C365053%2C374232%2C368303%2C375692%2C330632%2C297059%2C374250%2C276206%2C286212%2C350193%2C365036%2C373741%2C374405%2C373368%2C370846%2C364453%2C375713%2C369501%2C369165%2C368839%2C375433%2C373123%2C371555%2C371963%2C374142%2C372907&ab_client=a1%2Cc4%2Ce1%2Cf1%2Cg2%2Cf7&ab_group=94567%2C102754%2C181430&ab_feature=94567%2C102754&abflag=3&ssmix=a&device_type=ONEPLUS+A3010&device_brand=OnePlus&language=zh&os_api=26&os_version=8.0.0&uuid=864854034514328&openudid=9b35a4035eecee2c&manifest_version_code=676&resolution=1080*1920&dpi=420&update_version_code=67610&_rticket=1528706910264&plugin=10603&pos=5r_-9Onkv6e_eCQieCoDeCUfv7G_8fLz-vTp6Pn4v6esrK6zqKysqKyosb_x_On06ej5-L-nr6-zpa6srquqsb_88Pzt3vTp5L-nv3gkIngqA3glH7-xv_zw_O3R8vP69Ono-fi_p6ysrrOupauqqaSxv_zw_O3R_On06ej5-L-nr66zrairpKqv4A%3D%3D&fp=HrT_FlD_PMcIFlD5FSU1FYmeFrxO&rom_version=26&ts=1528706911&as=a265e371dff53b57de5999&mas=0073e8ef3f9a8b842da0ead7d35c0597ea2ee0ccce5e5d5db5' --task toutiao -H 'Accept-Encoding: gzip' -H 'X-SS-REQ-TICKET: 1528706910267' -H 'User-Agent: Dalvik/2.1.0 (Linux; U; Android 8.0.0; ONEPLUS A3010 Build/OPR1.170623.032) NewsArticle/6.7.6 okhttp/3.10.0.1' -H 'Cookie: odin_tt=210899a257b5fe787a3465e2220fb94d91d5ad34c77dee3560f93fccc82dd738cccb301770f633530fdd6ceea955983d; UM_distinctid=163ace3b0050-08fccf530af621-f1c0e26-49a10-163ace3b0093e8; CNZZDATA1271720685=1435124261-1527612007-%7C1527612007; CNZZDATA1264530760=119491224-1527609979-%7C1527612115; JSESSIONID=67814B7DDE08D5A9F3B3D684220CF3FB; alert_coverage=6; qh[360]=1; install_id=34903754482; ttreq=1$b7221ef01bd5ed7c030f5db45e959686c9ddd0d2' -H 'Host: is.snssdk.com' -H 'Connection: Keep-Alive' +curl --name "cip.cc" 'http://cip.cc' curl --name "ag-cn" 'https://appgrowing.cn/' -H 'authority: appgrowing.cn' -H 'cache-control: max-age=0' -H 'upgrade-insecure-requests: 1' -H 'user-agent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/70.0.3538.110 Safari/537.36' -H 'accept: text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,image/apng,*/*;q=0.8' -H 'accept-encoding: gzip, deflate, br' -H 'accept-language: zh' -H 'cookie: _ga=GA1.2.1371058419.1533104518; _gid=GA1.2.1726457685.1544092110; _gat_gtag_UA_4002880_19=1' -H 'if-none-match: W/"5c088f5a-ca6"' -H 'if-modified-since: Thu, 06 Dec 2018 02:54:18 GMT' --compressed \ No newline at end of file diff --git a/test.yaml b/test.yaml index b7e0637..3f50a7d 100644 --- a/test.yaml +++ b/test.yaml @@ -18,4 +18,4 @@ media : 55 spider_id : 73 catch_account_id : 123 -crontab: "2s" \ No newline at end of file +crontab: "1s" \ No newline at end of file diff --git a/toutiao.curl b/toutiao.curl new file mode 