@dwk/microsub 0.1.0-beta.0

package/dist/hfeed.js

@@ -0,0 +1,252 @@
+/**
+ * `@dwk/microsub` — `h-feed` / `h-entry` microformats parsing.
+ *
+ * When a followed source is an HTML page rather than a syndication feed, its
+ * entries are expressed as [microformats2](https://microformats.org/wiki/h-entry)
+ * `h-entry` items (usually inside an `h-feed`). This module extracts them into
+ * the same JF2 shape the other formats produce, using the Workers runtime's
+ * streaming `HTMLRewriter` rather than a regex or a bundled parser
+ * (`HTMLRewriter` is built into the runtime — zero script-size cost; see
+ * `spec/non-functional-requirements.md`). Because `HTMLRewriter` is a `workerd`
+ * global, this parser is async and exercised under the Workers test pool.
+ *
+ * It is a pragmatic extractor of the common `h-entry` properties — `u-url`,
+ * `p-name`, `e-content`, `dt-published`, `p-author` / nested `h-card`,
+ * `u-photo`, `p-category`, `u-in-reply-to`, `u-like-of` — not a full mf2 engine.
+ *
+ * @packageDocumentation
+ */
+function classes(value) {
+    return value === null ? [] : value.trim().split(/\s+/).filter(Boolean);
+}
+/**
+ * HTML void elements have no end tag, so `HTMLRewriter#onEndTag` throws on them.
+ * They also carry no text content — a `u-photo` / `dt-published` on one always
+ * takes its value from an attribute — so we commit immediately and never
+ * register an end-tag handler for them.
+ */
+const VOID_ELEMENTS = new Set([
+    "area",
+    "base",
+    "br",
+    "col",
+    "embed",
+    "hr",
+    "img",
+    "input",
+    "link",
+    "meta",
+    "param",
+    "source",
+    "track",
+    "wbr",
+]);
+function absolute(href, base) {
+    try {
+        return new URL(href, base).toString();
+    }
+    catch {
+        return href;
+    }
+}
+/** FNV-1a hash (base36), matching {@link ./jf2}'s fallback id derivation. */
+function hash(input) {
+    let h = 0x811c9dc5;
+    for (let i = 0; i < input.length; i++) {
+        h ^= input.charCodeAt(i);
+        h = Math.imul(h, 0x01000193);
+    }
+    return (h >>> 0).toString(36);
+}
+/** The first `(p|u|e|dt)-name` microformats class on an element, if any. */
+function propertyClass(classList) {
+    for (const cls of classList) {
+        const match = /^(p|u|e|dt)-(.+)$/.exec(cls);
+        if (match) {
+            return { name: match[2], format: match[1] };
+        }
+    }
+    return null;
+}
+function commitEntry(entry, baseUrl) {
+    const url = entry.url ? absolute(entry.url, baseUrl) : undefined;
+    const content = entry.content
+        ? { text: entry.content }
+        : undefined;
+    const author = entry.author
+        ? (() => {
+            const card = {
+                type: "card",
+                ...(entry.author.name ? { name: entry.author.name } : {}),
+                ...(entry.author.url
+                    ? { url: absolute(entry.author.url, baseUrl) }
+                    : {}),
+                ...(entry.author.photo
+                    ? { photo: absolute(entry.author.photo, baseUrl) }
+                    : {}),
+            };
+            return card.name || card.url || card.photo ? card : undefined;
+        })()
+        : undefined;
+    const id = url ?? hash(`${entry.name ?? ""}${entry.content ?? ""}`);
+    const out = { type: "entry", _id: id };
+    if (url)
+        out.url = url;
+    if (entry.published)
+        out.published = entry.published;
+    if (entry.name)
+        out.name = entry.name;
+    if (content)
+        out.content = content;
+    if (author)
+        out.author = author;
+    if (entry.category.length > 0)
+        out.category = entry.category;
+    if (entry.photo.length > 0)
+        out.photo = entry.photo.map((p) => absolute(p, baseUrl));
+    if (entry.inReplyTo)
+        out["in-reply-to"] = absolute(entry.inReplyTo, baseUrl);
+    if (entry.likeOf)
+        out["like-of"] = absolute(entry.likeOf, baseUrl);
+    return out;
+}
+/**
+ * Extract JF2 entries from an HTML document's `h-entry` microformats. Entries
+ * are returned in document order. Returns `[]` when the document has none.
