[pull] master from sherlock-project:master

docs/removed-sites.md

@@ -1892,4 +1892,81 @@ __2024-06-10 :__ Http request returns 403 forbidden, and tries to verify the con
 ```
 
 ## Alik.cz
-__2024-07-21 :__ Target is now BLACKLISTED from the default manifest due to the site recieving unnecessarily high traffic from Sherlock. This target is not permitted to be reactivited. Inclusion in unrelated manifests is not impacted, but it is discouraged.
+__2024-07-21 :__ Target is now BLACKLISTED from the default manifest due to the site recieving unnecessarily high traffic from Sherlock (by request of the site owners). This target is not permitted to be reactivited. Inclusion in unrelated manifests is not impacted, but it is discouraged.
+
+## 8tracks
+__2025-02-02 :__ Might be dead again. Nobody knows for sure.
+```json
+"8tracks": {
+    "errorType": "message",
+    "errorMsg": "\"available\":true",
+    "headers": {
+      "Accept-Language": "en-US,en;q=0.5"
+    },
+    "url": "https://8tracks.com/{}",
+    "urlProbe": "https://8tracks.com/users/check_username?login={}&format=jsonh",
+    "urlMain": "https://8tracks.com/",
+    "username_claimed": "blue"
+  }
+```
+
+## Shpock
+__2025-02-02 :__ Can likely be added back with a new endpoint (source username availability endpoint from mobile app reg flow?)
+```json
+"Shpock": {
+    "errorType": "status_code",
+    "url": "https://www.shpock.com/shop/{}/items",
+    "urlMain": "https://www.shpock.com/",
+    "username_claimed": "user"
+  }
+```
+
+## Twitch
+__2025-02-02 :__
+```json
+"Twitch": {
+    "errorType": "message",
+    "errorMsg": "components.availability-tracking.warn-unavailable.component",
+    "url": "https://www.twitch.tv/{}",
+    "urlMain": "https://www.twitch.tv/",
+    "urlProbe": "https://m.twitch.tv/{}",
+    "username_claimed": "jenny"
+  }
+```
+
+## Fiverr
+__2025-02-02 :__ Fiverr added CSRF protections that messed with this test
+```json
+"Fiverr": {
+    "errorMsg": "\"status\":\"success\"",
+    "errorType": "message",
+    "headers": {
+      "Content-Type": "application/json",
+      "Accept-Language": "en-US,en;q=0.9"
+    },
+    "regexCheck": "^[A-Za-z][A-Za-z\\d_]{5,14}$",
+    "request_method": "POST",
+    "request_payload": {
+      "username": "{}"
+    },
+    "url": "https://www.fiverr.com/{}",
+    "urlMain": "https://www.fiverr.com/",
+    "urlProbe": "https://www.fiverr.com/validate_username",
+    "username_claimed": "blueman"
+  }
+```
+
+## BabyRU
+__2025-02-02 :__ Just being problematic
+```json
+"babyRU": {
+    "errorMsg": [
+      "\u0421\u0442\u0440\u0430\u043d\u0438\u0446\u0430, \u043a\u043e\u0442\u043e\u0440\u0443\u044e \u0432\u044b \u0438\u0441\u043a\u0430\u043b\u0438, \u043d\u0435 \u043d\u0430\u0439\u0434\u0435\u043d\u0430",
+      "Доступ с вашего IP-адреса временно ограничен"
+    ],
+    "errorType": "message",
+    "url": "https://www.baby.ru/u/{}/",
+    "urlMain": "https://www.baby.ru/",
+    "username_claimed": "blue"
+  }
+```

sherlock_project/resources/data.json

@@ -30,17 +30,6 @@
     "urlMain": "https://www.7cups.com/",
     "username_claimed": "blue"
   },
-  "8tracks": {
-    "errorType": "message",
-    "errorMsg": "\"available\":true",
-    "headers": {
-      "Accept-Language": "en-US,en;q=0.5"
-    },
-    "url": "https://8tracks.com/{}",
-    "urlProbe": "https://8tracks.com/users/check_username?login={}&format=jsonh",
-    "urlMain": "https://8tracks.com/",
-    "username_claimed": "blue"
-  },
   "9GAG": {
     "errorType": "status_code",
     "url": "https://www.9gag.com/u/{}",
@@ -76,13 +65,6 @@
     "urlMain": "https://admireme.vip/",
     "username_claimed": "DemiDevil"
   },
-  "Air Pilot Life": {
-    "errorMsg": "Oops! That page doesn\u2019t exist or is private",
-    "errorType": "message",
-    "url": "https://airlinepilot.life/u/{}",
-    "urlMain": "https://airlinepilot.life/",
-    "username_claimed": "chris"
-  },
   "Airbit": {
     "errorType": "status_code",
     "url": "https://airbit.com/{}",
@@ -628,6 +610,14 @@
     "urlMain": "https://deviantart.com",
     "username_claimed": "blue"
   },
+  "DigitalSpy": {
+      "errorMsg": "The page you were looking for could not be found.",
+      "errorType": "message",
+      "url": "https://forums.digitalspy.com/profile/{}",
+      "urlMain": "https://forums.digitalspy.com/",
+      "username_claimed": "blue",
+      "regexCheck": "^\\w{3,20}$"
+    },
   "Discogs": {
     "errorType": "status_code",
     "url": "https://www.discogs.com/user/{}",
@@ -764,23 +754,6 @@
     "urlMain": "https://www.finanzfrage.net/",
     "username_claimed": "finanzfrage"
   },
-  "Fiverr": {
-    "errorMsg": "\"status\":\"success\"",
-    "errorType": "message",
-    "headers": {
