Add file-type and host filters with a 4chan default

CURRENT.md

@@ -40,6 +40,13 @@ mobile draft at 390px.
   - [x] C11 preview caching: Discord-hosted image/GIF previews (WebP 320px, hash-addressed, 5 MiB
     cap, one attempt) run after every walk in both modes. Live run `...-20260916090722`:
     `previewed: 2`; thumbs served over `/api/media/<scrape>/<item>/thumb` (200).
+  - [x] C39 File-type filter and the 4chan default. The Advanced drawer gained extension chips
+    (video mp4/webm/mov/mkv, image png/jpg/gif/webp/avif, audio mp3/ogg/wav/m4a/flac) matched on
+    URL or content type, plus a default-on **Hide 4chan links (usually dead)** toggle. Backend,
+    API (`ext`, `excludeHosts`) and MCP support; live: video count 7696 → 7383 with the default
+    exclusion, `ext=mp4`/`webm` return only those files. Also confirmed the mr-game-memorial group
+    has 924 attachment videos with 772 previews cached and 0 preview errors — the rest are just
+    pending in the 2,140-item backfill queue.
   - [x] C38 Filter layout swap. The common filters (filetype chips, source kind, server/DM
     origin, multi-select People popover) now sit in the always-visible bar; the Advanced search
     drawer holds aspect, resolution, channel, date range, cached-only, and the card-size slider.

docs/media-mcp.md

@@ -24,7 +24,8 @@ interface MediaService {
 ```
 
 `MediaQuery` (implemented): `kind`, `source`, `scrapeId`, `channelId`, `authorIds`, `origin`
-(`guild`/`dm`), `cachedOnly`, `aspect`, `resolution`, `after`, `before`, `limit`, `offset`. The
+(`guild`/`dm`), `cachedOnly`, `exts`, `excludeHosts`, `aspect`, `resolution`, `after`, `before`,
+`limit`, `offset`. The
 `aspect` and `resolution` values are the preset keys from `filter-presets.ts`; `after`/`before` are
 ISO times on `posted_at`. Planned: `guildId`, `text`.
 

docs/media.md

@@ -147,8 +147,8 @@ that deletion is irreversible. The delete button is disabled for a 5 second coun
 
 - **Media** tab: the bar carries the common filters — kind chips, source kind, server/DM origin,
   and the multi-select People popover. The **Advanced search** drawer has aspect ratio, resolution,
-  source channel, posted-date range, and cached-only, with **Apply filters** / **Reset** in the
+  source channel, posted-date range, file-type chips, cached-only, and a default-on hide for 4chan
-  bar.
+  links (they 404 almost always), with **Apply filters** / **Reset** in the bar.
   Nothing renders until **Load media** is pressed. A **Metadata** dropdown toggles which fields
   show in the small banner under each tile (User, Type, Date, Source, Resolution, Length). Video
   tiles carry a translucent play icon top-right. Discord serves no thumbnails for uploaded videos,

docs/webui.md

@@ -129,7 +129,9 @@ box live in Chat, not Call, and the connection stats are not chips yet.
 - Media lists scraped items. The bar shows the common filters: filetype chips, source kind
   (attachments/stickers/embeds), origin (server channels or DMs), and a multi-select **People**
   popover. The **Advanced search** drawer holds aspect ratio, resolution, source channel,
-  posted-date range, and **Show only cached previews**, plus the card-size slider. Filters do not
+  posted-date range, file-type chips (video mp4/webm/mov/mkv, image png/jpg/gif/webp/avif, audio),
+  **Show only cached previews**, and a default-on **Hide 4chan links (usually dead)**, plus the
+  card-size slider. Filters do not
   apply until **Apply filters** is pressed; **Reset** clears every filter and the zoom, unloads the
   grid, and brings back the initial Load media prompt.
