Collects google ads/answers + captcha solver test

This commit is contained in:
Rustem Kamalov
2024-05-06 01:13:07 +03:00
parent 482f725c10
commit 0a3c07b149
13 changed files with 283 additions and 92 deletions
+7 -14
View File
@@ -30,6 +30,7 @@ docker-compose up --build
| file | File extension to search (e.g. `PDF`, `DOC`) |
| site | Search within a specific website |
| limit | Limit the number of results
| answers | Include google answers as negative rank indexes (e.g. `true`, `false`)
### **Search**
### *Example request*
@@ -47,27 +48,18 @@ You can replace `google` to `yandex` or `baidu` in query to change search engine
"rank": 1,
"url": "https://en.wikipedia.org/wiki/%22Hello,_World!%22_program",
"title": "\"Hello, World!\" program",
"description": "A \"Hello, World!\" program is generally a computer program that ignores any input, and outputs or displays a message similar to \"Hello, World!\"."
"description": "A \"Hello, World!\" program is generally a computer program that ignores any input, and outputs or displays a message similar to \"Hello, World!\".",
"ad": false
},
]
```
### **Images**
### **Images** **[WIP]**
### *Example request*
Get 100 **Google** results for `golden puppy`:
```
GET http://127.0.0.1:7000/google/image?text=golden puppy&limit=100
```
### *Example response*
```JSON
[
{
"rank": 1,
"url": "https://en.wikipedia.org/wiki/%22Hello,_World!%22_program",
"title": "\"Hello, World!\" program",
"description": "A \"Hello, World!\" program is generally a computer program that ignores any input, and outputs or displays a message similar to \"Hello, World!\"."
},
]
```
## CLI <a name="cli"></a> ⌨️
* Use `-h` flag to see commands.
@@ -86,7 +78,8 @@ As a result you should get JSON output containting search results:
"rank": 1,
"url": "https://www.cyberoptik.net/blog/6-sure-fire-ways-to-get-banned-from-google/",
"title": "11 Sure-Fire Ways to Get Banned From Google | CyberOptik",
"description": "How To Get Banned From Google · 1. Cloaking: The Art of Deception · 2. Plagiarism: Because Originality is Overrated · 3. Keyword Stuffing: More is Always Better · 4 ..."
"description": "How To Get Banned From Google · 1. Cloaking: The Art of Deception · 2. Plagiarism: Because Originality is Overrated · 3. Keyword Stuffing: More is Always Better · 4 ...",
"ad": false
},
]
```
+13 -8
View File
@@ -1,7 +1,6 @@
package cmd
import (
"errors"
"fmt"
"strings"
@@ -13,16 +12,21 @@ import (
)
const (
version = "0.2.1"
version = "0.3"
defaultConfigFilename = "config"
envPrefix = "OPENSERP"
)
type Config struct {
App AppConfig `mapstructure:"app"`
GoogleConfig core.SearchEngineOptions `mapstructure:"google"`
YandexConfig core.SearchEngineOptions `mapstructure:"yandex"`
BaiduConfig core.SearchEngineOptions `mapstructure:"baidu"`
App AppConfig `mapstructure:"app"`
Config2Capcha Config2Captcha `mapstructure:"2captcha"`
GoogleConfig core.SearchEngineOptions `mapstructure:"google"`
YandexConfig core.SearchEngineOptions `mapstructure:"yandex"`
BaiduConfig core.SearchEngineOptions `mapstructure:"baidu"`
}
type Config2Captcha struct {
ApiKey string `mapstructure:"apikey"`
}
type AppConfig struct {
@@ -88,7 +92,7 @@ func initializeConfig(cmd *cobra.Command) error {
// 1. Config. Return an error if we cannot parse the config file.