-      "Content-Type": "application/json",
-      "Accept-Language": "en-US,en;q=0.9"
-    },
-    "regexCheck": "^[A-Za-z][A-Za-z\\d_]{5,14}$",
-    "request_method": "POST",
-    "request_payload": {
-      "username": "{}"
-    },
-    "url": "https://www.fiverr.com/{}",
-    "urlMain": "https://www.fiverr.com/",
-    "urlProbe": "https://www.fiverr.com/validate_username",
-    "username_claimed": "blueman"
-  },
   "Flickr": {
     "errorType": "status_code",
     "url": "https://www.flickr.com/people/{}",
@@ -911,10 +884,11 @@
     "username_claimed": "bob"
   },
   "Giphy": {
-    "errorType": "status_code",
+    "errorType": "message",
+    "errorMsg": "<title> GIFs - Find &amp; Share on GIPHY</title>",
     "url": "https://giphy.com/{}",
     "urlMain": "https://giphy.com/",
-    "username_claimed": "blue"
+    "username_claimed": "red"
   },
   "GitBook": {
     "errorType": "status_code",
@@ -1109,6 +1083,12 @@
     "urlMain": "https://hudsonrock.com",
     "username_claimed": "testadmin"
   },
+  "Hugging Face": {
+    "errorType": "status_code",
+    "url": "https://huggingface.co/{}",
+    "urlMain": "https://huggingface.co/",
+    "username_claimed": "Pasanlaksitha"
+  },
   "IFTTT": {
     "errorType": "status_code",
     "regexCheck": "^[A-Za-z0-9]{3,35}$",
@@ -1676,8 +1656,7 @@
     "username_claimed": "blue"
   },
   "ProductHunt": {
-    "errorMsg": "We seem to have lost this page",
-    "errorType": "message",
+    "errorType": "status_code",
     "url": "https://www.producthunt.com/@{}",
     "urlMain": "https://www.producthunt.com/",
     "username_claimed": "jenny"
@@ -1844,12 +1823,6 @@
     "urlMain": "https://www.shitpostbot.com/",
     "username_claimed": "blue"
   },
-  "Shpock": {
-    "errorType": "status_code",
-    "url": "https://www.shpock.com/shop/{}/items",
-    "urlMain": "https://www.shpock.com/",
-    "username_claimed": "user"
-  },
   "Signal": {
     "errorMsg": "Oops! That page doesn\u2019t exist or is private.",
     "errorType": "message",
@@ -1885,7 +1858,8 @@
     "username_claimed": "blue"
   },
   "SlideShare": {
-    "errorType": "status_code",
+    "errorType": "message",
+    "errorMsg": "<title>Username available</title>",
     "url": "https://slideshare.net/{}",
     "urlMain": "https://slideshare.net/",
     "username_claimed": "blue"
@@ -2036,13 +2010,6 @@
     "errorType": "status_code",
     "username_claimed": "blue"
   },
-  "TLDR Legal": {
-    "errorType": "status_code",
-    "regexCheck": "^[a-zA-Z0-9]{3,20}$",
-    "url": "https://tldrlegal.com/users/{}/",
-    "urlMain": "https://tldrlegal.com/",
-    "username_claimed": "kevin"
-  },
   "Topcoder": {
     "errorType": "status_code",
     "url": "https://profiles.topcoder.com/{}/",
@@ -2159,15 +2126,11 @@
     "urlMain": "https://tweakers.net",
     "username_claimed": "femme"
   },
-  "Twitch": {
-    "errorType": "status_code",
-    "url": "https://www.twitch.tv/{}",
-    "urlMain": "https://www.twitch.tv/",
-    "urlProbe": "https://m.twitch.tv/{}",
-    "username_claimed": "jenny"
-  },
   "Twitter": {
-    "errorMsg": "<div class=\"error-panel\"><span>User ",
+    "errorMsg": [
+      "<div class=\"error-panel\"><span>User ",
+      "<title>429 Too Many Requests</title>"
+    ],
     "errorType": "message",
     "regexCheck": "^[a-zA-Z0-9_]{1,15}$",
     "url": "https://x.com/{}",
@@ -2436,13 +2399,6 @@
     "urlMain": "http://www.authorstream.com/",
     "username_claimed": "blue"
   },
-  "babyRU": {
-    "errorMsg": "\u0421\u0442\u0440\u0430\u043d\u0438\u0446\u0430, \u043a\u043e\u0442\u043e\u0440\u0443\u044e \u0432\u044b \u0438\u0441\u043a\u0430\u043b\u0438, \u043d\u0435 \u043d\u0430\u0439\u0434\u0435\u043d\u0430",
-    "errorType": "message",
-    "url": "https://www.baby.ru/u/{}/",
-    "urlMain": "https://www.baby.ru/",
-    "username_claimed": "blue"
-  },
   "babyblogRU": {
     "errorType": "response_url",
     "errorUrl": "https://www.babyblog.ru/",
@@ -2853,6 +2809,13 @@
     "urlMain": "https://znanylekarz.pl",
     "username_claimed": "janusz-nowak"
   },
+  "v0.dev": {
+    "errorType": "message",
+    "errorMsg": "<title>v0 by Vercel</title>",
+    "url": "https://v0.dev/{}",
+    "urlMain": "https://v0.dev",
+    "username_claimed": "t3dotgg"
+  },
   "Bluesky": {
     "errorType": "status_code",
     "url": "https://bsky.app/profile/{}.bsky.social",

sherlock_project/sherlock.py

@@ -656,7 +656,7 @@ def main():
         metavar="JSON_FILE",
         dest="json_file",
         default=None,
-        help="Load data from a JSON file or an online, valid, JSON file.",
+        help="Load data from a JSON file or an online, valid, JSON file. Upstream PR numbers also accepted.",
     )
     parser.add_argument(
         "--timeout",