 - A **Card size** slider scales tiles from two-thirds of the old size up to one image per row; the
@@ -265,7 +267,7 @@ replaceable.
 | POST | `/api/audio/stop` | none | Stop speech and clear the queue |
 | POST | `/api/audio/capture` | `start` | Start or stop capture |
 | GET | `/api/transcript` | `limit?`, `sessionId?` | Sessions and captured segments |
-| GET | `/api/media` | `kind?`, `source?`, `channelId?`, `authorId?` (comma list), `origin?` (`guild`/`dm`), `cachedOnly?`, `aspect?`, `resolution?`, `after?`, `before?`, `limit?`, `offset?` | Media items across every scrape, newest first. `source` is `attachment`, `sticker`, or `embed`; `aspect` and `resolution` are the preset keys; `after`/`before` are ISO times on `posted_at`. Items carry `authorName`, `postedAt`, `width`, `height`, `aspectRatio`, `contentType`, `durationSecs`, `previewError`, `animUrl`; `localUrl`/`thumbUrl` when a file or preview exists |
+| GET | `/api/media` | `kind?`, `source?`, `channelId?`, `authorId?` (comma list), `origin?` (`guild`/`dm`), `cachedOnly?`, `ext?` (comma list), `excludeHosts?` (comma list), `aspect?`, `resolution?`, `after?`, `before?`, `limit?`, `offset?` | Media items across every scrape, newest first. `source` is `attachment`, `sticker`, or `embed`; `aspect` and `resolution` are the preset keys; `after`/`before` are ISO times on `posted_at`. Items carry `authorName`, `postedAt`, `width`, `height`, `aspectRatio`, `contentType`, `durationSecs`, `previewError`, `animUrl`; `localUrl`/`thumbUrl` when a file or preview exists |
 | GET | `/api/media/:scrapeId/:itemId/file` | none | The downloaded file, content-typed from its extension |
 | GET | `/api/media/:scrapeId/:itemId/thumb` | none | The cached preview (type from `preview_ext`) |
 | GET | `/api/media/:scrapeId/:itemId/stream` | none | Playback/full-view source: the local file, or the remote URL refreshed from the source message. Supports `Range`. Non-media upstream responses (provider pages) return `415` |

media-scrape/src/store.ts

@@ -108,6 +108,8 @@ export interface MediaFilterOptions {
     authorId?: string;
     authorIds?: string[];
     cachedOnly?: boolean;
+    exts?: string[];
+    excludeHosts?: string[];
     aspect?: string;
     resolution?: string;
     after?: string;
@@ -493,6 +495,17 @@ export class ScrapeStore {
         if (options.cachedOnly) {
             conditions.push("preview_hash IS NOT NULL");
         }
+        if (options.exts?.length) {
+            const clauses = options.exts.map(() => "(lower(url) LIKE ? OR lower(content_type) LIKE ?)");
+            conditions.push(`(${clauses.join(" OR ")})`);
+            for (const ext of options.exts) {
+                values.push(`%.${ext.toLowerCase()}%`, `%/${ext.toLowerCase()}%`);
+            }
+        }
+        for (const host of options.excludeHosts ?? []) {
+            conditions.push("lower(url) NOT LIKE ?");
+            values.push(`%${host.toLowerCase()}%`);
+        }
         if (options.after) {
             conditions.push("posted_at >= ?");
             values.push(options.after);

poc/public/media.js

@@ -195,11 +195,15 @@
             : "";
         const origin = state.filters.origin ? `origin=${state.filters.origin}&` : "";
         const cached = state.filters.cachedOnly ? "cachedOnly=1&" : "";
+        const exts = state.filters.exts.length ? `ext=${state.filters.exts.join(",")}&` : "";
+        const excludeHosts = state.filters.excludeHosts.length
+            ? `excludeHosts=${state.filters.excludeHosts.join(",")}&`
