@project-ava-supernova/core 0.2.71

package/dist/news/index.d.ts

@@ -0,0 +1,163 @@
+/** One story as an outlet actually published it. */
+export interface NewsHit {
+    /** The outlet's headline, verbatim. This is the bias evidence — in their words. */
+    title: string;
+    url: string;
+    /** A short verbatim excerpt. The ONLY text a quote may be drawn from. */
+    excerpt: string;
+    outlet: string | null;
+    age?: string | null;
+    thumbnail?: string | null;
+}
+/** News search, injected by the surface (the web supplies Brave News). */
+export type NewsSearchFn = (query: string, count?: number, freshness?: 'pd' | 'pw' | 'pm') => Promise<NewsHit[]>;
+/**
+ * Read a desk's front pages — the RSS feeds of real newsrooms. Injected by the
+ * surface (the web supplies `readDeskFeeds`).
+ *
+ * This is DISCOVERY, and it is deliberately not a search. A search is a pull:
+ * you only get back what you already suspected, and nobody thinks to ask for
+ * "Bangkok bar fire" at 6am. A front page is a push: the world editor put it
+ * there because it happened. So the newsroom SEES first (feeds), CHOOSES second
+ * (Ava), and only then VERIFIES with search (Brave) — which is the right tool for
+ * standing up a story she has already found, and the wrong one for deciding what
+ * the news is.
+ */
+export type NewsFeedFn = (desk: string) => Promise<NewsHit[]>;
+/** A finished article, on its way to the card + the drafts table. */
+export interface ArticleInput {
+    headline: string;
+    standfirst?: string;
+    body: string;
+    category: string;
+    sources: Array<{
+        outlet: string;
+        headline: string;
+        url: string;
+    }>;
+    quotes: Array<{
+        text: string;
+        speaker?: string;
+        outlet: string;
+        url: string;
+    }>;
+    coverage?: Record<string, unknown>;
+    unverified: string[];
+    ava_read?: string;
+    image_prompt?: string;
+}
+/** Article sink, injected by the surface (renders the card, saves the draft). */
+export interface ArticleStore {
+    save(article: ArticleInput): Promise<{
+        id: string | null;
+    }>;
+}
+/**
+ * One story on the front-page menu — a story she has SEEN, and thinks is worth
+ * writing. `why` is the whole point: the selection is the skill, and the operator
+ * is owed her reason, not a ranked list he has to re-judge himself.
+ */
+export interface StorySuggestion {
+    desk: string;
+    headline: string;
+    outlet: string;
+    /** The URL she saw it at. Checked against the fetched corpus — a story that
+     *  isn't in anything she pulled cannot reach the menu. */
+    url: string;
+    /** Why it's worth writing: what's actually at stake, or what everyone is missing. */
+    why: string;
+    /** Optional: how she'd come at it. */
+    angle?: string;
+}
+/** Story-menu sink, injected by the surface (renders the front page). */
+export interface StoryStore {
+    suggest(stories: StorySuggestion[]): Promise<void>;
+}
+/**
+ * Everything she fetched this turn, kept so write_article can CHECK her quotes
+ * against it. Populated by research_story / fact_check; read by write_article.
+ * Without this, "quote verbatim from a page you fetched" is an honour system —
+ * and an honour system is exactly what we refuse to ship.
+ */
+export interface FetchedCorpus {
+    hits: NewsHit[];
+}
+export interface CoverageCluster {
+    outlets: Array<{
+        outlet: string | null;
+        headline: string;
+        url: string;
+        excerpt: string;
+    }>;
+    syndicated: boolean;
+    wire: string | null;
+    confidence: 'high' | 'likely' | 'borderline' | 'n/a';
+    similarity: number;
+}
+/** Normalise for comparison: case, punctuation and spacing carry no meaning here. */
+export declare function normalise(text: string): string;
+/**
+ * Group hits by the underlying TEXT. Several outlets in one cluster = one report
+ * echoed, NOT corroboration.
+ *
+ * THRESHOLD — measured on live coverage, not taken from a paper. The first cut
+ * used 0.45 (the textbook value for full documents) and NEVER FIRED, because the
+ * search index returns a partial snippet of each article and different fragments
+ * get extracted from each copy. Measured:
+ *
+ *   independent reporting (Guardian / ESPN / Al Jazeera, same story) . 0.02–0.06
+ *   same-chain syndication (Sacramento Bee <-> Merced Sun-Star) ...... 0.175
+ *   republished copy (Sharewise <-> Yahoo Finance <-> Yahoo) ......... 0.32–0.41
+ *
+ * A clean gap sits between ~0.06 and ~0.17, so 0.15 separates them with headroom
+ * either side. Verified: it collapsed an 8-outlet Yahoo/Sharewise cluster and a
+ * 2-paper McClatchy cluster, while leaving 20 genuinely independent FIFA reports
+ * intact.
+ *
+ * Only ever run this over ONE story's coverage. On a broad category sweep it
+ * would cluster unrelated articles that merely share a template.
+ */
+export declare function clusterCoverage(hits: NewsHit[], threshold?: number): CoverageCluster[];
+/**
+ * The shape the Correspondent must reason over.
+ *
+ * `independent_sources` is the number that means something. The total is named
+ * to make it hard to misread as corroboration, because that misreading is the
+ * single worst thing this product could do.
+ */
+export declare function summariseCoverage(hits: NewsHit[]): {
+    independent_sources: number;
+    total_outlets_including_syndicated_copies: number;
+    wire_service_detected: string | null;
+    note: string;
+    coverage: CoverageCluster[];
+    headlines: {
+        outlet: string | null;
+        headline: string;
+        url: string;
+        age: string | null;
+    }[];
+};
+/**
+ * Does this quote actually appear in something she fetched?
+ *
+ * A fabricated or misattributed quote is a libel risk and the end of the
+ * product. So we do not ask her to be honest — we check. Anything she cannot
+ * point to in a fetched excerpt is refused, and she is told to fix it in the
+ * same turn.
+ *
+ * We compare on the NORMALISED text, so punctuation, casing and smart-quote
+ * mangling don't cause a false rejection — but the words must genuinely be
+ * there, in that order.
+ *
+ * The honest limitation, stated: the index gives us an EXCERPT, not the whole
+ * article. A real quote can legitimately live in the part we never saw. So a
+ * failure here does not mean she invented it — it means she cannot evidence it,
+ * and an unevidenced quote does not go in. The safe direction to fail in.
+ */
+export declare function verifyQuote(quote: string, corpus: FetchedCorpus): {
+    ok: boolean;
+    outlet?: string;
+    url?: string;
+};
+//# sourceMappingURL=index.d.ts.map