+ */
+export async function parseHFeed(html, baseUrl) {
+    if (html === "")
+        return [];
+    const entries = [];
+    const entryStack = [];
+    const cardStack = [];
+    const propStack = [];
+    const top = (arr) => arr[arr.length - 1];
+    const rewriter = new HTMLRewriter().on("*", {
+        element(el) {
+            const classList = classes(el.getAttribute("class"));
+            const isVoid = VOID_ELEMENTS.has(el.tagName);
+            if (classList.includes("h-entry") && !isVoid) {
+                const entry = { photo: [], category: [] };
+                entryStack.push(entry);
+                el.onEndTag(() => {
+                    const finished = entryStack.pop();
+                    if (finished)
+                        entries.push(commitEntry(finished, baseUrl));
+                });
+                return;
+            }
+            if (classList.includes("h-card") && !isVoid) {
+                const card = {};
+                const entry = top(entryStack);
+                // A bare p-author h-card attaches as the current entry's author.
+                if (entry && entry.author === undefined)
+                    entry.author = card;
+                cardStack.push(card);
+                el.onEndTag(() => {
+                    cardStack.pop();
+                });
+                return;
+            }
+            const prop = propertyClass(classList);
+            if (prop === null)
+                return;
+            const target = top(cardStack) ?? top(entryStack);
+            if (target === undefined)
+                return;
+            const frame = {
+                name: prop.name,
+                format: prop.format,
+                target,
+                buf: "",
+            };
+            // u-* and dt-* take their value from an attribute when present; that value
+            // is committed immediately, so no text capture (or end tag) is needed.
+            if (prop.format === "u") {
+                const attr = el.getAttribute("href") ??
+                    el.getAttribute("src") ??
+                    el.getAttribute("data") ??
+                    el.getAttribute("poster");
+                if (attr !== null) {
+                    assignProperty(frame, attr);
+                    return;
+                }
+            }
+            else if (prop.format === "dt") {
+                const attr = el.getAttribute("datetime") ?? el.getAttribute("value");
+                if (attr !== null) {
+                    assignProperty(frame, attr);
+                    return;
+                }
+            }
+            // Text-valued property: a void element has no text to capture.
+            if (isVoid)
+                return;
+            propStack.push(frame);
+            el.onEndTag(() => {
+                const finished = propStack.pop();
+                if (finished)
+                    assignProperty(finished, finished.buf.trim());
+            });
+        },
+        text(chunk) {
+            const frame = top(propStack);
+            if (frame)
+                frame.buf += chunk.text;
+        },
+    });
+    await rewriter.transform(new Response(html)).text();
+    return entries;
+}
+/** Write a resolved property value onto its target entry or card. */