+            : "";
         const aspect = state.filters.aspect ? `aspect=${state.filters.aspect}&` : "";
         const resolution = state.filters.resolution ? `resolution=${state.filters.resolution}&` : "";
         const after = state.filters.after ? `after=${state.filters.after}T00:00:00.000Z&` : "";
         const before = state.filters.before ? `before=${state.filters.before}T23:59:59.999Z&` : "";
-        return `${kind}${source}${channel}${authors}${origin}${cached}${aspect}${resolution}${after}${before}`;
+        return `${kind}${source}${channel}${authors}${origin}${cached}${exts}${excludeHosts}${aspect}${resolution}${after}${before}`;
     }
 
     function kindTotals(field) {
@@ -629,6 +633,8 @@
             authorIds: [],
             origin: "",
             cachedOnly: false,
+            exts: [],
+            excludeHosts: ["4chan.org", "i.4cdn.org"],
             aspect: "",
             resolution: "",
             after: "",
@@ -650,6 +656,8 @@
             authorIds: [...state.draft.authorIds],
             origin: state.draft.origin,
             cachedOnly: state.draft.cachedOnly,
+            exts: [...state.draft.exts],
+            excludeHosts: [...state.draft.excludeHosts],
             aspect: state.draft.aspect,
             resolution: state.draft.resolution,
             after: state.draft.after,
@@ -846,6 +854,34 @@
             })
         );
 
+        const typeRow = (label, extensions) => {
+            const row = el("div", "vk-media-toolbar");
+            row.append(el("span", "vk-media-status", label));
+            for (const extension of extensions) {
+                const chip = el(
+                    "button",
+                    `vk-chip ${draft.exts.includes(extension) ? "vk-chip-on" : "vk-chip-off"}`,
+                    extension
+                );
+                chip.type = "button";
+                chip.addEventListener("click", () => {
+                    draft.exts = draft.exts.includes(extension)
+                        ? draft.exts.filter(value => value !== extension)
+                        : [...draft.exts, extension];
+                    renderFilterDrawer();
+                });
+                row.append(chip);
+            }
+            return row;
+        };
+
+        filters.append(
+            el("div", "vk-media-status", "File types"),
+            typeRow("Video", ["mp4", "webm", "mov", "mkv"]),
+            typeRow("Image", ["png", "jpg", "gif", "webp", "avif"]),
+            typeRow("Audio", ["mp3", "ogg", "wav", "m4a", "flac"])
+        );
+
         const cachedRow = el("label", "vk-media-check");
         const cachedBox = document.createElement("input");
         cachedBox.type = "checkbox";
@@ -855,6 +891,16 @@
         });
         cachedRow.append(cachedBox, document.createTextNode(" Show only cached previews"));
         filters.append(cachedRow);
+
+        const deadRow = el("label", "vk-media-check");
+        const deadBox = document.createElement("input");
+        deadBox.type = "checkbox";
+        deadBox.checked = draft.excludeHosts.includes("4chan.org");
+        deadBox.addEventListener("change", () => {
+            draft.excludeHosts = deadBox.checked ? ["4chan.org", "i.4cdn.org"] : [];
+        });
+        deadRow.append(deadBox, document.createTextNode(" Hide 4chan links (usually dead)"));
+        filters.append(deadRow);
         filterDrawer.append(filters);
 
         // Zoom applies live; it is not part of the filter draft.