package/dist/news/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/news/index.ts"],"names":[],"mappings":"AAwBA,oDAAoD;AACpD,MAAM,WAAW,OAAO;IACtB,mFAAmF;IACnF,KAAK,EAAE,MAAM,CAAC;IACd,GAAG,EAAE,MAAM,CAAC;IACZ,yEAAyE;IACzE,OAAO,EAAE,MAAM,CAAC;IAChB,MAAM,EAAE,MAAM,GAAG,IAAI,CAAC;IACtB,GAAG,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IACpB,SAAS,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;CAC3B;AAED,0EAA0E;AAC1E,MAAM,MAAM,YAAY,GAAG,CACzB,KAAK,EAAE,MAAM,EACb,KAAK,CAAC,EAAE,MAAM,EACd,SAAS,CAAC,EAAE,IAAI,GAAG,IAAI,GAAG,IAAI,KAC3B,OAAO,CAAC,OAAO,EAAE,CAAC,CAAC;AAExB;;;;;;;;;;;GAWG;AACH,MAAM,MAAM,UAAU,GAAG,CAAC,IAAI,EAAE,MAAM,KAAK,OAAO,CAAC,OAAO,EAAE,CAAC,CAAC;AAE9D,qEAAqE;AACrE,MAAM,WAAW,YAAY;IAC3B,QAAQ,EAAE,MAAM,CAAC;IACjB,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,IAAI,EAAE,MAAM,CAAC;IACb,QAAQ,EAAE,MAAM,CAAC;IACjB,OAAO,EAAE,KAAK,CAAC;QAAE,MAAM,EAAE,MAAM,CAAC;QAAC,QAAQ,EAAE,MAAM,CAAC;QAAC,GAAG,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;IAClE,MAAM,EAAE,KAAK,CAAC;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,OAAO,CAAC,EAAE,MAAM,CAAC;QAAC,MAAM,EAAE,MAAM,CAAC;QAAC,GAAG,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;IAC/E,QAAQ,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;IACnC,UAAU,EAAE,MAAM,EAAE,CAAC;IACrB,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,YAAY,CAAC,EAAE,MAAM,CAAC;CACvB;AAED,iFAAiF;AACjF,MAAM,WAAW,YAAY;IAC3B,IAAI,CAAC,OAAO,EAAE,YAAY,GAAG,OAAO,CAAC;QAAE,EAAE,EAAE,MAAM,GAAG,IAAI,CAAA;KAAE,CAAC,CAAC;CAC7D;AAED;;;;GAIG;AACH,MAAM,WAAW,eAAe;IAC9B,IAAI,EAAE,MAAM,CAAC;IACb,QAAQ,EAAE,MAAM,CAAC;IACjB,MAAM,EAAE,MAAM,CAAC;IACf;8DAC0D;IAC1D,GAAG,EAAE,MAAM,CAAC;IACZ,qFAAqF;IACrF,GAAG,EAAE,MAAM,CAAC;IACZ,sCAAsC;IACtC,KAAK,CAAC,EAAE,MAAM,CAAC;CAChB;AAED,yEAAyE;AACzE,MAAM,WAAW,UAAU;IACzB,OAAO,CAAC,OAAO,EAAE,eAAe,EAAE,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC;CACpD;AAED;;;;;GAKG;AACH,MAAM,WAAW,aAAa;IAC5B,IAAI,EAAE,OAAO,EAAE,CAAC;CACjB;AASD,MAAM,WAAW,eAAe;IAC9B,OAAO,EAAE,KAAK,CAAC;QAAE,MAAM,EAAE,MAAM,GAAG,IAAI,CAAC;QAAC,QAAQ,EAAE,MAAM,CAAC;QAAC,GAAG,EAAE,MAAM,CAAC;QAAC,OAAO,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;IAC1F,UAAU,EAAE,OAAO,CAAC;IACpB,IAAI,EAAE,MAAM,GAAG,IAAI,CAAC;IACpB,UAAU,EAAE,MAAM,GAAG,QAAQ,GAAG,YAAY,GAAG,KAAK,CAAC;IACrD,UAAU,EAAE,MAAM,CAAC;CACpB;AAYD,qFAAqF;AACrF,wBAAgB,SAAS,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAE9C;AAgBD;;;;;;;;;;;;;;;;;;;;GAoBG;AACH,wBAAgB,eAAe,CAAC,IAAI,EAAE,OAAO,EAAE,EAAE,SAAS,SAAO,GAAG,eAAe,EAAE,CAkDpF;AAED;;;;;;GAMG;AACH,wBAAgB,iBAAiB,CAAC,IAAI,EAAE,OAAO,EAAE;;;;;;;;;;;;EAmBhD;AAID;;;;;;;;;;;;;;;;GAgBG;AACH,wBAAgB,WAAW,CAAC,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,aAAa,GAAG;IAAE,EAAE,EAAE,OAAO,CAAC;IAAC,MAAM,CAAC,EAAE,MAAM,CAAC;IAAC,GAAG,CAAC,EAAE,MAAM,CAAA;CAAE,CAWhH"}

