Directory: chips by kind and category over a grid of cover art

Feeds take their channel's first <itunes:category> into a new feeds.category
column; the migration drops ETag and Last-Modified once so every feed re-reads
on its normal schedule and picks one up. /api/popular and /api/directory carry
category and podcast (any audio or video enclosure). Directory becomes a grid of
cover-art tiles under a chip rail: All, Podcasts, Blogs, and a podcast's
categories once Podcasts is picked. Popular and Add a feed keep their rows.

Closes #4, closes #5, closes #6, closes #7.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

src/db.rs

@@ -17,6 +17,8 @@ CREATE TABLE IF NOT EXISTS feeds (
     url           TEXT NOT NULL,
     title         TEXT,
     image         TEXT,
+    -- The channel's first <itunes:category>, for the Directory.
+    category      TEXT,
     etag          TEXT,
     last_modified TEXT,
     last_checked  INTEGER,
@@ -170,6 +172,7 @@ fn migrate(conn: &Connection) -> Result<()> {
         ("users", "created", "INTEGER"),
         ("users", "last_login", "INTEGER"),
         ("feeds", "error_since", "INTEGER"),
+        ("feeds", "category", "TEXT"),
     ];
     let retired: &[(&str, &str)] = &[
         // Read state from before accounts, long since moved to entry_state. Two bugs came from
@@ -192,6 +195,13 @@ fn migrate(conn: &Connection) -> Result<()> {
         if !has(table, column)? {
             tracing::info!(table, column, "adding column");
             conn.execute_batch(&format!("ALTER TABLE {table} ADD COLUMN {column} {ty}"))?;
+            // A category is only read from a 200, and a feed with a validator mostly gets a
+            // 304, so most would never pick one up until the publisher changed something.
+            // Dropping the validators once makes each re-read on its normal schedule; leaving
+            // last_checked alone, unlike clear_validators, keeps them from all coming due at once.
+            if (*table, *column) == ("feeds", "category") {
+                conn.execute_batch("UPDATE feeds SET etag = NULL, last_modified = NULL")?;
+            }
         }
     }
     for (table, column) in retired {
@@ -208,6 +218,7 @@ fn migrate(conn: &Connection) -> Result<()> {
 pub struct FeedSummary {
     pub title: Option<String>,
     pub image: Option<String>,
+    pub category: Option<String>,
     /// Came from a subscribed OPML that no longer lists it, but it has downloads, so it
     /// was kept rather than removed.
     pub orphaned: bool,
@@ -256,7 +267,8 @@ impl Db {
         let conn = self.conn.lock().unwrap();
         let mut sum: FeedSummary = conn
             .query_row(
-                "SELECT title, image, last_checked, last_error, coalesce(orphaned, 0), error_since
+                "SELECT title, image, last_checked, last_error, coalesce(orphaned, 0), error_since,
+                        category
                  FROM feeds WHERE id = ?1",
                 [feed_id],
                 |r| {
@@ -267,6 +279,7 @@ impl Db {
                         last_error: r.get(3)?,
                         orphaned: r.get::<_, i64>(4)? != 0,
                         error_since: r.get(5)?,
+                        category: r.get(6)?,
                         ..Default::default()
                     })
                 },
@@ -328,11 +341,14 @@ impl Db {
         last_modified: Option<&str>,
         ttl_mins: Option<u64>,
         image: Option<&str>,
+        category: Option<&str>,
     ) -> Result<()> {
         let conn = self.conn.lock().unwrap();
+        // category is taken as it comes, unlike title and image: a show that leaves a category
+        // should leave the Directory's chip too.
         conn.execute(
-            "INSERT INTO feeds (id, url, title, etag, last_modified, last_checked, ttl_mins, last_error, image)
-             VALUES (?1, ?2, ?3, ?4, ?5, ?6, ?7, NULL, ?8)
+            "INSERT INTO feeds (id, url, title, etag, last_modified, last_checked, ttl_mins, last_error, image, category)
+             VALUES (?1, ?2, ?3, ?4, ?5, ?6, ?7, NULL, ?8, ?9)
              ON CONFLICT(id) DO UPDATE SET
                 url = excluded.url,
                 title = coalesce(excluded.title, feeds.title),
@@ -341,9 +357,10 @@ impl Db {
                 last_checked = excluded.last_checked,
                 ttl_mins = excluded.ttl_mins,
                 image = coalesce(excluded.image, feeds.image),
+                category = excluded.category,
                 last_error = NULL,
                 error_since = NULL",
-            rusqlite::params![feed_id, url, title, etag, last_modified, now(), ttl_mins.map(|t| t as i64), image],
+            rusqlite::params![feed_id, url, title, etag, last_modified, now(), ttl_mins.map(|t| t as i64), image, category],
         )?;
         Ok(())
     }
@@ -972,6 +989,17 @@ impl Db {
         Ok(out)
     }
 
+    /// Feeds with any audio or video enclosure: the Directory's Podcasts, with the rest Blogs.
+    /// Reaped files keep their rows, so a show whose files have all been purged still counts.
+    pub fn media_feeds(&self) -> Result<std::collections::HashSet<String>> {
+        let conn = self.conn.lock().unwrap();
+        let mut stmt = conn.prepare(
+            "SELECT DISTINCT feed_id FROM enclosures WHERE mime LIKE 'audio/%' OR mime LIKE 'video/%'",
+        )?;
+        let out = stmt.query_map([], |r| r.get(0))?.collect::<rusqlite::Result<_>>()?;
+        Ok(out)
+    }
+
     /// Who else would miss this file: subscribers other than `user_id` who have starred
     /// the item or have not read it yet. Deleting is deleting their copy too.
     pub fn others_wanting(&self, enclosure_id: i64, user_id: i64) -> Result<(i64, i64)> {
@@ -1441,6 +1469,31 @@ pub fn now() -> i64 {
 mod tests {
     use super::*;
 