poc/src/control-bridge.ts

@@ -883,6 +883,8 @@ function buildServer(): McpServer {
                 authorIds: z.array(z.string()).optional().describe("Restrict to these posters"),
                 origin: z.enum(["dm", "guild"]).optional().describe("Server channels or DMs"),
                 cachedOnly: z.boolean().optional().describe("Only items with a cached preview"),
+                exts: z.array(z.string()).optional().describe("Only these file extensions (mp4, webm, png, ...)"),
+                excludeHosts: z.array(z.string()).optional().describe("Hide URLs from these hosts"),
                 aspect: z
                     .enum(["vertical", "square", "wide43", "wide169", "ultrawide", "other"])
                     .optional()
@@ -905,6 +907,8 @@ function buildServer(): McpServer {
             authorIds,
             origin,
             cachedOnly,
+            exts,
+            excludeHosts,
             aspect,
             resolution,
             after,
@@ -921,6 +925,8 @@ function buildServer(): McpServer {
                     authorIds,
                     origin,
                     cachedOnly,
+                    exts,
+                    excludeHosts,
                     aspect,
                     resolution,
                     after,

poc/src/media.ts

@@ -638,6 +638,8 @@ export class MediaService {
         authorIds?: string[];
         origin?: "dm" | "guild";
         cachedOnly?: boolean;
+        exts?: string[];
+        excludeHosts?: string[];
         aspect?: string;
         resolution?: string;
         after?: string;
@@ -757,6 +759,8 @@ export class MediaService {
         authorIds?: string[];
         origin?: "dm" | "guild";
         cachedOnly?: boolean;
+        exts?: string[];
+        excludeHosts?: string[];
         aspect?: string;
         resolution?: string;
         after?: string;
@@ -782,6 +786,8 @@ export class MediaService {
                         authorId: options.authorId,
                         authorIds: options.authorIds,
                         cachedOnly: options.cachedOnly,
+                        exts: options.exts,
+                        excludeHosts: options.excludeHosts,
                         aspect: options.aspect,
                         resolution: options.resolution,
                         after: options.after,
@@ -833,6 +839,8 @@ export class MediaService {
             authorIds?: string[];
             origin?: "dm" | "guild";
             cachedOnly?: boolean;
+            exts?: string[];
+            excludeHosts?: string[];
             aspect?: string;
             resolution?: string;
             after?: string;

poc/src/webui.ts

@@ -893,6 +893,18 @@ export function startWebUi(deps: WebUiDeps): WebUiHandle {
                     ? (String(req.query.origin) as "dm" | "guild")
                     : undefined;
             const cachedOnly = req.query.cachedOnly === "1" || req.query.cachedOnly === "true";
+            const exts = req.query.ext
+                ? String(req.query.ext)
+                      .split(",")
+                      .map(value => value.trim().toLowerCase())
+                      .filter(Boolean)
+                : undefined;
+            const excludeHosts = req.query.excludeHosts
+                ? String(req.query.excludeHosts)
+                      .split(",")
+                      .map(value => value.trim().toLowerCase())
+                      .filter(Boolean)
+                : undefined;
             const aspect = req.query.aspect ? String(req.query.aspect) : undefined;
             const resolution = req.query.resolution ? String(req.query.resolution) : undefined;
             const after = req.query.after ? String(req.query.after) : undefined;
@@ -909,6 +921,8 @@ export function startWebUi(deps: WebUiDeps): WebUiHandle {
                     authorIds,
                     origin,
                     cachedOnly,
+                    exts,
+                    excludeHosts,
                     aspect,
                     resolution,
                     after,
@@ -944,6 +958,18 @@ export function startWebUi(deps: WebUiDeps): WebUiHandle {
                 req.query.origin === "dm" || req.query.origin === "guild"
                     ? (String(req.query.origin) as "dm" | "guild")
                     : undefined;
+            const exts = req.query.ext
+                ? String(req.query.ext)
+                      .split(",")
+                      .map(value => value.trim().toLowerCase())
+                      .filter(Boolean)
+                : undefined;
+            const excludeHosts = req.query.excludeHosts
+                ? String(req.query.excludeHosts)
+                      .split(",")
+                      .map(value => value.trim().toLowerCase())
+                      .filter(Boolean)
+                : undefined;
 
             res.json({
                 ok: true,
@@ -954,6 +980,8 @@ export function startWebUi(deps: WebUiDeps): WebUiHandle {
                     authorIds,
                     origin,
                     cachedOnly: req.query.cachedOnly === "1" || req.query.cachedOnly === "true",
+                    exts,
+                    excludeHosts,
                     aspect: req.query.aspect ? String(req.query.aspect) : undefined,
                     resolution: req.query.resolution ? String(req.query.resolution) : undefined,
                     after: req.query.after ? String(req.query.after) : undefined,