package/dist/news/index.js

@@ -0,0 +1,184 @@
+// ─── The Newsroom — types, syndication detection, quote verification ─────────
+//
+// Ava writes her OWN account of a story and cites the outlets she read. She is
+// NOT an aggregator — she never republishes another outlet's article. Facts are
+// free; expression is not. (See MidlevelU v. ACI, 11th Cir. 2021: an aggregator
+// that commercially summarised RSS feeds lost.)
+//
+// Two things in here are the whole product, and both are ENFORCED IN CODE rather
+// than merely requested in a prompt — because the same trick that makes
+// write_post reliable (don't ask the model to count characters; reject the post
+// when the count is wrong) is what makes this reliable:
+//
+//   1. SYNDICATION. Reuters writes it, 47 outlets republish it. Counting those
+//      as 47 sources launders one report into fake consensus. clusterCoverage()
+//      collapses them, so the model is handed `independent_sources`, not a
+//      flattering total it can misread.
+//
+//   2. QUOTES. A fabricated or misattributed quote is a libel risk and the end
+//      of the product. verifyQuote() checks every quote against the text she
+//      actually fetched this turn — so a quote she "remembers" cannot get into
+//      an article. She doesn't have to be honest. She has to be *checkable*.
+// ─── Syndication ─────────────────────────────────────────────────────────────
+const WIRE_SERVICES = [
+    'Reuters', 'Associated Press', 'AP', 'AFP', 'Agence France-Presse',
+    'PA Media', 'Bloomberg', 'dpa', 'ANSA', 'EFE',
+];
+function detectWire(texts) {
+    const blob = texts.join(' ');
+    for (const w of WIRE_SERVICES) {
+        // Word-boundary match so "AP" doesn't fire on "apple" or "capital".
+        const re = new RegExp(`(^|[^A-Za-z])${w.replace(/[.*+?^${}()|[\]\\]/g, '\\$&')}([^A-Za-z]|$)`);
+        if (re.test(blob))
+            return w;
+    }
+    return null;
+}
+/** Normalise for comparison: case, punctuation and spacing carry no meaning here. */
+export function normalise(text) {
+    return text.toLowerCase().replace(/[^\p{L}\p{N}\s]/gu, ' ').replace(/\s+/g, ' ').trim();
+}
+function shingles(text, n = 5) {
+    const words = normalise(text).split(' ').filter(Boolean);
+    const out = new Set();
+    for (let i = 0; i + n <= words.length; i++)
+        out.add(words.slice(i, i + n).join(' '));
+    return out;
+}
+function jaccard(a, b) {
+    if (a.size === 0 || b.size === 0)
+        return 0;
+    let inter = 0;
+    for (const x of a)
+        if (b.has(x))
+            inter++;
+    return inter / (a.size + b.size - inter);
+}
+/**
+ * Group hits by the underlying TEXT. Several outlets in one cluster = one report
+ * echoed, NOT corroboration.
+ *
+ * THRESHOLD — measured on live coverage, not taken from a paper. The first cut
+ * used 0.45 (the textbook value for full documents) and NEVER FIRED, because the
+ * search index returns a partial snippet of each article and different fragments
+ * get extracted from each copy. Measured:
+ *
+ *   independent reporting (Guardian / ESPN / Al Jazeera, same story) . 0.02–0.06
+ *   same-chain syndication (Sacramento Bee <-> Merced Sun-Star) ...... 0.175
+ *   republished copy (Sharewise <-> Yahoo Finance <-> Yahoo) ......... 0.32–0.41
+ *
+ * A clean gap sits between ~0.06 and ~0.17, so 0.15 separates them with headroom
+ * either side. Verified: it collapsed an 8-outlet Yahoo/Sharewise cluster and a
+ * 2-paper McClatchy cluster, while leaving 20 genuinely independent FIFA reports
+ * intact.
+ *
+ * Only ever run this over ONE story's coverage. On a broad category sweep it
+ * would cluster unrelated articles that merely share a template.
+ */
+export function clusterCoverage(hits, threshold = 0.15) {
+    const sigs = hits.map((h) => shingles(`${h.title} ${h.excerpt}`));
+    const clusterOf = new Array(hits.length).fill(-1);
+    const groups = [];
+    const peak = [];
+    for (let i = 0; i < hits.length; i++) {
+        if (clusterOf[i] !== -1)
+            continue;
+        const id = groups.length;
+        clusterOf[i] = id;
+        const members = [i];
+        let best = 0;
+        for (let j = i + 1; j < hits.length; j++) {
+            if (clusterOf[j] !== -1)
+                continue;
+            const s = jaccard(sigs[i], sigs[j]);
+            if (s >= threshold) {
+                clusterOf[j] = id;
+                members.push(j);
+                if (s > best)
+                    best = s;
+            }
+        }
+        groups.push(members);
+        peak.push(best);
+    }
+    return groups.map((members, idx) => {
+        const outlets = members.map((i) => ({
+            outlet: hits[i].outlet,
+            headline: hits[i].title,
+            url: hits[i].url,
+            excerpt: hits[i].excerpt.slice(0, 400),
+        }));
+        const similarity = peak[idx];
+        const syndicated = outlets.length > 1;
+        // A borderline call is an HONEST call — she is told to say "I can't tell"
+        // rather than pick an answer she can't support.
+        const confidence = !syndicated
+            ? 'n/a'
+            : similarity >= 0.30 ? 'high'
+                : similarity >= 0.20 ? 'likely'
+                    : 'borderline';
+        return {
+            outlets,
+            syndicated,
+            wire: detectWire(members.map((i) => `${hits[i].title} ${hits[i].excerpt}`)),
+            confidence,
+            similarity: Number(similarity.toFixed(3)),
+        };
+    });
+}
+/**
+ * The shape the Correspondent must reason over.
+ *
+ * `independent_sources` is the number that means something. The total is named
+ * to make it hard to misread as corroboration, because that misreading is the
+ * single worst thing this product could do.
+ */
+export function summariseCoverage(hits) {
+    const clusters = clusterCoverage(hits);
+    const wire = clusters.filter((c) => c.syndicated).map((c) => c.wire).find(Boolean) ?? null;
+    return {
+        independent_sources: clusters.length,
+        total_outlets_including_syndicated_copies: hits.length,
+        wire_service_detected: wire,
+        note: hits.length === 0
+            ? 'Nothing in the index reports this. That means nobody has reported it HERE — it does NOT mean the claim is false, and you must not say it does.'
+            : clusters.length === 1 && hits.length > 1
+                ? `All ${hits.length} outlets are running substantially the same text${wire ? ` (${wire})` : ''}. That is ONE report echoed — not ${hits.length} sources. Say so plainly in the article.`
+                : `${clusters.length} distinct piece(s) of reporting across ${hits.length} outlet(s).`,
+        coverage: clusters,
+        // Headlines side by side. THIS is the bias evidence — in the outlets' own
+        // words. No score, no lean, no verdict. Just what each of them chose to say.
+        headlines: hits.map((h) => ({ outlet: h.outlet, headline: h.title, url: h.url, age: h.age ?? null })),
+    };
+}
+// ─── Quote verification ──────────────────────────────────────────────────────
+/**
+ * Does this quote actually appear in something she fetched?
+ *
+ * A fabricated or misattributed quote is a libel risk and the end of the
+ * product. So we do not ask her to be honest — we check. Anything she cannot
+ * point to in a fetched excerpt is refused, and she is told to fix it in the
+ * same turn.
+ *
+ * We compare on the NORMALISED text, so punctuation, casing and smart-quote
+ * mangling don't cause a false rejection — but the words must genuinely be
+ * there, in that order.
+ *
+ * The honest limitation, stated: the index gives us an EXCERPT, not the whole
+ * article. A real quote can legitimately live in the part we never saw. So a
+ * failure here does not mean she invented it — it means she cannot evidence it,
+ * and an unevidenced quote does not go in. The safe direction to fail in.
+ */
+export function verifyQuote(quote, corpus) {
+    const needle = normalise(quote);
+    if (needle.length < 12)
+        return { ok: false }; // too short to be a meaningful match
+    for (const hit of corpus.hits) {
+        const hay = normalise(`${hit.title} ${hit.excerpt}`);
+        if (hay.includes(needle)) {
+            return { ok: true, outlet: hit.outlet ?? undefined, url: hit.url };
+        }
+    }
+    return { ok: false };
+}
+//# sourceMappingURL=index.js.map

