coto: decodear html entities

scraper-rs/Cargo.lock

@@ -604,6 +604,15 @@ version = "0.3.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "d77f7ec81a6d05a3abb01ab6eb7590f6083d08449fe5a1c8b1e620283546ccb7"
 
+[[package]]
+name = "html-escape"
+version = "0.2.13"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "6d1ad449764d627e22bfd7cd5e8868264fc9236e07c752972b4080cd351cb476"
+dependencies = [
+ "utf8-width",
+]
+
 [[package]]
 name = "http"
 version = "0.2.11"
@@ -1229,6 +1238,7 @@ dependencies = [
  "deadpool",
  "deadpool-sqlite",
  "futures",
+ "html-escape",
  "itertools",
  "nanoid",
  "quick-xml",
@@ -1614,6 +1624,12 @@ dependencies = [
  "percent-encoding",
 ]
 
+[[package]]
+name = "utf8-width"
+version = "0.1.7"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "86bd8d4e895da8537e5315b8254664e6b769c4ff3db18321b297a1e7004392e3"
+
 [[package]]
 name = "utf8parse"
 version = "0.2.1"

scraper-rs/Cargo.toml

@@ -14,6 +14,7 @@ cron = "0.12.0"
 deadpool = "0.10.0"
 deadpool-sqlite = "0.7.0"
 futures = "0.3.30"
+html-escape = "0.2.13"
 itertools = "0.12.0"
 nanoid = "0.4.0"
 quick-xml = "0.31.0"

scraper-rs/src/sites/coto.rs

@@ -53,7 +53,8 @@ pub fn parse(url: String, dom: &tl::VDom) -> Result<PrecioPoint, anyhow::Error>
         .filter_map(|h| h.get(dom.parser()))
         .find_map(|n| n.as_tag())
         .map(|t| t.inner_text(dom.parser()))
-        .map(|s| s.trim().to_string());
+        // https://github.com/catdevnull/preciazo/issues/24
+        .map(|s| html_escape::decode_html_entities(s.trim()).to_string());
 
     let image_url = dom
         .query_selector(".zoomImage1")