2018-11-29 17:33:44 +02:00
|
|
|
package twitterscraper
|
|
|
|
|
|
|
|
|
|
import (
|
2020-06-12 21:31:08 +08:00
|
|
|
"context"
|
2021-03-09 10:40:22 +02:00
|
|
|
"fmt"
|
2023-06-01 23:20:11 +03:00
|
|
|
"net/url"
|
2023-07-02 01:41:48 +03:00
|
|
|
"strconv"
|
2018-11-29 17:33:44 +02:00
|
|
|
)
|
|
|
|
|
|
2020-05-14 21:52:55 +03:00
|
|
|
// GetTweets returns channel with tweets for a given user.
|
2021-04-22 21:38:49 +03:00
|
|
|
func (s *Scraper) GetTweets(ctx context.Context, user string, maxTweetsNbr int) <-chan *TweetResult {
|
|
|
|
|
return getTweetTimeline(ctx, user, maxTweetsNbr, s.FetchTweets)
|
2020-12-12 23:33:57 +02:00
|
|
|
}
|
|
|
|
|
|
2020-05-14 21:52:55 +03:00
|
|
|
// FetchTweets gets tweets for a given user, via the Twitter frontend API.
|
2020-12-12 23:33:57 +02:00
|
|
|
func (s *Scraper) FetchTweets(user string, maxTweetsNbr int, cursor string) ([]*Tweet, string, error) {
|
|
|
|
|
userID, err := s.GetUserIDByScreenName(user)
|
2018-11-29 17:33:44 +02:00
|
|
|
if err != nil {
|
2020-12-11 20:58:49 +02:00
|
|
|
return nil, "", err
|
2018-11-29 17:33:44 +02:00
|
|
|
}
|
|
|
|
|
|
2023-07-02 01:41:48 +03:00
|
|
|
if s.isOpenAccount {
|
|
|
|
|
return s.FetchTweetsByUserIDLegacy(userID, maxTweetsNbr, cursor)
|
|
|
|
|
}
|
2023-05-10 06:09:11 -07:00
|
|
|
return s.FetchTweetsByUserID(userID, maxTweetsNbr, cursor)
|
|
|
|
|
}
|
|
|
|
|
|
2023-06-01 23:20:11 +03:00
|
|
|
// FetchTweetsByUserID gets tweets for a given userID, via the Twitter frontend GraphQL API.
|
2023-05-10 06:09:11 -07:00
|
|
|
func (s *Scraper) FetchTweetsByUserID(userID string, maxTweetsNbr int, cursor string) ([]*Tweet, string, error) {
|
|
|
|
|
if maxTweetsNbr > 200 {
|
|
|
|
|
maxTweetsNbr = 200
|
|
|
|
|
}
|
|
|
|
|
|
2023-06-01 23:20:11 +03:00
|
|
|
req, err := s.newRequest("GET", "https://twitter.com/i/api/graphql/UGi7tjRPr-d_U3bCPIko5Q/UserTweets")
|
2020-05-13 17:35:44 +02:00
|
|
|
if err != nil {
|
2020-12-11 20:58:49 +02:00
|
|
|
return nil, "", err
|
2020-05-13 17:35:44 +02:00
|
|
|
}
|
|
|
|
|
|
2023-06-01 23:20:11 +03:00
|
|
|
variables := map[string]interface{}{
|
|
|
|
|
"userId": userID,
|
|
|
|
|
"count": maxTweetsNbr,
|
|
|
|
|
"includePromotedContent": false,
|
|
|
|
|
"withQuickPromoteEligibilityTweetFields": false,
|
|
|
|
|
"withVoice": true,
|
|
|
|
|
"withV2Timeline": true,
|
|
|
|
|
}
|
|
|
|
|
features := map[string]interface{}{
|
|
|
|
|
"rweb_lists_timeline_redesign_enabled": true,
|
|
|
|
|
"responsive_web_graphql_exclude_directive_enabled": true,
|
|
|
|
|
"verified_phone_label_enabled": false,
|
|
|
|
|
"creator_subscriptions_tweet_preview_api_enabled": true,
|
|
|
|
|
"responsive_web_graphql_timeline_navigation_enabled": true,
|
|
|
|
|
"responsive_web_graphql_skip_user_profile_image_extensions_enabled": false,
|
|
|
|
|
"tweetypie_unmention_optimization_enabled": true,
|
|
|
|
|
"vibe_api_enabled": true,
|
|
|
|
|
"responsive_web_edit_tweet_api_enabled": true,
|
|
|
|
|
"graphql_is_translatable_rweb_tweet_is_translatable_enabled": true,