package/dist/news/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/news/index.ts"],"names":[],"mappings":"AAAA,gFAAgF;AAChF,EAAE;AACF,+EAA+E;AAC/E,gFAAgF;AAChF,gFAAgF;AAChF,gDAAgD;AAChD,EAAE;AACF,iFAAiF;AACjF,wEAAwE;AACxE,gFAAgF;AAChF,wDAAwD;AACxD,EAAE;AACF,+EAA+E;AAC/E,gFAAgF;AAChF,2EAA2E;AAC3E,wCAAwC;AACxC,EAAE;AACF,+EAA+E;AAC/E,6EAA6E;AAC7E,+EAA+E;AAC/E,6EAA6E;AAyF7E,gFAAgF;AAEhF,MAAM,aAAa,GAAG;IACpB,SAAS,EAAE,kBAAkB,EAAE,IAAI,EAAE,KAAK,EAAE,sBAAsB;IAClE,UAAU,EAAE,WAAW,EAAE,KAAK,EAAE,MAAM,EAAE,KAAK;CAC9C,CAAC;AAUF,SAAS,UAAU,CAAC,KAAe;IACjC,MAAM,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAC7B,KAAK,MAAM,CAAC,IAAI,aAAa,EAAE,CAAC;QAC9B,oEAAoE;QACpE,MAAM,EAAE,GAAG,IAAI,MAAM,CAAC,gBAAgB,CAAC,CAAC,OAAO,CAAC,qBAAqB,EAAE,MAAM,CAAC,eAAe,CAAC,CAAC;QAC/F,IAAI,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC;YAAE,OAAO,CAAC,CAAC;IAC9B,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED,qFAAqF;AACrF,MAAM,UAAU,SAAS,CAAC,IAAY;IACpC,OAAO,IAAI,CAAC,WAAW,EAAE,CAAC,OAAO,CAAC,mBAAmB,EAAE,GAAG,CAAC,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;AAC1F,CAAC;AAED,SAAS,QAAQ,CAAC,IAAY,EAAE,CAAC,GAAG,CAAC;IACnC,MAAM,KAAK,GAAG,SAAS,CAAC,IAAI,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;IACzD,MAAM,GAAG,GAAG,IAAI,GAAG,EAAU,CAAC;IAC9B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,IAAI,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE;QAAE,GAAG,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;IACrF,OAAO,GAAG,CAAC;AACb,CAAC;AAED,SAAS,OAAO,CAAC,CAAc,EAAE,CAAc;IAC7C,IAAI,CAAC,CAAC,IAAI,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,KAAK,CAAC;QAAE,OAAO,CAAC,CAAC;IAC3C,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,KAAK,MAAM,CAAC,IAAI,CAAC;QAAE,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;YAAE,KAAK,EAAE,CAAC;IACzC,OAAO,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,GAAG,KAAK,CAAC,CAAC;AAC3C,CAAC;AAED;;;;;;;;;;;;;;;;;;;;GAoBG;AACH,MAAM,UAAU,eAAe,CAAC,IAAe,EAAE,SAAS,GAAG,IAAI;IAC/D,MAAM,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC;IAClE,MAAM,SAAS,GAAG,IAAI,KAAK,CAAS,IAAI,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;IAC1D,MAAM,MAAM,GAAe,EAAE,CAAC;IAC9B,MAAM,IAAI,GAAa,EAAE,CAAC;IAE1B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACrC,IAAI,SAAS,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC;YAAE,SAAS;QAClC,MAAM,EAAE,GAAG,MAAM,CAAC,MAAM,CAAC;QACzB,SAAS,CAAC,CAAC,CAAC,GAAG,EAAE,CAAC;QAClB,MAAM,OAAO,GAAG,CAAC,CAAC,CAAC,CAAC;QACpB,IAAI,IAAI,GAAG,CAAC,CAAC;QACb,KAAK,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACzC,IAAI,SAAS,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC;gBAAE,SAAS;YAClC,MAAM,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;YACpC,IAAI,CAAC,IAAI,SAAS,EAAE,CAAC;gBACnB,SAAS,CAAC,CAAC,CAAC,GAAG,EAAE,CAAC;gBAClB,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;gBAChB,IAAI,CAAC,GAAG,IAAI;oBAAE,IAAI,GAAG,CAAC,CAAC;YACzB,CAAC;QACH,CAAC;QACD,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;QACrB,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAClB,CAAC;IAED,OAAO,MAAM,CAAC,GAAG,CAAC,CAAC,OAAO,EAAE,GAAG,EAAE,EAAE;QACjC,MAAM,OAAO,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;YAClC,MAAM,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM;YACtB,QAAQ,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,KAAK;YACvB,GAAG,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG;YAChB,OAAO,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC;SACvC,CAAC,CAAC,CAAC;QACJ,MAAM,UAAU,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC;QAC7B,MAAM,UAAU,GAAG,OAAO,CAAC,MAAM,GAAG,CAAC,CAAC;QACtC,0EAA0E;QAC1E,gDAAgD;QAChD,MAAM,UAAU,GAAkC,CAAC,UAAU;YAC3D,CAAC,CAAC,KAAK;YACP,CAAC,CAAC,UAAU,IAAI,IAAI,CAAC,CAAC,CAAC,MAAM;gBAC7B,CAAC,CAAC,UAAU,IAAI,IAAI,CAAC,CAAC,CAAC,QAAQ;oBAC/B,CAAC,CAAC,YAAY,CAAC;QAEjB,OAAO;YACL,OAAO;YACP,UAAU;YACV,IAAI,EAAE,UAAU,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,KAAK,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC,CAAC;YAC3E,UAAU;YACV,UAAU,EAAE,MAAM,CAAC,UAAU,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC;SAC1C,CAAC;IACJ,CAAC,CAAC,CAAC;AACL,CAAC;AAED;;;;;;GAMG;AACH,MAAM,UAAU,iBAAiB,CAAC,IAAe;IAC/C,MAAM,QAAQ,GAAG,eAAe,CAAC,IAAI,CAAC,CAAC;IACvC,MAAM,IAAI,GAAG,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,IAAI,CAAC;IAE3F,OAAO;QACL,mBAAmB,EAAE,QAAQ,CAAC,MAAM;QACpC,yCAAyC,EAAE,IAAI,CAAC,MAAM;QACtD,qBAAqB,EAAE,IAAI;QAC3B,IAAI,EACF,IAAI,CAAC,MAAM,KAAK,CAAC;YACf,CAAC,CAAC,gJAAgJ;YAClJ,CAAC,CAAC,QAAQ,CAAC,MAAM,KAAK,CAAC,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC;gBACxC,CAAC,CAAC,OAAO,IAAI,CAAC,MAAM,mDAAmD,IAAI,CAAC,CAAC,CAAC,KAAK,IAAI,GAAG,CAAC,CAAC,CAAC,EAAE,qCAAqC,IAAI,CAAC,MAAM,0CAA0C;gBACzL,CAAC,CAAC,GAAG,QAAQ,CAAC,MAAM,0CAA0C,IAAI,CAAC,MAAM,aAAa;QAC5F,QAAQ,EAAE,QAAQ;QAClB,0EAA0E;QAC1E,6EAA6E;QAC7E,SAAS,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,CAAC,MAAM,EAAE,QAAQ,EAAE,CAAC,CAAC,KAAK,EAAE,GAAG,EAAE,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,CAAC,CAAC,GAAG,IAAI,IAAI,EAAE,CAAC,CAAC;KACtG,CAAC;AACJ,CAAC;AAED,gFAAgF;AAEhF;;;;;;;;;;;;;;;;GAgBG;AACH,MAAM,UAAU,WAAW,CAAC,KAAa,EAAE,MAAqB;IAC9D,MAAM,MAAM,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC;IAChC,IAAI,MAAM,CAAC,MAAM,GAAG,EAAE;QAAE,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,CAAC,CAAC,qCAAqC;IAEnF,KAAK,MAAM,GAAG,IAAI,MAAM,CAAC,IAAI,EAAE,CAAC;QAC9B,MAAM,GAAG,GAAG,SAAS,CAAC,GAAG,GAAG,CAAC,KAAK,IAAI,GAAG,CAAC,OAAO,EAAE,CAAC,CAAC;QACrD,IAAI,GAAG,CAAC,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAC;YACzB,OAAO,EAAE,EAAE,EAAE,IAAI,EAAE,MAAM,EAAE,GAAG,CAAC,MAAM,IAAI,SAAS,EAAE,GAAG,EAAE,GAAG,CAAC,GAAG,EAAE,CAAC;QACrE,CAAC;IACH,CAAC;IACD,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,CAAC;AACvB,CAAC"}

