From f56e0ec20799ddb7948dab988153d9c7796cc974 Mon Sep 17 00:00:00 2001
From: annieversary <annie@versary.town>
Date: Fri, 9 Sep 2022 15:05:35 +0200
Subject: [PATCH] add regex scraper

---
 Cargo.toml      |  4 +++-
 src/scraping.rs | 38 ++++++++++++++++++++++++++++++++++++--
 2 files changed, 39 insertions(+), 3 deletions(-)

diff --git a/Cargo.toml b/Cargo.toml
index 5dee8d3..453caa8 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -7,10 +7,12 @@ edition = "2021"
 default = []
 
 inventory = ["dep:inventory"]
-scraping = ["dep:scraper"]
+scraping = ["dep:scraper", "dep:lazy_static", "dep:regex"]
 
 [dependencies]
 inventory = { version = "0.3", optional = true }
+lazy_static = { version = "1.4.0", optional = true }
+regex = { version = "1.6.0", optional = true }
 scraper = { version = "0.13.0", optional = true }
 tracing = "0.1.35"
 
diff --git a/src/scraping.rs b/src/scraping.rs
index 1269770..13d356e 100644
--- a/src/scraping.rs
+++ b/src/scraping.rs
@@ -1,6 +1,10 @@
+use regex::Regex;
 use scraper::{ElementRef, Html};
 
-/// Gets all classes from an html
+/// Gets all classes from an html, using the scraper crate
+///
+/// If you have a templated html source and it's failing to get some classes, you might want
+/// to use `get_classes_regex` instead.
 ///
 /// ```
 /// # use zephyr::{*, scraping::*};
@@ -31,12 +35,32 @@ pub fn get_classes(html: &str) -> Vec<String> {
     classes
 }
 
+lazy_static::lazy_static! {
+    static ref STYLE_REGEX: Regex =
+        Regex::new(r#"(?:class|className)=(?:["']\W+\s*(?:\w+)\()?["']([^'"]+)['"]"#).unwrap();
+}
+
+/// Gets all classes from an html, using a regex
+///
+/// It's less accurate than `get_classes`, meaning it will find more false positives.
+/// Use this if you have a templated html source and don't mind generating more classes than there actually are.
+pub fn get_classes_regex(html: &str) -> Vec<&str> {
+    let mut classes = vec![];
+    for capture in STYLE_REGEX.captures_iter(html) {
+        if let Some(group) = capture.get(1) {
+            classes.push(group.as_str())
+        }
+    }
+
+    classes
+}
+
 #[cfg(test)]
 mod tests {
     use super::*;
 
     #[test]
-    fn test_parse() {
+    fn test_parse_scraper() {
         let c = get_classes(
             "<h1 class=\"hey hello\">Hello, <i class=\"hiii\">world!</i></h1>
 <h1 class=\"hey hello\">Hello, <i class=\"hiii\">world!</i></h1>",
@@ -44,4 +68,14 @@ mod tests {
 
         assert_eq!(c, vec!["hey hello", "hiii", "hey hello", "hiii",]);
     }
+
+    #[test]
+    fn test_parse_regex() {
+        let c = get_classes_regex(
+            "<h1 class=\"hey hello\">Hello, <i class=\"hiii\">world!</i></h1>
+<h1 class=\"hey hello\">Hello, <i class=\"hiii\">world!</i></h1>",
+        );
+
+        assert_eq!(c, vec!["hey hello", "hiii", "hey hello", "hiii",]);
+    }
 }