100644 index 0000000..79f54ce --- /dev/null +++ b/toutiao.curl @@ -0,0 +1 @@ +curl --name "tt-information" 'http://is.snssdk.com/2/article/information/v24/?latitude=22.831367&longitude=113.511515&group_id=6565653745026204168&item_id=6565653745026204168&aggr_type=1&context=1&from_category=news_game&article_page=0&iid=34903754482&device_id=41148471494&ac=wifi&channel=oppo-cpa&aid=13&app_name=news_article&version_code=676&version_name=6.7.6&device_platform=android&ab_version=304489%2C261579%2C373245%2C360501%2C374617%2C366851%2C356335%2C345191%2C271178%2C357704%2C326524%2C326532%2C292723%2C366036%2C323233%2C371779%2C346557%2C351090%2C319958%2C372620%2C362184%2C214069%2C31643%2C333971%2C366873%2C374962%2C372618%2C280449%2C281298%2C366489%2C325619%2C373770%2C357402%2C361073%2C362402%2C290191%2C370014%2C353484%2C375739%2C373725%2C295827%2C353305%2C375426%2C374426%2C239095%2C360541%2C344347%2C170988%2C371590%2C368831%2C368827%2C368775%2C374117%2C365053%2C374232%2C368303%2C375692%2C330632%2C297059%2C374250%2C276206%2C286212%2C350193%2C365036%2C373741%2C374405%2C373368%2C370846%2C364453%2C375713%2C369501%2C369165%2C368839%2C375433%2C373123%2C371555%2C371963%2C374142%2C372907&ab_client=a1%2Cc4%2Ce1%2Cf1%2Cg2%2Cf7&ab_group=94567%2C102754%2C181430&ab_feature=94567%2C102754&abflag=3&ssmix=a&device_type=ONEPLUS+A3010&device_brand=OnePlus&language=zh&os_api=26&os_version=8.0.0&uuid=864854034514328&openudid=9b35a4035eecee2c&manifest_version_code=676&resolution=1080*1920&dpi=420&update_version_code=67610&_rticket=1528706910264&plugin=10603&pos=5r_-9Onkv6e_eCQieCoDeCUfv7G_8fLz-vTp6Pn4v6esrK6zqKysqKyosb_x_On06ej5-L-nr6-zpa6srquqsb_88Pzt3vTp5L-nv3gkIngqA3glH7-xv_zw_O3R8vP69Ono-fi_p6ysrrOupauqqaSxv_zw_O3R_On06ej5-L-nr66zrairpKqv4A%3D%3D&fp=HrT_FlD_PMcIFlD5FSU1FYmeFrxO&rom_version=26&ts=1528706911&as=a265e371dff53b57de5999&mas=0073e8ef3f9a8b842da0ead7d35c0597ea2ee0ccce5e5d5db5' --task toutiao -H 'Accept-Encoding: gzip' -H 'X-SS-REQ-TICKET: 1528706910267' -H 'User-Agent: Dalvik/2.1.0 (Linux; U; Android 8.0.0; ONEPLUS A3010 Build/OPR1.170623.032) NewsArticle/6.7.6 okhttp/3.10.0.1' -H 'Cookie: odin_tt=210899a257b5fe787a3465e2220fb94d91d5ad34c77dee3560f93fccc82dd738cccb301770f633530fdd6ceea955983d; UM_distinctid=163ace3b0050-08fccf530af621-f1c0e26-49a10-163ace3b0093e8; CNZZDATA1271720685=1435124261-1527612007-%7C1527612007; CNZZDATA1264530760=119491224-1527609979-%7C1527612115; JSESSIONID=67814B7DDE08D5A9F3B3D684220CF3FB; alert_coverage=6; qh[360]=1; install_id=34903754482; ttreq=1$b7221ef01bd5ed7c030f5db45e959686c9ddd0d2' -H 'Host: is.snssdk.com' -H 'Connection: Keep-Alive' \ No newline at end of file