package/dist/papers/arxiv-client.d.ts

@@ -0,0 +1,21 @@
+import type { Paper } from './types.js';
+export declare function fetchPaperByArxivId(arxivId: string, signal?: AbortSignal): Promise<Paper | null>;
+export interface ArxivSearchOptions {
+    /** Free-text query searching title + abstract. */
+    q: string;
+    /** arXiv category filter (e.g. 'cs.AI', 'q-bio'). Optional. */
+    category?: string;
+    /** Sort order. arXiv's only useful ones are these two. */
+    sortBy?: 'relevance' | 'submittedDate';
+    perPage?: number;
+    page?: number;
+}
+export declare function searchArxiv(opts: ArxivSearchOptions, signal?: AbortSignal): Promise<Paper[]>;
+/**
+ * Deterministic full-text PDF URL. No API call needed — arXiv guarantees
+ * `https://arxiv.org/pdf/{id}.pdf` resolves for any extant arxiv_id.
+ * The `paper_fetch_full_text` tool downloads from this URL and runs the
+ * bytes through a PDF-to-text extractor.
+ */
+export declare function arxivPdfUrl(arxivId: string): string;
+//# sourceMappingURL=arxiv-client.d.ts.map

package/dist/papers/arxiv-client.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"arxiv-client.d.ts","sourceRoot":"","sources":["../../src/papers/arxiv-client.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,KAAK,EAAe,MAAM,YAAY,CAAC;AA6FrD,wBAAsB,mBAAmB,CAAC,OAAO,EAAE,MAAM,EAAE,MAAM,CAAC,EAAE,WAAW,GAAG,OAAO,CAAC,KAAK,GAAG,IAAI,CAAC,CAgBtG;AAED,MAAM,WAAW,kBAAkB;IACjC,kDAAkD;IAClD,CAAC,EAAE,MAAM,CAAC;IACV,+DAA+D;IAC/D,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,0DAA0D;IAC1D,MAAM,CAAC,EAAE,WAAW,GAAG,eAAe,CAAC;IACvC,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,IAAI,CAAC,EAAE,MAAM,CAAC;CACf;AAED,wBAAsB,WAAW,CAAC,IAAI,EAAE,kBAAkB,EAAE,MAAM,CAAC,EAAE,WAAW,GAAG,OAAO,CAAC,KAAK,EAAE,CAAC,CA0BlG;AAED;;;;;GAKG;AACH,wBAAgB,WAAW,CAAC,OAAO,EAAE,MAAM,GAAG,MAAM,CAEnD"}