+function assignProperty(frame, value) {
+    if (value === "")
+        return;
+    const target = frame.target;
+    // Card target (author).
+    if (isCard(target)) {
+        if (frame.name === "name" && target.name === undefined)
+            target.name = value;
+        else if (frame.name === "url" && target.url === undefined)
+            target.url = value;
+        else if (frame.name === "photo" && target.photo === undefined)
+            target.photo = value;
+        return;
+    }
+    // Entry target.
+    switch (frame.name) {
+        case "url":
+            target.url ??= value;
+            break;
+        case "name":
+            target.name ??= value;
+            break;
+        case "content":
+            target.content ??= value;
+            break;
+        case "summary":
+            target.content ??= value;
+            break;
+        case "published":
+            target.published ??= value;
+            break;
+        case "photo":
+            target.photo.push(value);
+            break;
+        case "category":
+            target.category.push(value);
+            break;
+        case "in-reply-to":
+            target.inReplyTo ??= value;
+            break;
+        case "like-of":
+            target.likeOf ??= value;
+            break;
+        default:
+            break;
+    }
+}
+function isCard(target) {
+    return !("photo" in target && Array.isArray(target.photo));
+}
+//# sourceMappingURL=hfeed.js.map

package/dist/hfeed.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"hfeed.js","sourceRoot":"","sources":["../src/hfeed.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;GAiBG;AA+BH,SAAS,OAAO,CAAC,KAAoB;IACnC,OAAO,KAAK,KAAK,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;AACzE,CAAC;AAED;;;;;GAKG;AACH,MAAM,aAAa,GAAG,IAAI,GAAG,CAAC;IAC5B,MAAM;IACN,MAAM;IACN,IAAI;IACJ,KAAK;IACL,OAAO;IACP,IAAI;IACJ,KAAK;IACL,OAAO;IACP,MAAM;IACN,MAAM;IACN,OAAO;IACP,QAAQ;IACR,OAAO;IACP,KAAK;CACN,CAAC,CAAC;AAEH,SAAS,QAAQ,CAAC,IAAY,EAAE,IAAY;IAC1C,IAAI,CAAC;QACH,OAAO,IAAI,GAAG,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,QAAQ,EAAE,CAAC;IACxC,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,IAAI,CAAC;IACd,CAAC;AACH,CAAC;AAED,6EAA6E;AAC7E,SAAS,IAAI,CAAC,KAAa;IACzB,IAAI,CAAC,GAAG,UAAU,CAAC;IACnB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACtC,CAAC,IAAI,KAAK,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC;QACzB,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,EAAE,UAAU,CAAC,CAAC;IAC/B,CAAC;IACD,OAAO,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC;AAChC,CAAC;AAED,4EAA4E;AAC5E,SAAS,aAAa,CACpB,SAAmB;IAEnB,KAAK,MAAM,GAAG,IAAI,SAAS,EAAE,CAAC;QAC5B,MAAM,KAAK,GAAG,mBAAmB,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;QAC5C,IAAI,KAAK,EAAE,CAAC;YACV,OAAO,EAAE,IAAI,EAAE,KAAK,CAAC,CAAC,CAAW,EAAE,MAAM,EAAE,KAAK,CAAC,CAAC,CAAe,EAAE,CAAC;QACtE,CAAC;IACH,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED,SAAS,WAAW,CAAC,KAAmB,EAAE,OAAe;IACvD,MAAM,GAAG,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;IACjE,MAAM,OAAO,GAA2B,KAAK,CAAC,OAAO;QACnD,CAAC,CAAC,EAAE,IAAI,EAAE,KAAK,CAAC,OAAO,EAAE;QACzB,CAAC,CAAC,SAAS,CAAC;IACd,MAAM,MAAM,GAA0B,KAAK,CAAC,MAAM;QAChD,CAAC,CAAC,CAAC,GAAG,EAAE;YACJ,MAAM,IAAI,GAAc;gBACtB,IAAI,EAAE,MAAM;gBACZ,GAAG,CAAC,KAAK,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,KAAK,