err := v.ReadInConfig()
if err != nil {
err = errors.New(fmt.Sprintf("Cannot read config: %v", err))
err = fmt.Errorf("cannot read config: %v", err)
logrus.Warn(err)
}
@@ -107,7 +111,7 @@ func initializeConfig(cmd *cobra.Command) error {
// Dump Viper values to config struct
err = v.Unmarshal(&config)
if err != nil {
return errors.New(fmt.Sprintf("Cannot unmarshall config: %v", err))
return fmt.Errorf("cannot unmarshall config: %v", err)
}
if config.App.IsDebug {
@@ -128,4 +132,5 @@ func init() {
RootCmd.PersistentFlags().BoolVarP(&config.App.IsLeakless, "leakless", "l", false, "Use leakless mode to insure browser instances are closed after search")
RootCmd.PersistentFlags().BoolVarP(&config.App.IsRawRequests, "raw", "r", false, "Disable browser usage, use HTTP requests")
RootCmd.PersistentFlags().BoolVarP(&config.App.IsLeaveHead, "leave", "", false, "Leave browser and tabs opened after search is made")
RootCmd.PersistentFlags().StringVarP(&config.Config2Capcha.ApiKey, "2captcha_key", "", "", "2 captcha api key")
}
+5 -4
View File
@@ -54,10 +54,11 @@ func searchBrowser(engineType string, query core.Query) ([]core.SearchResult, er
var engine core.SearchEngine
opts := core.BrowserOpts{
IsHeadless: !config.App.IsBrowserHead, // Disable headless if browser head mode is set
IsLeakless: config.App.IsLeakless,
Timeout: time.Second * time.Duration(config.App.Timeout),
LeavePageOpen: config.App.IsLeaveHead,
IsHeadless: !config.App.IsBrowserHead, // Disable headless if browser head mode is set
IsLeakless: config.App.IsLeakless,
Timeout: time.Second * time.Duration(config.App.Timeout),
LeavePageOpen: config.App.IsLeaveHead,
CaptchaSolverApiKey: config.Config2Capcha.ApiKey,
}
if config.App.IsDebug {
+5 -4
View File
@@ -21,10 +21,11 @@ var serveCMD = &cobra.Command{
func serve(cmd *cobra.Command, args []string) {
opts := core.BrowserOpts{
IsHeadless: !config.App.IsBrowserHead, // Disable headless if browser head mode is set
IsLeakless: config.App.IsLeakless,
Timeout: time.Second * time.Duration(config.App.Timeout),
LeavePageOpen: config.App.IsLeaveHead,
IsHeadless: !config.App.IsBrowserHead, // Disable headless if browser head mode is set
IsLeakless: config.App.IsLeakless,
Timeout: time.Second * time.Duration(config.App.Timeout),
LeavePageOpen: config.App.IsLeaveHead,
CaptchaSolverApiKey: config.Config2Capcha.ApiKey,
}
if config.App.IsDebug {
+11 -6
View File
@@ -6,15 +6,20 @@ app:
timeout: 15
head: false
leakless: false
leave_head: false
2captcha:
apikey: "123123123123123"
google:
rate_requests: 4 # Number of requests per Minute
rate_burst: 2 # Number of non-ratelimited requests per Minute
rate_requests: 4 # Number of requests per Minute
rate_burst: 2 # Number of non-ratelimited requests per Minute
captcha: true
yandex:
rate_requests: 4
rate_burst: 2
rate_requests: 4
rate_burst: 2
baidu:
rate_requests: 4
rate_burst: 2
rate_requests: 4
rate_burst: 2
+28
View File
@@ -0,0 +1,28 @@
package core
import (
api2captcha "github.com/2captcha/2captcha-go"
