Merge pull request #1661 from liimee/eng-tw

Add twitter engine

2 files changed, 80 insertions, 0 deletions

diff --git a/searx/engines/twitter.py b/searx/engines/twitter.py
new file mode 100644
index 000000000..3ebe34be1
--- /dev/null
+++ b/searx/engines/twitter.py
@@ -0,0 +1,75 @@
+# SPDX-License-Identifier: AGPL-3.0-or-later
+# lint: pylint
+"""Twitter (microblogging platform)"""
+
+from json import loads
+from urllib.parse import urlencode
+from datetime import datetime
+
+about = {
+    "website": 'https://twitter.com',
+    "wikidata_id": None,
+    "official_api_documentation": 'https://developer.twitter.com/en/docs/twitter-api',
+    "use_official_api": True,
+    "require_api_key": False,
+    "results": 'JSON',
+}
+
+categories = ['social media']
+
+url = "https://api.twitter.com"
+search_url = (
+    "{url}/2/search/adaptive.json?{query}&tweet_mode=extended&query_source=typed_query&pc=1&spelling_corrections=1"
+)
+
+
+def request(query, params):
+    params['url'] = search_url.format(url=url, query=urlencode({'q': query}))
+
+    params['headers'] = {
+        # This token is used in the Twitter web interface (twitter.com). Without this header, the API doesn't work.
+        # The value of the token has never changed (or maybe once a long time ago).
+        # https://github.com/zedeus/nitter/blob/5f31e86e0e8578377fa7d5aeb9631bbb2d35ef1e/src/consts.nim#L5
+        'Authorization': (
+            "Bearer AAAAAAAAAAAAAAAAAAAAAPYXBAAAAAAACLXUNDekMxqa8h%2F40K4moUkGsoc%3DTYfbDKb"
+            "T3jJPCEVnMYqilB28NHfOPqkca3qaAxGfsyKCs0wRbw"
+        )
+    }
+
+    return params
+
+
+def response(resp):
+    results = []
+
+    json_res = loads(resp.text)['globalObjects']
+
+    for tweet in json_res['tweets'].values():
+        text = tweet['full_text']
+        display = tweet['display_text_range']
+
+        img_src = tweet.get('extended_entities', {}).get('media', [{}])[0].get('media_url_https')
+        if img_src:
+            img_src += "?name=thumb"
+
+        results.append(
+            {
+                'url': 'https://twitter.com/i/web/status/' + tweet['id_str'],
+                'title': (text[:40] + '...') if len(text) > 40 else text,
+                'content': text[display[0] : display[1]],
+                'img_src': img_src,
+                'publishedDate': datetime.strptime(tweet['created_at'], '%a %b %d %H:%M:%S %z %Y'),
+            }
+        )
+
+    for user in json_res['users'].values():
+        results.append(
+            {
+                'title': user['name'],
+                'content': user['description'],
+                'url': 'https://twitter.com/' + user['screen_name'],
+                'img_src': user['profile_image_url_https'],
+            }
+        )
+
+    return results
diff --git a/searx/settings.yml b/searx/settings.yml
index 9530a4420..a61be1071 100644
--- a/searx/settings.yml
+++ b/searx/settings.yml
@@ -1425,6 +1425,11 @@ engines:
   #     - 2000
   #     - 5000
 
+  - name: twitter
+    shortcut: tw
+    engine: twitter
+    disabled: true
+
   # maybe in a fun category
   #  - name: uncyclopedia
   #    engine: mediawiki