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;gBACzD,GAAG,CAAC,KAAK,CAAC,MAAM,CAAC,GAAG;oBAClB,CAAC,CAAC,EAAE,GAAG,EAAE,QAAQ,CAAC,KAAK,CAAC,MAAM,CAAC,GAAG,EAAE,OAAO,CAAC,EAAE;oBAC9C,CAAC,CAAC,EAAE,CAAC;gBACP,GAAG,CAAC,KAAK,CAAC,MAAM,CAAC,KAAK;oBACpB,CAAC,CAAC,EAAE,KAAK,EAAE,QAAQ,CAAC,KAAK,CAAC,MAAM,CAAC,KAAK,EAAE,OAAO,CAAC,EAAE;oBAClD,CAAC,CAAC,EAAE,CAAC;aACR,CAAC;YACF,OAAO,IAAI,CAAC,IAAI,IAAI,IAAI,CAAC,GAAG,IAAI,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,SAAS,CAAC;QAChE,CAAC,CAAC,EAAE;QACN,CAAC,CAAC,SAAS,CAAC;IACd,MAAM,EAAE,GAAG,GAAG,IAAI,IAAI,CAAC,GAAG,KAAK,CAAC,IAAI,IAAI,EAAE,GAAG,KAAK,CAAC,OAAO,IAAI,EAAE,EAAE,CAAC,CAAC;IAEpE,MAAM,GAAG,GAA4B,EAAE,IAAI,EAAE,OAAO,EAAE,GAAG,EAAE,EAAE,EAAE,CAAC;IAChE,IAAI,GAAG;QAAE,GAAG,CAAC,GAAG,GAAG,GAAG,CAAC;IACvB,IAAI,KAAK,CAAC,SAAS;QAAE,GAAG,CAAC,SAAS,GAAG,KAAK,CAAC,SAAS,CAAC;IACrD,IAAI,KAAK,CAAC,IAAI;QAAE,GAAG,CAAC,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC;IACtC,IAAI,OAAO;QAAE,GAAG,CAAC,OAAO,GAAG,OAAO,CAAC;IACnC,IAAI,MAAM;QAAE,GAAG,CAAC,MAAM,GAAG,MAAM,CAAC;IAChC,IAAI,KAAK,CAAC,QAAQ,CAAC,MAAM,GAAG,CAAC;QAAE,GAAG,CAAC,QAAQ,GAAG,KAAK,CAAC,QAAQ,CAAC;IAC7D,IAAI,KAAK,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC;QACxB,GAAG,CAAC,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,QAAQ,CAAC,CAAC,EAAE,OAAO,CAAC,CAAC,CAAC;IAC3D,IAAI,KAAK,CAAC,SAAS;QAAE,GAAG,CAAC,aAAa,CAAC,GAAG,QAAQ,CAAC,KAAK,CAAC,SAAS,EAAE,OAAO,CAAC,CAAC;IAC7E,IAAI,KAAK,CAAC,MAAM;QAAE,GAAG,CAAC,SAAS,CAAC,GAAG,QAAQ,CAAC,KAAK,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;IACnE,OAAO,GAA0B,CAAC;AACpC,CAAC;AAED;;;GAGG;AACH,MAAM,CAAC,KAAK,UAAU,UAAU,CAC9B,IAAY,EACZ,OAAe;IAEf,IAAI,IAAI,KAAK,EAAE;QAAE,OAAO,EAAE,CAAC;IAE3B,MAAM,OAAO,GAAe,EAAE,CAAC;IAC/B,MAAM,UAAU,GAAmB,EAAE,CAAC;IACtC,MAAM,SAAS,GAAkB,EAAE,CAAC;IACpC,MAAM,SAAS,GAAgB,EAAE,CAAC;IAElC,MAAM,GAAG,GAAG,CAAI,GAAQ,EAAiB,EAAE,CAAC,GAAG,CAAC,GAAG,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAEhE,MAAM,QAAQ,GAAG,IAAI,YAAY,EAAE,CAAC,EAAE,CAAC,GAAG,EAAE;QAC1C,OAAO,CAAC,EAAE;YACR,MAAM,SAAS,GAAG,OAAO,CAAC,EAAE,CAAC,YAAY,CAAC,OAAO,CAAC,CAAC,CAAC;YACpD,MAAM,MAAM,GAAG,aAAa,CAAC,GAAG,CAAC,EAAE,CAAC,OAAO,CAAC,CAAC;YAE7C,IAAI,SAAS,CAAC,QAAQ,CAAC,SAAS,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC;gBAC7C,MAAM,KAAK,GAAiB,EAAE,KAAK,EAAE,EAAE,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC;gBACxD,UAAU,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;gBACvB,EAAE,CAAC,QAAQ,CAAC,GAAG,EAAE;oBACf,MAAM,QAAQ,GAAG,UAAU,CAAC,GAAG,EAAE,CAAC;oBAClC,IAAI,QAAQ;wBAAE,OAAO,CAAC,IAAI,CAAC,WAAW,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC,CAAC;gBAC7D,CAAC,CAAC,CAAC;gBACH,OAAO;YACT,CAAC;YAED,IAAI,SAAS,CAAC,QAAQ,CAAC,QAAQ,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC;gBAC5C,MAAM,IAAI,GAAgB,EAAE,CAAC;gBAC7B,MAAM,KAAK,GAAG,GAAG,CAAC,UAAU,CAAC,CAAC;gBAC9B,iEAAiE;gBACjE,IAAI,KAAK,IAAI,KAAK,CAAC,MAAM,KAAK,SAAS;oBAAE,KAAK,CAAC,MAAM,GAAG,IAAI,CAAC;gBAC7D,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;gBACrB,EAAE,CAAC,QAAQ,CAAC,GAAG,EAAE;oBACf,SAAS,CAAC,GAAG,EAAE,CAAC;gBAClB,CAAC,CAAC,CAAC;gBACH,OAAO;YACT,CAAC;YAED,MAAM,IAAI,GAAG,aAAa,CAAC,SAAS,CAAC,CAAC;YACtC,IAAI,IAAI,KAAK,IAAI;gBAAE,OAAO;YAC1B,MAAM,MAAM,GAAG,GAAG,CAAC,SAAS,CAAC,IAAI,GAAG,CAAC,UAAU,CAAC,CAAC;YACjD,IAAI,MAAM,KAAK,SAAS;gBAAE,OAAO;YAEjC,MAAM,KAAK,GAAc;gBACvB,IAAI,EAAE,IAAI,CAAC,IAAI;gBACf,MAAM,EAAE,IAAI,CAAC,MAAM;gBACnB,MAAM;gBACN,GAAG,EAAE,EAAE;aACR,CAAC;YAEF,2EAA2E;YAC3E,uEAAuE;YACvE,IAAI,IAAI,CAAC,MAAM,KAAK,GAAG,EAAE,CAAC;gBACxB,MAAM,IAAI,GACR,EAAE,CAAC,YAAY,CAAC,MAAM,CAAC;oBACvB,EAAE,CAAC,YAAY,CAAC,KAAK,CAAC;oBACtB,EAAE,CAAC,YAAY,CAAC,MAAM,CAAC;oBACvB,EAAE,CAAC,YAAY,CAAC,QAAQ,CAAC,CAAC;gBAC5B,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;oBAClB,cAAc,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC;oBAC5B,OAAO;gBACT,CAAC;YACH,CAAC;iBAAM,IAAI,IAAI,CAAC,MAAM,KAAK,IAAI,EAAE,CAAC;gBAChC,MAAM,IAAI,GAAG,EAAE,CAAC,YAAY,CAAC,UAAU,CAAC,IAAI,EAAE,CAAC,YAAY,CAAC,OAAO,CAAC,CAAC;gBACrE,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;oBAClB,cAAc,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC;oBAC5B,OAAO;gBACT,CAAC;YACH,CAAC;YAED,+DAA+D;YAC/D,IAAI,MAAM;gBAAE,OAAO;YAEnB,SAAS,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YACtB,EAAE,CAAC,QAAQ,CAAC,GAAG,EAAE;gBACf,MAAM,QAAQ,GAAG,SAAS,CAAC,GAAG,EAAE,CAAC;gBACjC,IAAI,QAAQ;oBAAE,cAAc,CAAC,QAAQ,EAAE,QAAQ,CAAC,GAAG,CAAC,IAAI,EAAE,CAAC,CAAC;YAC9D,CAAC,CAAC,CAAC;QACL,CAAC;QACD,IAAI,CAAC,KAAK;YACR,MAAM,KAAK,GAAG,GAAG,CAAC,SAAS,CAAC,CAAC;YAC7B,IAAI,KAAK;gBAAE,KAAK,CAAC,GAAG,IAAI,KAAK,CAAC,IAAI,CAAC;QACrC,CAAC;KACF,CAAC,CAAC;IAEH,MAAM,QAAQ,CAAC,SAAS,CAAC,IAAI,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;IACpD,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,qEAAqE;AACrE,SAAS,cAAc,CAAC,KAAgB,EAAE,KAAa;IACrD,IAAI,KAAK,KAAK,EAAE;QAAE,OAAO;IACzB,MAAM,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC;IAC5B,wBAAwB;IACxB,IAAI,MAAM,CAAC,MAAM,CAAC,EAAE,CAAC;QACnB,IAAI,KAAK,CAAC,IAAI,KAAK,MAAM,IAAI,MAAM,CAAC,IAAI,KAAK,SAAS;YAAE,MAAM,CAAC,IAAI,GAAG,KAAK,CAAC;aACvE,IAAI,KAAK,CAAC,IAAI,KAAK,KAAK,IAAI,MAAM,CAAC,GAAG,KAAK,SAAS;YACvD,MAAM,CAAC,GAAG,GAAG,KAAK,CAAC;aAChB,IAAI,KAAK,CAAC,IAAI,KAAK,OAAO,IAAI,MAAM,CAAC,KAAK,KAAK,SAAS;YAC3D,MAAM,CAAC,KAAK,GAAG,KAAK,CAAC;QACvB,OAAO;IACT,CAAC;IACD,gBAAgB;IAChB,QAAQ,KAAK,CAAC,IAAI,EAAE,CAAC;QACnB,KAAK,KAAK;YACR,MAAM,CAAC,GAAG,KAAK,KAAK,CAAC;YACrB,MAAM;QACR,KAAK,MAAM;YACT,MAAM,CAAC,IAAI,KAAK,KAAK,CAAC;YACtB,MAAM;QACR,KAAK,SAAS;YACZ,MAAM,CAAC,OAAO,KAAK,KAAK,CAAC;YACzB,MAAM;QACR,KAAK,SAAS;YACZ,MAAM,CAAC,OAAO,KAAK,KAAK,CAAC;YACzB,MAAM;QACR,KAAK,WAAW;YACd,MAAM,CAAC,SAAS,KAAK,KAAK,CAAC;YAC3B,MAAM;QACR,KAAK,OAAO;YACV,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YACzB,MAAM;QACR,KAAK,UAAU;YACb,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YAC5B,MAAM;QACR,KAAK,aAAa;YAChB,MAAM,CAAC,SAAS,KAAK,KAAK,CAAC;YAC3B,MAAM;QACR,KAAK,SAAS;YACZ,MAAM,CAAC,MAAM,KAAK,KAAK,CAAC;YACxB,MAAM;QACR;YACE,MAAM;IACV,CAAC;AACH,CAAC;AAED,SAAS,MAAM,CAAC,MAAkC;IAChD,OAAO,CAAC,CAAC,OAAO,IAAI,MAAM,IAAI,KAAK,CAAC,OAAO,CAAE,MAAuB,CAAC,KAAK,CAAC,CAAC,CAAC;AAC/E,CAAC"}

package/dist/index.d.ts

@@ -0,0 +1,39 @@
+/**
+ * `@dwk/microsub` — a [Microsub](https://indieweb.org/Microsub-spec) server: the
+ * IndieWeb **read side**.
+ *
+ * Endpoint package: exports a factory returning a `fetch`-compatible handler,
+ * mountable under a path prefix so it composes with other `@dwk` packages in one
+ * Worker. It manages feed **subscriptions** organised into **channels**, polls
+ * and parses sources server-side (Atom / RSS / JSON Feed / `h-feed`), and serves
+ * a normalised **JF2** timeline to reader clients (Monocle, Together,
+ * Indigenous) — keeping the user's reading state on infrastructure they own.
