Merge pull request #111 from bershadskiy/sanitize-YouTube-links

#101
Lurk · May 14, 2024 · 9f8eeca · 9f8eeca
2 parents df28a63 + 21c634b
commit 9f8eeca
Show file tree

Hide file tree

Showing 3 changed files with 108 additions and 9 deletions.
diff --git a/readme.md b/readme.md
@@ -53,6 +53,8 @@ params = [
     'utm_medium', # Identifies what type of link was used
     'utm_term', # Identifies search terms
     'utm_content', # Identifies what specifically was clicked to bring the user to the site
+    "youtube.com``si", # YouTube specific Source identifier using "{domain}``{param}" pattern
+    "youtu.be``si", # YouTube specific Source identifier using "{domain}``{param}" pattern
 ]
 # Which exit params in URL should be unwrapped
 exit = [

diff --git a/src/clink.rs b/src/clink.rs
@@ -35,7 +35,13 @@ impl Clink {
         let mut res = str.to_string();
         for link in self.finder.links(str) {
             let mut l = Url::parse(self.unwrap_exit_params(link.as_str()).as_str()).unwrap();
-            let query: Vec<(_, _)> = self.process_query(l.query_pairs());
+            let query: Vec<(_, _)> = self.process_query(
+                l.query_pairs(),
+                l.domain()
+                    .unwrap_or_default()
+                    .strip_prefix("www.")
+                    .or(l.domain()),
+            );
             l.set_query(None);
             for pair in query {
                 l.query_pairs_mut()
@@ -46,16 +52,20 @@ impl Clink {
         res
     }
 
-    fn process_query(&self, query: url::form_urlencoded::Parse<'_>) -> Vec<(String, String)> {
+    fn process_query(
+        &self,
+        query: url::form_urlencoded::Parse<'_>,
+        domain: Option<&str>,
+    ) -> Vec<(String, String)> {
         match self.config.mode {
-            Mode::Remove => self.filter(query),
-            Mode::Replace => self.replace(query),
+            Mode::Remove => self.filter(query, domain),
+            Mode::Replace => self.replace(query, domain),
             Mode::YourMom => {
                 let date = Utc::now();
                 if date.month() == 5 && date.day() == 9 {
-                    self.filter(query)
+                    self.filter(query, domain)
                 } else {
-                    let mut tmp = self.filter(query);
+                    let mut tmp = self.filter(query, domain);
                     tmp.push(("utm_source".to_string(), "your_mom".to_string()));
                     tmp
                 }
@@ -82,18 +92,42 @@ impl Clink {
         }
     }
 
-    fn filter(&self, query: url::form_urlencoded::Parse<'_>) -> Vec<(String, String)> {
+    fn filter(
+        &self,
+        query: url::form_urlencoded::Parse<'_>,
+        domain: Option<&str>,
+    ) -> Vec<(String, String)> {
         query
-            .filter(|p| !self.config.params.contains::<Rc<str>>(&p.0.clone().into()))
+            .filter(|p| {
+                let global_absent = !self.config.params.contains::<Rc<str>>(&p.0.clone().into());
+                global_absent
+                    && if let Some(domain_val) = domain {
+                        let param_name = format!("{}``{}", domain_val, p.0);
+                        return !self.config.params.contains::<Rc<str>>(&param_name.into());
+                    } else {
+                        true
+                    }
+            })
             .map(|p| (p.0.to_string(), p.1.to_string()))
             .collect()
     }
 