)
type CaptchaSolver struct {
client *api2captcha.Client
}
func NewSolver(apikey string) *CaptchaSolver {
cs := CaptchaSolver{}
cs.client = api2captcha.NewClient(apikey)
return &cs
}
func (cs *CaptchaSolver) SolveReCaptcha2(sitekey, pageUrl, dataS string) (string, error) {
cap := api2captcha.ReCaptcha{
SiteKey: sitekey,
Url: pageUrl,
DataS: dataS,
Invisible: false,
Action: "verify",
}
req := cap.ToRequest()
req.SetProxy("HTTPS", "login:password@IP_address:PORT")
return cs.client.Solve(req)
}
+20
View File
@@ -0,0 +1,20 @@
package core
import (
"testing"
)
var (
API_KEY = ""
)
func Test2Captcha(t *testing.T) {
solver := NewSolver(API_KEY)
sitekey := "6LfwuyUTAAAAAOAmoS0fdqijC2PbbdH4kjq62Y1b"
url := "https://www.google.com/sorry/index?continue=https://www.google.de/search%3Fhl%3DDE%26lr%3Dlang_de%26nfpr%3D1%26num%3D500%26pws%3D0%26q%3Dwhere%2Bwhy%2Beach&hl=DE&q=EgRegw55GObHiq4GIjDqmzFKayGXrS2-s9ooWfcskhpK8-6tIjWSaSvhxd3f5eAyUXj7lYq2DYLDXB8ASz0yAXJaAUM"
datas := "Ghk0n7ZQNDS0c7ES53eef_YBfSdfeXnyRD0p2OR0R4Dg91CUXKS_hio5Do6TpJ8sHhhOat_NymTASZGe1gqAjP7w9dSvhvRT7QXsrdziO3JPngLDSRzDdjT42GDcSbO0kzInlDPxe1yy2t4yifo9xHpMnlZU7pTVNTQUIXqOMLHAR-iERi6aoSQDQ4d-88-jW3LEinquxEut0OhHG2l2stwG9AnCmNvCsUNJda-H24saFlOh5csK9KNXeeQmpr6at52_skMIMiLXSlY56vYFVCRMkXLQdAM"
resp, err := solver.SolveReCaptcha2(sitekey, url, datas)
if err != nil || resp == "" {
t.Fatalf("Failed to solve recaptchaV2: %s", err)
}
}
+19 -12
View File
@@ -9,14 +9,15 @@ import (
"github.com/gofiber/fiber/v2"
)
var ErrCaptcha = errors.New("Captcha detected")
var ErrSearchTimeout = errors.New("Timeout. Cannot find element on page")
var ErrCaptcha = errors.New("captcha detected")
var ErrSearchTimeout = errors.New("timeout. Cannot find element on page")
type SearchResult struct {
Rank int `json:"rank"`
URL string `json:"url"`
Title string `json:"title"`
Description string `json:"description"`
Ad bool `json:"ad"`
}
func ConvertSearchResultsMap(searchResultsMap map[string]SearchResult) *[]SearchResult {
@@ -39,6 +40,7 @@ type Query struct {
Filetype string // File extension to search.
Site string // Search site
Limit int // Limit the number of results
Answers bool // Include question and answers from SERP page to results with negative indexes
}
func (q Query) IsEmpty() bool {
@@ -48,23 +50,27 @@ func (q Query) IsEmpty() bool {
return false
}
func (q *Query) InitFromContext(c *fiber.Ctx) error {
q.Text = c.Query("text")
q.LangCode = c.Query("lang")
q.DateInterval = c.Query("date")
q.Filetype = c.Query("file")
q.Site = c.Query("site")
func (searchQuery *Query) InitFromContext(reqCtx *fiber.Ctx) error {
searchQuery.Text = reqCtx.Query("text")
searchQuery.LangCode = reqCtx.Query("lang")
searchQuery.DateInterval = reqCtx.Query("date")
searchQuery.Filetype = reqCtx.Query("file")
searchQuery.Site = reqCtx.Query("site")
limit, err := strconv.Atoi(c.Query("limit", "25"))
limit, err := strconv.Atoi(reqCtx.Query("limit", "25"))
if err != nil {
return err
}
q.Limit = limit