|
|
|
|
|
"view_counts_everywhere_api_enabled": true,
|
|
|
|
|
"longform_notetweets_consumption_enabled": true,
|
|
|
|
|
"tweet_awards_web_tipping_enabled": false,
|
|
|
|
|
"freedom_of_speech_not_reach_fetch_enabled": true,
|
|
|
|
|
"standardized_nudges_misinfo": true,
|
|
|
|
|
"tweet_with_visibility_results_prefer_gql_limited_actions_policy_enabled": false,
|
|
|
|
|
"interactive_text_enabled": true,
|
|
|
|
|
"responsive_web_text_conversations_enabled": false,
|
|
|
|
|
"longform_notetweets_rich_text_read_enabled": true,
|
|
|
|
|
"longform_notetweets_inline_media_enabled": false,
|
|
|
|
|
"responsive_web_enhance_cards_enabled": false,
|
|
|
|
|
}
|
|
|
|
|
|
2020-12-11 20:58:49 +02:00
|
|
|
if cursor != "" {
|
2023-06-01 23:20:11 +03:00
|
|
|
variables["cursor"] = cursor
|
2020-12-11 20:58:49 +02:00
|
|
|
}
|
2020-05-13 17:35:44 +02:00
|
|
|
|
2023-06-01 23:20:11 +03:00
|
|
|
query := url.Values{}
|
|
|
|
|
query.Set("variables", mapToJSONString(variables))
|
|
|
|
|
query.Set("features", mapToJSONString(features))
|
|
|
|
|
req.URL.RawQuery = query.Encode()
|
|
|
|
|
|
|
|
|
|
var timeline timelineV2
|
2020-12-12 23:33:57 +02:00
|
|
|
err = s.RequestAPI(req, &timeline)
|
2020-02-12 10:45:19 +02:00
|
|
|
if err != nil {
|
2020-12-11 20:58:49 +02:00
|
|
|
return nil, "", err
|
2020-02-12 10:45:19 +02:00
|
|
|
}
|
2018-11-29 17:33:44 +02:00
|
|
|
|
2021-07-16 11:08:43 +03:00
|
|
|
tweets, nextCursor := timeline.parseTweets()
|
2020-12-11 20:58:49 +02:00
|
|
|
return tweets, nextCursor, nil
|
2018-11-29 17:33:44 +02:00
|
|
|
}
|
2021-03-09 10:40:22 +02:00
|
|
|
|
2023-07-02 01:41:48 +03:00
|
|
|
// FetchTweetsByUserIDLegacy gets tweets for a given userID, via the Twitter frontend legacy API.
|
|
|
|
|
func (s *Scraper) FetchTweetsByUserIDLegacy(userID string, maxTweetsNbr int, cursor string) ([]*Tweet, string, error) {
|
|
|
|
|
if maxTweetsNbr > 200 {
|
|
|
|
|
maxTweetsNbr = 200
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
req, err := s.newRequest("GET", "https://api.twitter.com/2/timeline/profile/"+userID+".json")
|
2021-03-09 10:40:22 +02:00
|
|
|
if err != nil {
|
2023-07-02 01:41:48 +03:00
|
|
|
return nil, "", err
|
2021-03-09 10:40:22 +02:00
|
|
|
}
|
|
|
|
|
|
2023-07-02 01:41:48 +03:00
|
|
|
q := req.URL.Query()
|
|
|
|
|
q.Add("count", strconv.Itoa(maxTweetsNbr))
|
|
|
|
|
q.Add("userId", userID)
|
|
|
|
|
if cursor != "" {
|
|
|
|
|
q.Add("cursor", cursor)
|
2023-06-01 23:20:11 +03:00
|
|
|
}
|
2023-07-02 01:41:48 +03:00
|
|
|
req.URL.RawQuery = q.Encode()
|
2023-06-01 23:20:11 +03:00
|
|
|
|
2023-07-02 01:41:48 +03:00
|
|
|
var timeline timelineV1
|
|
|
|
|
err = s.RequestAPI(req, &timeline)
|
|
|
|
|
if err != nil {
|
|
|
|
|
return nil, "", err
|
2023-06-01 23:20:11 +03:00
|
|
|
}
|
|
|
|
|
|
2023-07-02 01:41:48 +03:00
|
|
|
tweets, nextCursor := timeline.parseTweets()
|
|
|
|
|
return tweets, nextCursor, nil
|
|
|
|
|
}
|
2023-06-01 23:20:11 +03:00
|
|
|
|
2023-07-02 01:41:48 +03:00
|
|
|
// GetTweet get a single tweet by ID.