+    #[test]
+    fn adding_category_drops_validators_once_and_keeps_the_schedule() {
+        let conn = Connection::open_in_memory().unwrap();
+        conn.execute_batch(SCHEMA).unwrap();
+        // A database from before the column, holding a feed that would answer 304.
+        conn.execute_batch(
+            "ALTER TABLE feeds DROP COLUMN category;
+             INSERT INTO feeds (id, url, etag, last_modified, last_checked) VALUES ('f','u','e','lm',5);",
+        )
+        .unwrap();
+        let row = |conn: &Connection| -> (Option<String>, Option<String>, Option<i64>) {
+            conn.query_row("SELECT etag, last_modified, last_checked FROM feeds", [], |r| {
+                Ok((r.get(0)?, r.get(1)?, r.get(2)?))
+            })
+            .unwrap()
+        };
+        migrate(&conn).unwrap();
+        assert_eq!(row(&conn), (None, None, Some(5)), "re-read on its normal schedule");
+
+        // Only the once: every later open keeps the validators the next poll stored.
+        conn.execute_batch("UPDATE feeds SET etag = 'e2'").unwrap();
+        migrate(&conn).unwrap();
+        assert_eq!(row(&conn).0.as_deref(), Some("e2"));
+    }
+
     #[test]
     fn every_sort_column_runs_and_orders_both_ways() {
         let db = Db::memory().unwrap();

src/feed.rs

@@ -11,6 +11,8 @@ pub struct ParsedFeed {
     pub title: Option<String>,
     pub ttl_mins: Option<u64>,
     pub image: Option<String>,
+    /// The channel's first `<itunes:category>`, for the Directory's chips.
+    pub category: Option<String>,
     pub entries: Vec<Entry>,
 }
 
@@ -535,6 +537,12 @@ fn from_rss(ch: rss::Channel, bytes: &[u8]) -> ParsedFeed {
             .and_then(|i| i.image())
             .map(str::to_owned)
             .or_else(|| ch.image().map(|i| i.url().to_owned())),
+        // Only the iTunes one: Apple's list is fixed, while a plain <category> is freeform and
+        // would fill the Directory with one-off tags. The top level only, for the same reason.
+        category: ch
+            .itunes_ext()
+            .and_then(|i| i.categories().first())
+            .and_then(|c| non_empty(Some(c.text().trim()))),
         entries,
     }
 }
@@ -591,6 +599,7 @@ fn from_atom(feed: atom_syndication::Feed) -> ParsedFeed {
         title: non_empty(Some(feed.title().as_str())),
         ttl_mins: None,
         image: feed.logo().or_else(|| feed.icon()).map(str::to_owned),
+        category: None,
         entries,
     }
 }
@@ -708,6 +717,7 @@ mod tests {
 
         assert_eq!(feed.title.as_deref(), Some("Test Cast"));
         assert_eq!(feed.ttl_mins, Some(45));
+        assert_eq!(feed.category.as_deref(), Some("Technology"), "the first, top level only");
         assert_eq!(feed.entries.len(), 3);
 
         let ep = &feed.entries[0];

src/main.rs

@@ -1084,6 +1084,7 @@ async fn scan_one(
         last_modified.as_deref(),
         parsed.ttl_mins,
         parsed.image.as_deref(),
+        parsed.category.as_deref(),
     )?;
 
     let policy = policy_for(ctx, id, feed_cfg)?;

src/web.rs

@@ -627,6 +627,11 @@ struct PopularRow {
     subscribers: i64,
     /// Yours already. Everyone counts, you included, so your own feeds are listed too.
     subscribed: bool,
+    /// The feed's own iTunes category, if it names one; most blogs do not.
+    category: Option<String>,
+    /// Any audio or video enclosure. Unlike category, every feed has an answer, so the
+    /// Directory's Podcasts and Blogs between them hold everything.
+    podcast: bool,
 }
 
 /// Every feed that may be listed, with everyone counted, you included, most subscribers
@@ -638,6 +643,7 @@ fn popular(state: &WebState, user_id: i64) -> Result<Vec<PopularRow>> {
     let mine: std::collections::HashSet<String> =
         db.subscriptions_for(user_id)?.into_iter().map(|s| s.feed_id).collect();
     let counts = db.subscriber_counts()?;
+    let media = db.media_feeds()?;
     let catalogue = crate::subscriptions(&state.ctx)?;
     let by_id: std::collections::HashMap<&str, &crate::config::Feed> =
         catalogue.iter().map(|s| (s.id.as_str(), &s.cfg)).collect();
@@ -657,7 +663,15 @@ fn popular(state: &WebState, user_id: i64) -> Result<Vec<PopularRow>> {
         }
         let sum = db.feed_summary(&s.id)?;
         let subscribed = mine.contains(&s.id);
-        out.push(PopularRow { id: s.id.clone(), title: sum.title, image: sum.image, subscribers: n, subscribed });
+        out.push(PopularRow {
+            id: s.id.clone(),
+            title: sum.title,
+            image: sum.image,
+            subscribers: n,
+            subscribed,
+            category: sum.category,
+            podcast: media.contains(&s.id),
+        });
     }
     out.sort_by(|a, b| b.subscribers.cmp(&a.subscribers).then_with(|| sort_name(a).cmp(&sort_name(b))));
     Ok(out)