searchQuery.Limit = limit
if q.IsEmpty() {
return errors.New("Query cannot be empty")
searchQuery.Answers, err = strconv.ParseBool(reqCtx.Query("answers", "0"))
if err != nil {
return err
}
if searchQuery.IsEmpty() {
return errors.New("Query cannot be empty")
}
return nil
}
@@ -73,6 +79,7 @@ type SearchEngineOptions struct {
RateTime int64 `mapstructure:"rate_seconds"`
RateBurst int `mapstructure:"rate_burst"`
SelectorTimeout int64 `mapstructure:"selector_timeout"` // CSS selector timeout in seconds
IsSolveCaptcha bool `mapstructure:"captcha"`
}
func (o *SearchEngineOptions) Init() {
+4 -5
View File
@@ -2,7 +2,6 @@ package core
import (
"context"
"errors"
"fmt"
"strings"
@@ -52,9 +51,9 @@ func NewServer(host string, port int, searchEngines ...SearchEngine) *Server {
if err != nil {
switch err {
case ErrCaptcha:
err = errors.New(fmt.Sprintf("Captcha found, please stop sending requests for a while\n%s", err))
err = fmt.Errorf("captcha found, please stop sending requests for a while\n%s", err)
case ErrSearchTimeout:
err = errors.New(fmt.Sprintf("%s", err))
err = fmt.Errorf("%s", err)
}
logrus.Errorf("Error during %s search: %s", locEngine.Name(), err)
@@ -87,9 +86,9 @@ func NewServer(host string, port int, searchEngines ...SearchEngine) *Server {
if err != nil {
switch err {
case ErrCaptcha:
err = errors.New(fmt.Sprintf("Captcha found, please stop sending requests for a while\n%s", err))
err = fmt.Errorf("captcha found, please stop sending requests for a while: %s", err)
case ErrSearchTimeout:
err = errors.New(fmt.Sprintf("%s", err))
err = fmt.Errorf("%s", err)
}
logrus.Errorf("Error during %s search: %s", locEngine.Name(), err)
+1
View File
@@ -3,6 +3,7 @@ module github.com/karust/openserp
go 1.20
require (
github.com/2captcha/2captcha-go v1.1.4
github.com/PuerkitoBio/goquery v1.8.1
github.com/corpix/uarand v0.2.0
github.com/go-rod/rod v0.113.3
+2
View File
@@ -36,6 +36,8 @@ cloud.google.com/go/storage v1.8.0/go.mod h1:Wv1Oy7z6Yz3DshWRJFhqM/UCfaWIRTdp0RX
cloud.google.com/go/storage v1.10.0/go.mod h1:FLPqc6j+Ki4BU591ie1oL6qBQGu2Bl/tZ9ullr3+Kg0=
cloud.google.com/go/storage v1.14.0/go.mod h1:GrKmX003DSIwi9o29oFT7YDnHYwZoctc3fOKtUw0Xmo=
dmitri.shuralyov.com/gpu/mtl v0.0.0-20190408044501-666a987793e9/go.mod h1:H6x//7gZCb22OMCxBHrMx7a5I7Hp++hsVxbQ4BYO7hU=
github.com/2captcha/2captcha-go v1.1.4 h1:Fm62VPvVhEHYQ8AI+/uquiTg41ml9f8ASjUkVuBvHcE=
github.com/2captcha/2captcha-go v1.1.4/go.mod h1:hYOq+KVOq/0zAG6OTYW7Y313qDkHv58CcaOyjdBQSco=
github.com/BurntSushi/toml v0.3.1/go.mod h1:xHWCNGjB5oqiDr8zfno3MHue2Ht5sIBksp03qcyfWMU=
github.com/BurntSushi/xgb v0.0.0-20160522181843-27f122750802/go.mod h1:IVnqGOEym/WlBOVXweHU+Q+/VP0lqqI8lqeDx9IjBqo=
github.com/PuerkitoBio/goquery v1.8.1 h1:uQxhNlArOIdbrH1tr0UXwdVFgDcZDrZVdcpygAcwmWM=
+167 -38
View File
@@ -6,6 +6,7 @@ import (
"regexp"
"strconv"
"strings"
"time"
"github.com/go-rod/rod"
"github.com/go-rod/rod/lib/proto"
@@ -24,7 +25,6 @@ func New(browser core.Browser, opts core.SearchEngineOptions) *Google {