|
|
|
|
|
func (s *Scraper) GetTweet(id string) (*Tweet, error) {
|
|
|
|
|
if s.isOpenAccount {
|
|
|
|
|
req, err := s.newRequest("GET", "https://api.twitter.com/2/timeline/conversation/"+id+".json")
|
|
|
|
|
if err != nil {
|
|
|
|
|
return nil, err
|
|
|
|
|
}
|
2023-06-18 19:13:26 +02:00
|
|
|
|
2023-07-02 01:41:48 +03:00
|
|
|
var timeline timelineV1
|
|
|
|
|
err = s.RequestAPI(req, &timeline)
|
|
|
|
|
if err != nil {
|
|
|
|
|
return nil, err
|
|
|
|
|
}
|
2023-06-18 19:13:26 +02:00
|
|
|
|
2023-07-02 01:41:48 +03:00
|
|
|
tweets, _ := timeline.parseTweets()
|
|
|
|
|
for _, tweet := range tweets {
|
|
|
|
|
if tweet.ID == id {
|
|
|
|
|
return tweet, nil
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
} else {
|
|
|
|
|
req, err := s.newRequest("GET", "https://twitter.com/i/api/graphql/VWFGPVAGkZMGRKGe3GFFnA/TweetDetail")
|
|
|
|
|
if err != nil {
|
|
|
|
|
return nil, err
|
|
|
|
|
}
|
2023-06-18 19:13:26 +02:00
|
|
|
|
2023-07-02 01:41:48 +03:00
|
|
|
variables := map[string]interface{}{
|
|
|
|
|
"focalTweetId": id,
|
|
|
|
|
"with_rux_injections": false,
|
|
|
|
|
"includePromotedContent": true,
|
|
|
|
|
"withCommunity": true,
|
|
|
|
|
"withQuickPromoteEligibilityTweetFields": true,
|
|
|
|
|
"withBirdwatchNotes": true,
|
|
|
|
|
"withVoice": true,
|
|
|
|
|
"withV2Timeline": true,
|
|
|
|
|
}
|
2023-06-18 19:13:26 +02:00
|
|
|
|
2023-07-02 01:41:48 +03:00
|
|
|
features := map[string]interface{}{
|
|
|
|
|
"rweb_lists_timeline_redesign_enabled": true,
|
|
|
|
|
"responsive_web_graphql_exclude_directive_enabled": true,
|
|
|
|
|
"verified_phone_label_enabled": false,
|
|
|
|
|
"creator_subscriptions_tweet_preview_api_enabled": true,
|
|
|
|
|
"responsive_web_graphql_timeline_navigation_enabled": true,
|
|
|
|
|
"responsive_web_graphql_skip_user_profile_image_extensions_enabled": false,
|
|
|
|
|
"tweetypie_unmention_optimization_enabled": true,
|
|
|
|
|
"responsive_web_edit_tweet_api_enabled": true,
|
|
|
|
|
"graphql_is_translatable_rweb_tweet_is_translatable_enabled": true,
|
|
|
|
|
"view_counts_everywhere_api_enabled": true,
|
|
|
|
|
"longform_notetweets_consumption_enabled": true,
|
|
|
|
|
"tweet_awards_web_tipping_enabled": false,
|
|
|
|
|
"freedom_of_speech_not_reach_fetch_enabled": true,
|
|
|
|
|
"standardized_nudges_misinfo": true,
|
|
|
|
|
"tweet_with_visibility_results_prefer_gql_limited_actions_policy_enabled": false,
|
|
|
|
|
"longform_notetweets_rich_text_read_enabled": true,
|
|
|
|
|
"longform_notetweets_inline_media_enabled": true,
|
|
|
|
|
"responsive_web_enhance_cards_enabled": false,
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
query := url.Values{}
|
|
|
|
|
query.Set("variables", mapToJSONString(variables))
|
|
|
|
|
query.Set("features", mapToJSONString(features))
|
|
|
|
|
req.URL.RawQuery = query.Encode()
|
|
|
|
|
|
|
|
|
|
var conversation threadedConversation
|
|
|
|
|
|
|
|
|
|
// Surprisingly, if bearerToken2 is not set, then animated GIFs are not
|
|
|
|
|
// present in the response for tweets with a GIF + a photo like this one:
|
|
|
|
|
// https://twitter.com/Twitter/status/1580661436132757506
|
|
|
|
|
curBearerToken := s.bearerToken
|
|
|
|
|
if curBearerToken != bearerToken2 {
|
|
|
|
|
s.setBearerToken(bearerToken2)
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
err = s.RequestAPI(req, &conversation)
|
|
|
|
|
|
|
|
|
|
if curBearerToken != bearerToken2 {
|
|
|
|
|
s.setBearerToken(curBearerToken)
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if err != nil {
|
|
|
|
|
return nil, err
|
|
|
|
|
}
|
2021-03-09 10:40:22 +02:00
|
|
|
|
2023-07-02 01:41:48 +03:00
|
|
|
tweets := conversation.parse()
|
|
|
|
|
for _, tweet := range tweets {
|
|
|
|
|
if tweet.ID == id {
|
|
|
|
|
return tweet, nil
|
|
|
|
|
}
|
2021-03-09 10:40:22 +02:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
return nil, fmt.Errorf("tweet with ID %s not found", id)
|
|
|
|
|
}
|