-    fn replace(&self, query: url::form_urlencoded::Parse<'_>) -> Vec<(String, String)> {
+    fn replace(
+        &self,
+        query: url::form_urlencoded::Parse<'_>,
+        domain: Option<&str>,
+    ) -> Vec<(String, String)> {
         query
             .map(|p| {
                 if self.config.params.contains::<Rc<str>>(&p.0.clone().into()) {
                     (p.0.to_string(), self.config.replace_to.clone())
+                } else if let Some(domain_val) = domain {
+                    let param_name = format!("{}``{}", domain_val, p.0);
+                    if self.config.params.contains::<Rc<str>>(&param_name.into()) {
+                        (p.0.to_string(), self.config.replace_to.clone())
+                    } else {
+                        (p.0.to_string(), p.1.to_string())
+                    }
                 } else {
                     (p.0.to_string(), p.1.to_string())
                 }
@@ -267,6 +301,64 @@ mod find_and_replace {
             "https://test.test/?foo=clink"
         );
     }
+
+    #[test]
+    fn youtube_sanitize() {
+        let clink = Clink::new(ClinkConfig::default());
+
+        assert_eq!(
+            clink.find_and_replace("https://youtu.be/dQw4w9WgXcQ?si=NblIBgit-qHN7MoH",),
+            "https://youtu.be/dQw4w9WgXcQ"
+        );
+
+        assert_eq!(
+            clink.find_and_replace("https://www.youtu.be/dQw4w9WgXcQ?si=NblIBgit-qHN7MoH",),
+            "https://www.youtu.be/dQw4w9WgXcQ"
+        );
+
+        assert_eq!(
+            clink.find_and_replace("https://youtu.be/dQw4w9WgXcQ?si=NblIBgit-qHN7MoH&t=69",),
+            "https://youtu.be/dQw4w9WgXcQ?t=69"
+        );
+
+        assert_eq!(
+            clink.find_and_replace(
+                "https://youtu.be/dQw4w9WgXcQ?si=NblIBgit-qHN7MoH&t=69&fbclid=clid",
+            ),
+            "https://youtu.be/dQw4w9WgXcQ?t=69"
+        );
+
+        assert_eq!(
+            clink.find_and_replace("https://test.test/dQw4w9WgXcQ?si=NblIBgit-qHN7MoH&t=69",),
+            "https://test.test/dQw4w9WgXcQ?si=NblIBgit-qHN7MoH&t=69"
+        );
+
+        let clink = Clink::new(ClinkConfig::new(Mode::Replace));
+        assert_eq!(
+            clink.find_and_replace(
+                "https://test.test/?fbclid=dsadsa&utm_source=fafa&utm_campaign=fafas&utm_medium=adsa&si=qweasd",
+            ),
+            "https://test.test/?fbclid=clink&utm_source=clink&utm_campaign=clink&utm_medium=clink&si=qweasd"
+        );
+
+        assert_eq!(
+            clink.find_and_replace(
+                "https://youtu.be/?fbclid=dsadsa&utm_source=fafa&utm_campaign=fafas&utm_medium=adsa&si=qweasd",
+            ),
+            "https://youtu.be/?fbclid=clink&utm_source=clink&utm_campaign=clink&utm_medium=clink&si=clink"
+        );
+
+        let clink = Clink::new(ClinkConfig::new(Mode::YourMom));
+        assert_eq!(
+            clink.find_and_replace("https://test.test/?si=dsadsa",),
+            "https://test.test/?si=dsadsa&utm_source=your_mom"
+        );
+
+        assert_eq!(
+            clink.find_and_replace("https://youtu.be/?si=dsadsa",),
+            "https://youtu.be/?utm_source=your_mom"
+        );
+    }
 }
 
 #[cfg(test)]

diff --git a/src/config.rs b/src/config.rs
@@ -9,6 +9,9 @@ use serde::{Deserialize, Serialize};
 
 use crate::mode::Mode;
 
+/// add query param, that must be replaced within any domain
+/// to specify domain specific params use format
+/// "{domain}``{param}"
 fn get_default_params() -> HashSet<Rc<str>> {
     HashSet::from([
         "fbclid".into(),
@@ -21,6 +24,8 @@ fn get_default_params() -> HashSet<Rc<str>> {
         "utm_medium".into(),
         "utm_term".into(),
         "utm_content".into(),
+        "youtube.com``si".into(),
+        "youtu.be``si".into(),
     ])
 }