gogl := Google{Browser: browser}
opts.Init()
gogl.SearchEngineOptions = opts
gogl.rgxpGetDigits = regexp.MustCompile("\\d")
return &gogl
}
@@ -67,7 +67,26 @@ func (gogl *Google) getTotalResults(page *rod.Page) (int, error) {
return total, nil
}
func (gogl *Google) isCaptcha(page *rod.Page) bool {
func (gogl *Google) solveCaptcha(page *rod.Page, sitekey, datas string) bool {
logrus.Debugf("Solve google Captcha: sitekey=%s, datas=%s, url=%s", sitekey, datas, page.MustInfo().URL)
resp, err := gogl.CaptchaSolver.SolveReCaptcha2(sitekey, page.MustInfo().URL, datas)
if err != nil {
logrus.Errorf("Error solving google captcha: %s", err)
return false
}
logrus.Debug("Resp:", resp)
_, err = page.Eval(fmt.Sprintf(`;(() => { document.getElementById("g-recaptcha-response").innerHTML="%s"; submitCallback(); })();`, resp))
if err != nil {
logrus.Errorf("Error setting captcha response: %s", err)
return false
}
return true
}
func (gogl *Google) checkCaptcha(page *rod.Page) bool {
captchaDiv, err := page.Timeout(gogl.GetSelectorTimeout()).Search("div[data-sitekey]")
if err != nil {
return false
@@ -85,15 +104,34 @@ func (gogl *Google) isCaptcha(page *rod.Page) bool {
return false
}
logrus.Debug("Google Captcha:", *sitekey, *dataS)
if gogl.IsSolveCaptcha {
return !gogl.solveCaptcha(page, *sitekey, *dataS)
}
return true
}
func (gogl *Google) preparePage(page *rod.Page) {
// Remove "similar queries" lists
page.Eval(";(() => { document.querySelectorAll(`div[data-initq]`).forEach( el => el.remove()); })();")
_, err := page.Eval(";(() => { document.querySelectorAll(`div[data-initq]`).forEach( el => el.remove()); })();")
if err != nil {
logrus.Errorf("Error preparing the page: %s", err)
}
}
func (gogl *Google) acceptCookies(page *rod.Page) {
diaglogBtns, err := page.Timeout(gogl.Timeout / 10).Search("div[role='dialog'][aria-modal] button")
if err != nil {
logrus.Errorf("Cannot find cookie consent: %s", err)
return
}
btnElms, err := diaglogBtns.All()
if err != nil {
logrus.Errorf("Cannot get cookie consent buttons: %s", err)
return
}
btnElms[3].Click(proto.InputMouseButtonLeft, 1)
}
func (gogl *Google) Search(query core.Query) ([]core.SearchResult, error) {
logrus.Tracef("Start Google search, query: %+v", query)
@@ -109,19 +147,23 @@ func (gogl *Google) Search(query core.Query) ([]core.SearchResult, error) {
gogl.preparePage(page)
// Check first if there captcha
if gogl.isCaptcha(page) {
if gogl.checkCaptcha(page) {
logrus.Errorf("Google captcha occurred during: %s", url)
return nil, core.ErrCaptcha
}
// Accept cookie consent to get google answers
if query.Answers {
gogl.acceptCookies(page)
}
// Find all results
results, err := page.Timeout(gogl.Timeout).Search("div[data-hveid][data-ved][lang], div[data-surl][jsaction]")
results, err := page.Timeout(gogl.Timeout).Search("div[data-hveid]")
if err != nil {
logrus.Errorf("Cannot parse search results: %s", err)
return nil, core.ErrSearchTimeout
}
// Check why no results, maybe captcha?
