add more fields

crates/blend-db/migrations/20240422000000_feeds.sql

@@ -5,6 +5,7 @@ CREATE TABLE IF NOT EXISTS feeds (
   url_site TEXT,
   title TEXT,
   title_display TEXT,
+  favicon_url TEXT,
   favicon_b64 BLOB,
   published_at DATETIME,
   updated_at DATETIME,

crates/blend-db/migrations/20240430000000_entries.sql

@@ -4,12 +4,13 @@ CREATE TABLE IF NOT EXISTS entries (
   id TEXT NOT NULL,
   url TEXT,
   title TEXT,
-  summary TEXT,
+  summary_html TEXT,
   content_html TEXT,
   content_scraped_html TEXT,
   published_at DATETIME,
   updated_at DATETIME,
   read_at DATETIME,
+  saved_at DATETIME,
   UNIQUE(feed_uuid, id),
   CONSTRAINT fk_feed FOREIGN KEY (feed_uuid) REFERENCES feeds(uuid) ON DELETE CASCADE
 );

crates/blend-db/src/model/entry.rs

@@ -12,12 +12,13 @@ pub struct Entry {
     pub id: String,
     pub url: String,
     pub title: Option<String>,
-    pub summary: Option<String>,
+    pub summary_html: Option<String>,
     #[sqlx(default)]
     pub content_html: Option<String>,
     #[sqlx(default)]
     pub content_scraped_html: Option<String>,
     pub published_at: Option<DateTime<Utc>>,
     pub updated_at: Option<DateTime<Utc>>,
     pub read_at: Option<DateTime<Utc>>,
+    pub saved_at: Option<DateTime<Utc>>,
 }

crates/blend-db/src/model/feed.rs

@@ -13,7 +13,8 @@ pub struct Feed {
     pub url_site: Option<String>,
     pub title: Option<String>,
     pub title_display: Option<String>,
-    pub favicon_b64: Option<Vec<u8>>,
+    pub favicon_b64: Option<String>,
+    pub favicon_url: Option<String>,
     pub published_at: Option<DateTime<Utc>>,
     pub updated_at: Option<DateTime<Utc>>,
 }

crates/blend-db/src/repo/entry.rs

@@ -14,7 +14,7 @@ pub struct CreateEntryParams {
     pub id: String,
     pub url: Option<String>,
     pub title: Option<String>,
-    pub summary: Option<String>,
+    pub summary_html: Option<String>,
     pub content_html: Option<String>,
     pub published_at: Option<DateTime<Utc>>,
     pub updated_at: Option<DateTime<Utc>>,
@@ -86,7 +86,7 @@ impl EntryRepo {
         let el = filter.sort.query_elements();
         let el_inv = filter.sort.query_elements_inverse();
 
-        let mut query = QueryBuilder::<Sqlite>::new("SELECT uuid, feed_uuid, id, url, title, summary, published_at, updated_at, read_at FROM entries WHERE 1=1");
+        let mut query = QueryBuilder::<Sqlite>::new("SELECT uuid, feed_uuid, id, url, title, summary_html, published_at, updated_at, read_at FROM entries WHERE 1=1");
 
         match filter.view {
             View::All => query.push(""),
@@ -176,14 +176,14 @@ impl EntryRepo {
             return Ok(vec![]);
         }
 
-        let mut query = QueryBuilder::<Sqlite>::new("INSERT INTO entries (feed_uuid, uuid, id, url, title, summary, content_html, published_at, updated_at) ");
+        let mut query = QueryBuilder::<Sqlite>::new("INSERT INTO entries (feed_uuid, uuid, id, url, title, summary_html, content_html, published_at, updated_at) ");
         query.push_values(entries.iter(), |mut b, entry| {
             b.push_bind(feed_uuid)
                 .push_bind(uuid::Uuid::new_v4())
                 .push_bind(entry.id.clone())
                 .push_bind(entry.url.clone())
                 .push_bind(entry.title.clone())
-                .push_bind(entry.summary.clone())
+                .push_bind(entry.summary_html.clone())
                 .push_bind(entry.content_html.clone())
                 .push_bind(entry.published_at)
                 .push_bind(entry.updated_at);
@@ -194,7 +194,7 @@ impl EntryRepo {
             DO UPDATE SET
                 url = excluded.url,
                 title = excluded.title,
-                summary = excluded.summary,
+                summary_html = excluded.summary_html,
                 content_html = excluded.content_html,
                 updated_at = excluded.updated_at
             RETURNING uuid

crates/blend-feed/Cargo.toml

@@ -10,7 +10,6 @@ license.workspace = true
 ammonia = "4.0"
 chrono = { workspace = true, features = ["serde"] }
 feed-rs = { workspace = true }
-html5ever = "0.27"
 reqwest = { workspace = true }
 serde = { workspace = true }
 serde_json = { workspace = true }

crates/blend-feed/src/lib.rs

@@ -20,6 +20,7 @@ async fn get_feed(url: &str) -> FeedResult<feed_rs::model::Feed> {
 pub async fn parse_feed(url: &str) -> FeedResult<ParsedFeed> {
     let feed = get_feed(url).await?;
 
+    // Parse favicon URL to use until we can convert the remote image into binary data stored in the db
     let favicon_url = feed
         .icon
         .or_else(|| feed.logo)

crates/blend-feed/src/sanitize/html.rs

@@ -1,6 +1,25 @@
 use ammonia::Builder;
+use std::collections::HashSet;
 
-/// Extract safe and sanitized HTML elements.
+const REMOVE_TAGS: [&str; 1] = ["article"];
+
+// Sanitize HTML input, allowing only safe elements
 pub fn sanitize_html(src: &str) -> String {
-    Builder::default().clean(src).to_string()
+    Builder::default().rm_tags(HashSet::from(REMOVE_TAGS)).clean(src).to_string()
+}
+
+#[cfg(test)]
+mod test {
+    use super::*;
+
+    #[test]
+    fn it_keeps_only_safe_elements() {
+        let src = r#"<article><p>Some body text that we <em>want</em> to keep.</p><p class="read-more">[<a href="https://example.com">Read More</a>]</p><script>alert("gotcha")</script><style>body { display: none }</style></article>"#;
+
+        let parsed = sanitize_html(src);
+        assert_eq!(
+            parsed,
+            r#"<p>Some body text that we <em>want</em> to keep.</p><p>[<a href="https://example.com" rel="noopener noreferrer">Read More</a>]</p>"#
+        );
+    }
 }

crates/blend-feed/src/sanitize/mod.rs

@@ -6,3 +6,5 @@ pub use text::sanitize_text;
 
 mod html;
 pub use html::sanitize_html;
+
+// TODO: use `.url_relative(UrlRelative::RewriteWithBase(...))` with ammonia and pass in site URL to rewrite relative URLs

crates/blend-feed/src/sanitize/text.rs

@@ -1,6 +1,6 @@
 use ammonia::Builder;
 
-/// Sanitize HTML input, allowing only text.
+/// Sanitize HTML input, allowing only plain text
 pub fn sanitize_text(src: &str) -> String {
     Builder::empty().clean(src).to_string()
 }

crates/blend-worker/src/handler.rs

@@ -20,7 +20,7 @@ pub async fn fetch_entries(
             id: entry.id,
             url: entry.url,
             title: entry.title,
-            summary: entry.summary_html,
+            summary_html: entry.summary_html,
             content_html: entry.content_html,
             published_at: entry.published_at,
             updated_at: entry.updated_at,