package/dist/papers/arxiv-client.js

@@ -0,0 +1,130 @@
+import { arxivCategoryToDiscipline } from './discipline.js';
+import { logger } from '../core/logger.js';
+/**
+ * arXiv API client. Free, no auth, returns Atom XML.
+ *
+ * Two operations:
+ *   - `fetchPaperByArxivId(id)` — single paper detail
+ *   - `searchArxiv(query, opts)` — search by title / abstract / category
+ *
+ * The full PDF body is at `https://arxiv.org/pdf/{id}.pdf` — no API
+ * call needed; the URL is deterministic. Used by the
+ * `paper_fetch_full_text` tool when the Tutor needs the body.
+ *
+ * Courtesy: arXiv asks for ≥3s between bulk requests. For interactive
+ * use (single paper fetches) that's not needed; this client doesn't
+ * impose a delay because callers don't loop on it. Cron jobs that
+ * crawl many papers should add their own throttling.
+ */
+const ARXIV_API_BASE = 'http://export.arxiv.org/api/query';
+/**
+ * Lift an arXiv ID out of the entry's `<id>` URL. The URL is always
+ * `http://arxiv.org/abs/<id>v<version>`; we strip the version because
+ * downstream callers want the canonical version-less ID for matching.
+ */
+function arxivIdFromEntryId(entryId) {
+    const m = entryId.match(/abs\/([^v]+)(?:v\d+)?$/);
+    return m ? m[1] : entryId;
+}
+/**
+ * Tiny Atom XML parser scoped to arXiv's response shape. arXiv guarantees
+ * a stable Atom feed so we don't need a full XML parser; targeted regex
+ * extraction is reliable here and avoids pulling in xml2js / fast-xml-parser
+ * just for two operations.
+ */
+function parseArxivAtom(xml) {
+    const entries = [];
+    const entryBlocks = xml.match(/<entry>[\s\S]*?<\/entry>/g) ?? [];
+    for (const block of entryBlocks) {
+        const id = (block.match(/<id>([^<]+)<\/id>/)?.[1] ?? '').trim();
+        const title = (block.match(/<title>([\s\S]*?)<\/title>/)?.[1] ?? '').replace(/\s+/g, ' ').trim();
+        const summary = (block.match(/<summary>([\s\S]*?)<\/summary>/)?.[1] ?? '').replace(/\s+/g, ' ').trim();
+        const published = (block.match(/<published>([^<]+)<\/published>/)?.[1] ?? '').trim();
+        const primaryCategory = (block.match(/<arxiv:primary_category[^/]*term="([^"]+)"/)?.[1]
+            ?? block.match(/<category[^/]*term="([^"]+)"/)?.[1]
+            ?? '').trim();
+        const authors = [...block.matchAll(/<author>\s*<name>([^<]+)<\/name>/g)].map(m => m[1].trim());
+        const pdfMatch = block.match(/<link[^>]*type="application\/pdf"[^>]*href="([^"]+)"/);
+        entries.push({
+            id,
+            title,
+            summary,
+            authors,
+            published,
+            primaryCategory,
+            pdfUrl: pdfMatch?.[1],
+        });
+    }
+    return entries;
+}
+function entryToPaper(entry) {
+    const arxiv_id = arxivIdFromEntryId(entry.id);
+    const authors = entry.authors.map(name => ({ name }));
+    const year = entry.published ? new Date(entry.published).getFullYear() : undefined;
+    return {
+        arxiv_id,
+        title: entry.title,
+        authors,
+        abstract: entry.summary,
+        year,
+        primary_url: `https://arxiv.org/abs/${arxiv_id}`,
+        oa_pdf_url: entry.pdfUrl ?? `https://arxiv.org/pdf/${arxiv_id}.pdf`,
+        arxiv_category: entry.primaryCategory || undefined,
+        discipline: entry.primaryCategory ? arxivCategoryToDiscipline(entry.primaryCategory) : undefined,
+        source: 'arxiv',
+    };
+}
+export async function fetchPaperByArxivId(arxivId, signal) {
+    const url = `${ARXIV_API_BASE}?id_list=${encodeURIComponent(arxivId)}`;
+    try {
+        const res = await fetch(url, { signal });
+        if (!res.ok) {
+            logger.warn(`[arxiv] fetch failed ${res.status} for ${arxivId}`);
+            return null;
+        }
+        const xml = await res.text();
+        const entries = parseArxivAtom(xml);
+        if (entries.length === 0)
+            return null;
+        return entryToPaper(entries[0]);
+    }
+    catch (err) {
+        logger.warn(`[arxiv] fetch error for ${arxivId}: ${err instanceof Error ? err.message : String(err)}`);
+        return null;
+    }
+}
+export async function searchArxiv(opts, signal) {
+    const perPage = Math.min(opts.perPage ?? 20, 50);
+    const page = Math.max(opts.page ?? 0, 0);
+    const start = page * perPage;
+    const queryParts = [`all:${opts.q}`];
+    if (opts.category)
+        queryParts.push(`cat:${opts.category}`);
+    const search_query = queryParts.join('+AND+');
+    const sortBy = opts.sortBy ?? 'relevance';
+    const sortOrder = 'descending';
+    const url = `${ARXIV_API_BASE}?search_query=${encodeURIComponent(search_query)}&start=${start}&max_results=${perPage}&sortBy=${sortBy}&sortOrder=${sortOrder}`;
+    try {
+        const res = await fetch(url, { signal });
+        if (!res.ok) {
+            logger.warn(`[arxiv] search failed ${res.status} for query "${opts.q}"`);
+            return [];
+        }
+        const xml = await res.text();
+        return parseArxivAtom(xml).map(entryToPaper);
+    }
+    catch (err) {
+        logger.warn(`[arxiv] search error for "${opts.q}": ${err instanceof Error ? err.message : String(err)}`);
+        return [];
+    }
+}
+/**
+ * Deterministic full-text PDF URL. No API call needed — arXiv guarantees
+ * `https://arxiv.org/pdf/{id}.pdf` resolves for any extant arxiv_id.
+ * The `paper_fetch_full_text` tool downloads from this URL and runs the
+ * bytes through a PDF-to-text extractor.
+ */
+export function arxivPdfUrl(arxivId) {
+    return `https://arxiv.org/pdf/${arxivId}.pdf`;
+}
+//# sourceMappingURL=arxiv-client.js.map