+ *
+ * It consumes the DPoP-bound IndieAuth access tokens issued by `@dwk/indieauth`
+ * (same authorization as `@dwk/micropub`), so a token minted for a different
+ * `me` cannot read here. Subscriptions, timeline, and read-state live in D1 — a
+ * strongly-consistent store, never KV. Polling runs off the read path on a
+ * Cron-triggered queue; every outbound fetch is SSRF-guarded.
+ *
+ * @see spec/packages/microsub.md
+ * @packageDocumentation
+ */
+export { createMicrosub } from "./handler";
+export type { MicrosubHandler } from "./handler";
+export { createMicrosubPoller } from "./poll";
+export type { MicrosubScheduledHandler } from "./poll";
+export { createMicrosubQueueConsumer } from "./consumer";
+export type { MicrosubQueueConsumer, ConsumerOptions } from "./consumer";
+export { resolveConfig } from "./config";
+export type { MicrosubConfig, MicrosubEnv, ResolvedConfig } from "./config";
+export { createMicrosubStore, NOTIFICATIONS_CHANNEL, type MicrosubStore, type MicrosubStoreEnv, type ChannelRecord, type FollowRecord, type StoredItem, type ItemPage, type ListOptions, type FeedCache, } from "./store";
+export { authorize, tokenFromHeader, hasScope, type AuthEnv, type AuthResult, type AuthSuccess, type AuthFailure, } from "./auth";
+export { parseFeed, type Jf2Entry, type Jf2Author, type Jf2Content, } from "./jf2";
+export { parseHFeed } from "./hfeed";
+export { discoverFeed, fetchFeed, type DiscoveredFeed, type FetchedFeed, type DiscoveryOptions, } from "./discovery";
+export { safeFetch, assertPublicUrl, isPrivateOrReservedHost, SsrfError, type SsrfReason, } from "./safe-fetch";
+export type { FetchLike } from "./fetch";
+export type { MicrosubJob, PollJob } from "./queue";
+export { MicrosubLogEvent } from "./log";
+export type { Logger, Metrics } from "@dwk/log";
+//# sourceMappingURL=index.d.ts.map

package/dist/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;GAmBG;AAEH,OAAO,EAAE,cAAc,EAAE,MAAM,WAAW,CAAC;AAC3C,YAAY,EAAE,eAAe,EAAE,MAAM,WAAW,CAAC;AAEjD,OAAO,EAAE,oBAAoB,EAAE,MAAM,QAAQ,CAAC;AAC9C,YAAY,EAAE,wBAAwB,EAAE,MAAM,QAAQ,CAAC;AAEvD,OAAO,EAAE,2BAA2B,EAAE,MAAM,YAAY,CAAC;AACzD,YAAY,EAAE,qBAAqB,EAAE,eAAe,EAAE,MAAM,YAAY,CAAC;AAEzE,OAAO,EAAE,aAAa,EAAE,MAAM,UAAU,CAAC;AACzC,YAAY,EAAE,cAAc,EAAE,WAAW,EAAE,cAAc,EAAE,MAAM,UAAU,CAAC;AAE5E,OAAO,EACL,mBAAmB,EACnB,qBAAqB,EACrB,KAAK,aAAa,EAClB,KAAK,gBAAgB,EACrB,KAAK,aAAa,EAClB,KAAK,YAAY,EACjB,KAAK,UAAU,EACf,KAAK,QAAQ,EACb,KAAK,WAAW,EAChB,KAAK,SAAS,GACf,MAAM,SAAS,CAAC;AAEjB,OAAO,EACL,SAAS,EACT,eAAe,EACf,QAAQ,EACR,KAAK,OAAO,EACZ,KAAK,UAAU,EACf,KAAK,WAAW,EAChB,KAAK,WAAW,GACjB,MAAM,QAAQ,CAAC;AAEhB,OAAO,EACL,SAAS,EACT,KAAK,QAAQ,EACb,KAAK,SAAS,EACd,KAAK,UAAU,GAChB,MAAM,OAAO,CAAC;AAEf,OAAO,EAAE,UAAU,EAAE,MAAM,SAAS,CAAC;AAErC,OAAO,EACL,YAAY,EACZ,SAAS,EACT,KAAK,cAAc,EACnB,KAAK,WAAW,EAChB,KAAK,gBAAgB,GACtB,MAAM,aAAa,CAAC;AAErB,OAAO,EACL,SAAS,EACT,eAAe,EACf,uBAAuB,EACvB,SAAS,EACT,KAAK,UAAU,GAChB,MAAM,cAAc,CAAC;AAEtB,YAAY,EAAE,SAAS,EAAE,MAAM,SAAS,CAAC;AACzC,YAAY,EAAE,WAAW,EAAE,OAAO,EAAE,MAAM,SAAS,CAAC;AAEpD,OAAO,EAAE,gBAAgB,EAAE,MAAM,OAAO,CAAC;AACzC,YAAY,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,UAAU,CAAC"}

package/dist/index.js

@@ -0,0 +1,32 @@
+/**
+ * `@dwk/microsub` — a [Microsub](https://indieweb.org/Microsub-spec) server: the
+ * IndieWeb **read side**.
