From 8416fe9acfe2cab2c4813f4a1a87bda556698b7d Mon Sep 17 00:00:00 2001
From: Finn Bear <finnbearlabs@gmail.com>
Date: Sun, 22 Sep 2024 13:56:35 -0700
Subject: [PATCH] 0.7.28 - reduce domain name false positives (pii).

---
 Cargo.toml     |  2 +-
 src/context.rs | 12 +++++++++---
 src/pii.rs     | 10 +++++++---
 3 files changed, 17 insertions(+), 7 deletions(-)

diff --git a/Cargo.toml b/Cargo.toml
index 76a672b..acadb23 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -1,7 +1,7 @@
 [package]
 name = "rustrict"
 authors = ["Finn Bear"]
-version = "0.7.27"
+version = "0.7.28"
 edition = "2021"
 license = "MIT OR Apache-2.0"
 repository = "https://github.com/finnbear/rustrict/"
diff --git a/src/context.rs b/src/context.rs
index ceb75af..055b14c 100644
--- a/src/context.rs
+++ b/src/context.rs
@@ -8,9 +8,9 @@ use std::time::{Duration, Instant};
 
 /// Context is useful for taking moderation actions on a per-user basis i.e. each user would get
 /// their own Context.
-/// 
+///
 /// # Recommendation
-/// 
+///
 /// Use this as a reference implementation e.g. by copying and adapting it.
 #[derive(Clone)]
 #[cfg_attr(feature = "serde", derive(serde::Serialize, serde::Deserialize))]
@@ -868,6 +868,12 @@ mod tests {
         let json = serde_json::to_value(&ctx).unwrap();
         let only_safe_until = &json["only_safe_until"];
         let unix = only_safe_until.as_i64().unwrap();
-        assert!(unix > 1000 + SystemTime::now().duration_since(SystemTime::UNIX_EPOCH).unwrap().as_millis() as i64)
+        assert!(
+            unix > 1000
+                + SystemTime::now()
+                    .duration_since(SystemTime::UNIX_EPOCH)
+                    .unwrap()
+                    .as_millis() as i64
+        )
     }
 }
diff --git a/src/pii.rs b/src/pii.rs
index ac68c80..5266661 100644
--- a/src/pii.rs
+++ b/src/pii.rs
@@ -5,10 +5,10 @@ use std::borrow::Cow;
 lazy_static! {
     static ref PHONE : Regex = Regex::new(r#"(\+\d{1,2})?\s*\(?\d{3}\)?[\s\.-]*\d{3}[\s\.-]*\d{4}"#).unwrap();
     static ref IP_ADDRESS : Regex  = Regex::new(r#"(?:[0-9]{1,3}\.){3}[0-9]{1,3}"#).unwrap();
-    static ref EMAIL_ADDRESS : Regex = Regex::new(r#"(?i)[a-z0-9_\-]{3,}\s*(@|[\[\(\s]at[\s\)\]])\s*[a-z0-9_\-]{5,}\s*(\.|dot)\s*[a-z]{2,3}"#).unwrap();
+    static ref EMAIL_ADDRESS : Regex = Regex::new(r#"(?i)[a-z0-9_\-]{3,}\s*(@|[\[\(\s]at[\s\)\]])\s*[a-z0-9_\-]{5,}\s*(\.|dot)\s*(com|net|org|gov|biz|co|us|ru|uk|de|se|to|tv|io|info|online|site)"#).unwrap();
     //static ref ADDRESS : Regex = Regex::new(r#"(?i)\d+[ ](?:[A-Za-z0-9\.-]+ )+(?:Avenue|Lane|Road|Boulevard|Drive|Street|Ave|Dr|Rd|Blvd|Ln|St)\.?(\s+#[0-9]{1,5})?"#).unwrap();
     static ref NAME : Regex = Regex::new(r#"(?i)(real\s)?name\s+is:?\s[a-zA-Z]+(\s[a-zA-z]+)?"#).unwrap();
-    static ref URL : Regex = Regex::new(r#"(?i)(https?:?/*)?[a-zA-Z0-9]{4,}\.[a-zA-Z]{2,3}"#).unwrap();
+    static ref URL : Regex = Regex::new(r#"(?i)(https?:?/*)?[a-zA-Z0-9]{4,}\.(com|net|org|gov|biz|co|us|ru|uk|de|se|to|tv|io|info|online|site)"#).unwrap();
 }
 
 /// Returns [`s`] with personally-identifiable information censored out, and a `true` if
@@ -103,7 +103,11 @@ mod tests {
         for line in include_str!("./safe.txt")
             .lines()
             .chain(include_str!("./false_positives.txt").lines())
-            .chain(r#"1234 Have 1234"#.lines())
+            .chain(
+                r#"1234 Have 1234
+            gmail.zzz"#
+                    .lines(),
+            )
         {
             assert!(!has_pii(line), "{line}");
         }