package/dist/papers/arxiv-client.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"arxiv-client.js","sourceRoot":"","sources":["../../src/papers/arxiv-client.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,yBAAyB,EAAE,MAAM,iBAAiB,CAAC;AAC5D,OAAO,EAAE,MAAM,EAAE,MAAM,mBAAmB,CAAC;AAE3C;;;;;;;;;;;;;;;GAeG;AAEH,MAAM,cAAc,GAAG,mCAAmC,CAAC;AAY3D;;;;GAIG;AACH,SAAS,kBAAkB,CAAC,OAAe;IACzC,MAAM,CAAC,GAAG,OAAO,CAAC,KAAK,CAAC,wBAAwB,CAAC,CAAC;IAClD,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC;AAC5B,CAAC;AAED;;;;;GAKG;AACH,SAAS,cAAc,CAAC,GAAW;IACjC,MAAM,OAAO,GAAqB,EAAE,CAAC;IACrC,MAAM,WAAW,GAAG,GAAG,CAAC,KAAK,CAAC,2BAA2B,CAAC,IAAI,EAAE,CAAC;IACjE,KAAK,MAAM,KAAK,IAAI,WAAW,EAAE,CAAC;QAChC,MAAM,EAAE,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,mBAAmB,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;QAChE,MAAM,KAAK,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,4BAA4B,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;QACjG,MAAM,OAAO,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,gCAAgC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;QACvG,MAAM,SAAS,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,iCAAiC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;QACrF,MAAM,eAAe,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,4CAA4C,CAAC,EAAE,CAAC,CAAC,CAAC;eAClF,KAAK,CAAC,KAAK,CAAC,8BAA8B,CAAC,EAAE,CAAC,CAAC,CAAC;eAChD,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;QAChB,MAAM,OAAO,GAAG,CAAC,GAAG,KAAK,CAAC,QAAQ,CAAC,mCAAmC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;QAC/F,MAAM,QAAQ,GAAG,KAAK,CAAC,KAAK,CAAC,sDAAsD,CAAC,CAAC;QACrF,OAAO,CAAC,IAAI,CAAC;YACX,EAAE;YACF,KAAK;YACL,OAAO;YACP,OAAO;YACP,SAAS;YACT,eAAe;YACf,MAAM,EAAE,QAAQ,EAAE,CAAC,CAAC,CAAC;SACtB,CAAC,CAAC;IACL,CAAC;IACD,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,SAAS,YAAY,CAAC,KAAqB;IACzC,MAAM,QAAQ,GAAG,kBAAkB,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;IAC9C,MAAM,OAAO,GAAkB,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;IACrE,MAAM,IAAI,GAAG,KAAK,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,IAAI,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC;IACnF,OAAO;QACL,QAAQ;QACR,KAAK,EAAE,KAAK,CAAC,KAAK;QAClB,OAAO;QACP,QAAQ,EAAE,KAAK,CAAC,OAAO;QACvB,IAAI;QACJ,WAAW,EAAE,yBAAyB,QAAQ,EAAE;QAChD,UAAU,EAAE,KAAK,CAAC,MAAM,IAAI,yBAAyB,QAAQ,MAAM;QACnE,cAAc,EAAE,KAAK,CAAC,eAAe,IAAI,SAAS;QAClD,UAAU,EAAE,KAAK,CAAC,eAAe,CAAC,CAAC,CAAC,yBAAyB,CAAC,KAAK,CAAC,eAAe,CAAC,CAAC,CAAC,CAAC,SAAS;QAChG,MAAM,EAAE,OAAO;KAChB,CAAC;AACJ,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,mBAAmB,CAAC,OAAe,EAAE,MAAoB;IAC7E,MAAM,GAAG,GAAG,GAAG,cAAc,YAAY,kBAAkB,CAAC,OAAO,CAAC,EAAE,CAAC;IACvE,IAAI,CAAC;QACH,MAAM,GAAG,GAAG,MAAM,KAAK,CAAC,GAAG,EAAE,EAAE,MAAM,EAAE,CAAC,CAAC;QACzC,IAAI,CAAC,GAAG,CAAC,EAAE,EAAE,CAAC;YACZ,MAAM,CAAC,IAAI,CAAC,wBAAwB,GAAG,CAAC,MAAM,QAAQ,OAAO,EAAE,CAAC,CAAC;YACjE,OAAO,IAAI,CAAC;QACd,CAAC;QACD,MAAM,GAAG,GAAG,MAAM,GAAG,CAAC,IAAI,EAAE,CAAC;QAC7B,MAAM,OAAO,GAAG,cAAc,CAAC,GAAG,CAAC,CAAC;QACpC,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,IAAI,CAAC;QACtC,OAAO,YAAY,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC;IAClC,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,MAAM,CAAC,IAAI,CAAC,2BAA2B,OAAO,KAAK,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QACvG,OAAO,IAAI,CAAC;IACd,CAAC;AACH,CAAC;AAaD,MAAM,CAAC,KAAK,UAAU,WAAW,CAAC,IAAwB,EAAE,MAAoB;IAC9E,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,OAAO,IAAI,EAAE,EAAE,EAAE,CAAC,CAAC;IACjD,MAAM,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC;IACzC,MAAM,KAAK,GAAG,IAAI,GAAG,OAAO,CAAC;IAE7B,MAAM,UAAU,GAAG,CAAC,OAAO,IAAI,CAAC,CAAC,EAAE,CAAC,CAAC;IACrC,IAAI,IAAI,CAAC,QAAQ;QAAE,UAAU,CAAC,IAAI,CAAC,OAAO,IAAI,CAAC,QAAQ,EAAE,CAAC,CAAC;IAC3D,MAAM,YAAY,GAAG,UAAU,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IAE9C,MAAM,MAAM,GAAG,IAAI,CAAC,MAAM,IAAI,WAAW,CAAC;IAC1C,MAAM,SAAS,GAAG,YAAY,CAAC;IAE/B,MAAM,GAAG,GAAG,GAAG,cAAc,iBAAiB,kBAAkB,CAAC,YAAY,CAAC,UAAU,KAAK,gBAAgB,OAAO,WAAW,MAAM,cAAc,SAAS,EAAE,CAAC;IAE/J,IAAI,CAAC;QACH,MAAM,GAAG,GAAG,MAAM,KAAK,CAAC,GAAG,EAAE,EAAE,MAAM,EAAE,CAAC,CAAC;QACzC,IAAI,CAAC,GAAG,CAAC,EAAE,EAAE,CAAC;YACZ,MAAM,CAAC,IAAI,CAAC,yBAAyB,GAAG,CAAC,MAAM,eAAe,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC;YACzE,OAAO,EAAE,CAAC;QACZ,CAAC;QACD,MAAM,GAAG,GAAG,MAAM,GAAG,CAAC,IAAI,EAAE,CAAC;QAC7B,OAAO,cAAc,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC;IAC/C,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,MAAM,CAAC,IAAI,CAAC,6BAA6B,IAAI,CAAC,CAAC,MAAM,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QACzG,OAAO,EAAE,CAAC;IACZ,CAAC;AACH,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,WAAW,CAAC,OAAe;IACzC,OAAO,yBAAyB,OAAO,MAAM,CAAC;AAChD,CAAC"}