if results == nil {
return nil, nil
}
@@ -132,47 +174,134 @@ func (gogl *Google) Search(query core.Query) ([]core.SearchResult, error) {
}
logrus.Infof("%d SERP results", totalResults)
resultElements, err := results.All()
searchResultElems, err := results.All()
if err != nil {
return nil, err
}
for i, r := range resultElements {
// Get URL
link, err := r.Element("a")
if err != nil {
rank := 0
for _, resEl := range searchResultElems {
srchRes := core.SearchResult{}
attrs := strings.Join(resEl.MustDescribe().Attributes, " ")
if strings.Contains(attrs, "data-text-ad") {
// 1. Parse ads
srchRes.Ad = true
// Get URL
link, err := resEl.Element("a")
if err != nil {
logrus.Debug("No link found")
continue
}
link.MoveMouseOut()
href, err := link.Property("href")
if err != nil {
logrus.Debug("No `href` tag found")
continue
}
srchRes.URL = href.String()
// Get title
srchRes.Title = link.MustText()
// Get description
text := resEl.MustText()
textSliced := strings.Split(text, "\n")
srchRes.Description = strings.Join(textSliced[4:], "\n")
rank += 1
} else if query.Answers && strings.Contains(attrs, "data-ulkwtsb") && !strings.Contains(attrs, "data-ispaa") {
// 2. Parse answer boxes
answerEls, err := resEl.Page().Search("div[data-hveid][data-ulkwtsb] div[data-q]")
if err != nil {
logrus.Debugf("Error while parsing answer box 1: %s", err.Error())
continue
}
answers, err := answerEls.All()
if err != nil {
logrus.Debugf("Error while parsing answer box 2: %s", err.Error())
continue
}
logrus.Infof("%d answers found", len(answers))
// Unvail answer contents
for _, answ := range answers {
answ.Click(proto.InputMouseButtonLeft, 1)
answ.Focus()
//answ.Page().WaitRepaint()
}
time.Sleep(time.Millisecond * 2000)
for i, answ := range answers {
answerText := strings.Split(answ.MustText(), "\n")
if len(answerText) < 2 {
logrus.Debugf("Short answer text: %s", answerText)
continue
}
// Get URL
link, err := answ.Element("a")
if err != nil {
logrus.Debug("No answer link found")
continue
}
link.MoveMouseOut()
href, err := link.Property("href")
if err != nil {
logrus.Debug("No answer `href` tag found")
continue
}
srchRes.URL = href.String()
srchRes.Title = answerText[0]
srchRes.Description = strings.Join(answerText[1:len(answerText)-2], "\n")
srchRes.Rank = -1 * (i + 1)
searchResults = append(searchResults, srchRes)
}
continue
}
linkText, err := link.Property("href")
if err != nil {
logrus.Error("No `href` tag found")
}
} else if strings.Contains(attrs, "data-ved") && strings.Contains(attrs, "lang") {
// 3. Parse regular search results
// Get URL
link, err := resEl.Element("a")
if err != nil {
continue
}
href, err := link.Property("href")
if err != nil {
logrus.Debug("No `href` tag found")
}
srchRes.URL = href.String()
rank += 1
// Get title
titleTag, err := link.Element("h3")
if err != nil {
logrus.Error("No `h3` tag found")
continue
}
// Get title
titleTag, err := link.Element("h3")
if err != nil {
continue
}
title, err := titleTag.Text()
if err != nil {
logrus.Error("Cannot extract text from title")
title = "No title"
}
srchRes.Title, err = titleTag.Text()
if err != nil {
logrus.Debug("Cannot extract text from title")
}
// Get description
text := resEl.MustText()
textSliced := strings.Split(text, "\n")
srchRes.Description = strings.Join(textSliced[4:], "\n")
// Get description
// doesn't catch all
descTag, err := r.Element(`div[data-sncf~="1"]`)
desc := ""
if err != nil {
logrus.Trace(`No description 'div[data-sncf~="1"]' tag found`)
} else {
desc = descTag.MustText()
//fmt.Println(i, attrs)
continue
}
gR := core.SearchResult{Rank: i + 1, URL: linkText.String(), Title: title, Description: desc}
searchResults = append(searchResults, gR)
srchRes.Rank = rank
searchResults = append(searchResults, srchRes)
}
return searchResults, nil
@@ -212,7 +341,7 @@ func (gogl *Google) SearchImage(query core.Query) ([]core.SearchResult, error) {
// Check why no results
if results == nil {
if gogl.isCaptcha(page) {
if gogl.checkCaptcha(page) {
logrus.Errorf("Google captcha occurred during: %s", url)
return *core.ConvertSearchResultsMap(searchResultsMap), core.ErrCaptcha
}
+1 -1
View File
@@ -10,7 +10,7 @@ import (
var browser *core.Browser
func init() {
opts := core.BrowserOpts{IsHeadless: false, IsLeakless: false, Timeout: time.Second * 5, LeavePageOpen: true}
opts := core.BrowserOpts{IsHeadless: true, IsLeakless: false, Timeout: time.Second * 5, LeavePageOpen: false}
browser, _ = core.NewBrowser(opts)
}