+ *
+ * Endpoint package: exports a factory returning a `fetch`-compatible handler,
+ * mountable under a path prefix so it composes with other `@dwk` packages in one
+ * Worker. It manages feed **subscriptions** organised into **channels**, polls
+ * and parses sources server-side (Atom / RSS / JSON Feed / `h-feed`), and serves
+ * a normalised **JF2** timeline to reader clients (Monocle, Together,
+ * Indigenous) — keeping the user's reading state on infrastructure they own.
+ *
+ * It consumes the DPoP-bound IndieAuth access tokens issued by `@dwk/indieauth`
+ * (same authorization as `@dwk/micropub`), so a token minted for a different
+ * `me` cannot read here. Subscriptions, timeline, and read-state live in D1 — a
+ * strongly-consistent store, never KV. Polling runs off the read path on a
+ * Cron-triggered queue; every outbound fetch is SSRF-guarded.
+ *
+ * @see spec/packages/microsub.md
+ * @packageDocumentation
+ */
+export { createMicrosub } from "./handler";
+export { createMicrosubPoller } from "./poll";
+export { createMicrosubQueueConsumer } from "./consumer";
+export { resolveConfig } from "./config";
+export { createMicrosubStore, NOTIFICATIONS_CHANNEL, } from "./store";
+export { authorize, tokenFromHeader, hasScope, } from "./auth";
+export { parseFeed, } from "./jf2";
+export { parseHFeed } from "./hfeed";
+export { discoverFeed, fetchFeed, } from "./discovery";
+export { safeFetch, assertPublicUrl, isPrivateOrReservedHost, SsrfError, } from "./safe-fetch";
+export { MicrosubLogEvent } from "./log";
+//# sourceMappingURL=index.js.map

package/dist/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;GAmBG;AAEH,OAAO,EAAE,cAAc,EAAE,MAAM,WAAW,CAAC;AAG3C,OAAO,EAAE,oBAAoB,EAAE,MAAM,QAAQ,CAAC;AAG9C,OAAO,EAAE,2BAA2B,EAAE,MAAM,YAAY,CAAC;AAGzD,OAAO,EAAE,aAAa,EAAE,MAAM,UAAU,CAAC;AAGzC,OAAO,EACL,mBAAmB,EACnB,qBAAqB,GAStB,MAAM,SAAS,CAAC;AAEjB,OAAO,EACL,SAAS,EACT,eAAe,EACf,QAAQ,GAKT,MAAM,QAAQ,CAAC;AAEhB,OAAO,EACL,SAAS,GAIV,MAAM,OAAO,CAAC;AAEf,OAAO,EAAE,UAAU,EAAE,MAAM,SAAS,CAAC;AAErC,OAAO,EACL,YAAY,EACZ,SAAS,GAIV,MAAM,aAAa,CAAC;AAErB,OAAO,EACL,SAAS,EACT,eAAe,EACf,uBAAuB,EACvB,SAAS,GAEV,MAAM,cAAc,CAAC;AAKtB,OAAO,EAAE,gBAAgB,EAAE,MAAM,OAAO,CAAC"}

package/dist/jf2.d.ts

@@ -0,0 +1,69 @@
+/**
+ * `@dwk/microsub` — feed → JF2 normalisation.
+ *
+ * Microsub serves a single normalised timeline regardless of a source's wire
+ * format. This module turns the four feed formats a reader meets — JSON Feed,
+ * Atom, RSS 2.0, and `h-feed` microformats (the last parsed upstream in
+ * {@link ./hfeed}) — into [JF2](https://jf2.spec.indieweb.org/) `entry` objects.
+ *
+ * Everything here is **pure**: a feed body (and the format-sniffing on its
+ * content type) in, an array of JF2 entries out, each carrying a stable `_id`
+ * derived from the source's own identifier so the store can dedupe across polls.
+ * The XML formats go through the dependency-free reader in {@link ./xml}.
+ *
+ * @packageDocumentation
+ */
+/** A JF2 author card. */
+export interface Jf2Author {
+    readonly type: "card";
+    readonly name?: string;
+    readonly url?: string;
+    readonly photo?: string;
+}
+/** Structured JF2 content (`html` and/or its plain-text rendering). */
+export interface Jf2Content {
+    readonly html?: string;
+    readonly text?: string;
+}
+/** A normalised JF2 timeline entry. */
+export interface Jf2Entry {
+    readonly type: "entry";
+    /**
+     * Stable per-entry identifier, derived from the source's own id/guid/url. The
+     * store keys timeline rows on this so re-polling a feed does not duplicate
+     * entries.
+     */
+    readonly _id: string;
+    readonly url?: string;
+    readonly published?: string;
+    readonly name?: string;
+    readonly summary?: string;
+    readonly content?: Jf2Content;
+    readonly author?: Jf2Author;
+    readonly category?: readonly string[];
+    readonly photo?: readonly string[];
+    readonly "in-reply-to"?: string;
+    readonly "like-of"?: string;
+    /** Read flag, attached by the store on read (never by the parser). */
+    readonly _is_read?: boolean;
+}
+/**
+ * Order a feed's entries oldest-first for insertion, so the newest entry is
+ * appended last and receives the highest paging `seq`.
+ *
+ * When the entries carry dates they are sorted chronologically (a feed's own
+ * order is not trusted). When none do — some `h-feed` pages — the feed is
+ * assumed newest-first and simply reversed. The sort is stable, so undated
+ * entries keep their relative position among dated ones.
+ */
+export declare function orderEntriesForInsert(entries: readonly Jf2Entry[]): Jf2Entry[];
+/**
+ * Parse a fetched feed body into JF2 entries, sniffing the format from the
+ * content type and the document root. Returns `[]` for an unrecognised or
+ * unparseable body (the caller treats an empty parse as "nothing new").
+ *
+ * `h-feed` HTML is **not** handled here — it needs the runtime's `HTMLRewriter`
+ * and lives in {@link ./hfeed}; this module is pure and runtime-free.
+ */
+export declare function parseFeed(body: string, contentType: string, baseUrl: string): Jf2Entry[];
+//# sourceMappingURL=jf2.d.ts.map

package/dist/jf2.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"jf2.d.ts","sourceRoot":"","sources":["../src/jf2.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;GAcG;AAWH,yBAAyB;AACzB,MAAM,WAAW,SAAS;IACxB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,QAAQ,CAAC,IAAI,CAAC,EAAE,MAAM,CAAC;IACvB,QAAQ,CAAC,GAAG,CAAC,EAAE,MAAM,CAAC;IACtB,QAAQ,CAAC,KAAK,CAAC,EAAE,MAAM,CAAC;CACzB;AAED,uEAAuE;AACvE,MAAM,WAAW,UAAU;IACzB,QAAQ,CAAC,IAAI,CAAC,EAAE,MAAM,CAAC;IACvB,QAAQ,CAAC,IAAI,CAAC,EAAE,MAAM,CAAC;CACxB;AAED,uCAAuC;AACvC,MAAM,WAAW,QAAQ;IACvB,QAAQ,CAAC,IAAI,EAAE,OAAO,CAAC;IACvB;;;;OAIG;IACH,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAC;IACrB,QAAQ,CAAC,GAAG,CAAC,EAAE,MAAM,CAAC;IACtB,QAAQ,CAAC,SAAS,CAAC,EAAE,MAAM,CAAC;IAC5B,QAAQ,CAAC,IAAI,CAAC,EAAE,MAAM,CAAC;IACvB,QAAQ,CAAC,OAAO,CAAC,EAAE,MAAM,CAAC;IAC1B,QAAQ,CAAC,OAAO,CAAC,EAAE,UAAU,CAAC;IAC9B,QAAQ,CAAC,MAAM,CAAC,EAAE,SAAS,CAAC;IAC5B,QAAQ,CAAC,QAAQ,CAAC,EAAE,SAAS,MAAM,EAAE,CAAC;IACtC,QAAQ,CAAC,KAAK,CAAC,EAAE,SAAS,MAAM,EAAE,CAAC;IACnC,QAAQ,CAAC,aAAa,CAAC,EAAE,MAAM,CAAC;IAChC,QAAQ,CAAC,SAAS,CAAC,EAAE,MAAM,CAAC;IAC5B,sEAAsE;IACtE,QAAQ,CAAC,QAAQ,CAAC,EAAE,OAAO,CAAC;CAC7B;AAgRD;;;;;;;;GAQG;AACH,wBAAgB,qBAAqB,CACnC,OAAO,EAAE,SAAS,QAAQ,EAAE,GAC3B,QAAQ,EAAE,CAYZ;AAcD;;;;;;;GAOG;AACH,wBAAgB,SAAS,CACvB,IAAI,EAAE,MAAM,EACZ,WAAW,EAAE,MAAM,EACnB,OAAO,EAAE,MAAM,GACd,QAAQ,EAAE,CAYZ"}