package/dist/papers/crossref-client.d.ts

@@ -0,0 +1,3 @@
+import type { Paper } from './types.js';
+export declare function fetchPaperByDoi(doi: string, signal?: AbortSignal): Promise<Paper | null>;
+//# sourceMappingURL=crossref-client.d.ts.map

package/dist/papers/crossref-client.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"crossref-client.d.ts","sourceRoot":"","sources":["../../src/papers/crossref-client.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,KAAK,EAAe,MAAM,YAAY,CAAC;AA0GrD,wBAAsB,eAAe,CAAC,GAAG,EAAE,MAAM,EAAE,MAAM,CAAC,EAAE,WAAW,GAAG,OAAO,CAAC,KAAK,GAAG,IAAI,CAAC,CAmB9F"}

package/dist/papers/crossref-client.js

@@ -0,0 +1,96 @@
+import { inferDisciplineFromText } from './discipline.js';
+import { logger } from '../core/logger.js';
+/**
+ * Crossref DOI metadata client. Free, no auth, polite-pool email
+ * recommended (same shape as OpenAlex). Used as a fallback when a
+ * DOI lookup misses on OpenAlex — Crossref is the upstream source
+ * for most journal DOIs and occasionally has metadata before
+ * OpenAlex catches up.
+ */
+const CROSSREF_API_BASE = 'https://api.crossref.org';
+function politeMailto() {
+    return (typeof process !== 'undefined' && process.env?.CROSSREF_MAILTO) || '';
+}
+/**
+ * Crossref returns abstracts wrapped in JATS XML tags. Strip them
+ * down to plain prose. Best-effort — if the cleanup fails the raw
+ * abstract still flows through, just with the tag noise visible.
+ */
+function cleanCrossrefAbstract(raw) {
+    if (!raw)
+        return undefined;
+    return raw
+        .replace(/<jats:[^>]+>/g, '')
+        .replace(/<\/jats:[^>]+>/g, '')
+        .replace(/<[^>]+>/g, '')
+        .replace(/\s+/g, ' ')
+        .trim();
+}
+function authorName(a) {
+    if (a.name)
+        return a.name;
+    if (a.given && a.family)
+        return `${a.given} ${a.family}`;
+    if (a.family)
+        return a.family;
+    if (a.given)
+        return a.given;
+    return null;
+}
+function workToPaper(work) {
+    const title = work.title?.[0]?.trim();
+    if (!title || !work.DOI)
+        return null;
+    const authors = (work.author ?? [])
+        .map(a => {
+        const name = authorName(a);
+        if (!name)
+            return null;
+        return {
+            name,
+            orcid: a.ORCID?.replace(/^https?:\/\/orcid\.org\//, ''),
+            affiliation: a.affiliation?.[0]?.name,
+        };
+    })
+        .filter((a) => a !== null);
+    const year = work.issued?.['date-parts']?.[0]?.[0]
+        ?? work.published?.['date-parts']?.[0]?.[0];
+    const pdfLink = work.link?.find(l => l['content-type'] === 'application/pdf')?.URL;
+    const abstract = cleanCrossrefAbstract(work.abstract);
+    return {
+        doi: work.DOI,
+        title,
+        authors,
+        abstract,
+        year,
+        primary_url: work.URL ?? `https://doi.org/${work.DOI}`,
+        oa_pdf_url: pdfLink,
+        discipline: inferDisciplineFromText(title, abstract),
+        source: 'crossref',
+        citation_count: work['is-referenced-by-count'],
+    };
+}
+export async function fetchPaperByDoi(doi, signal) {
+    const params = new URLSearchParams();
+    const mailto = politeMailto();
+    if (mailto)
+        params.set('mailto', mailto);
+    const qs = params.toString();
+    const url = `${CROSSREF_API_BASE}/works/${encodeURIComponent(doi)}${qs ? `?${qs}` : ''}`;
+    try {
+        const res = await fetch(url, { signal });
+        if (!res.ok) {
+            logger.warn(`[crossref] fetch failed ${res.status} for ${doi}`);
+            return null;
+        }
+        const data = await res.json();
+        if (data.status !== 'ok' || !data.message)
+            return null;
+        return workToPaper(data.message);
+    }
+    catch (err) {
+        logger.warn(`[crossref] fetch error for ${doi}: ${err instanceof Error ? err.message : String(err)}`);
+        return null;
+    }
+}
+//# sourceMappingURL=crossref-client.js.map

package/dist/papers/crossref-client.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"crossref-client.js","sourceRoot":"","sources":["../../src/papers/crossref-client.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,uBAAuB,EAAE,MAAM,iBAAiB,CAAC;AAC1D,OAAO,EAAE,MAAM,EAAE,MAAM,mBAAmB,CAAC;AAE3C;;;;;;GAMG;AAEH,MAAM,iBAAiB,GAAG,0BAA0B,CAAC;AAErD,SAAS,YAAY;IACnB,OAAO,CAAC,OAAO,OAAO,KAAK,WAAW,IAAI,OAAO,CAAC,GAAG,EAAE,eAAe,CAAC,IAAI,EAAE,CAAC;AAChF,CAAC;AA+BD;;;;GAIG;AACH,SAAS,qBAAqB,CAAC,GAAuB;IACpD,IAAI,CAAC,GAAG;QAAE,OAAO,SAAS,CAAC;IAC3B,OAAO,GAAG;SACP,OAAO,CAAC,eAAe,EAAE,EAAE,CAAC;SAC5B,OAAO,CAAC,iBAAiB,EAAE,EAAE,CAAC;SAC9B,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC;SACvB,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC;SACpB,IAAI,EAAE,CAAC;AACZ,CAAC;AAED,SAAS,UAAU,CAAC,CAAiB;IACnC,IAAI,CAAC,CAAC,IAAI;QAAE,OAAO,CAAC,CAAC,IAAI,CAAC;IAC1B,IAAI,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC,MAAM;QAAE,OAAO,GAAG,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC,MAAM,EAAE,CAAC;IACzD,IAAI,CAAC,CAAC,MAAM;QAAE,OAAO,CAAC,CAAC,MAAM,CAAC;IAC9B,IAAI,CAAC,CAAC,KAAK;QAAE,OAAO,CAAC,CAAC,KAAK,CAAC;IAC5B,OAAO,IAAI,CAAC;AACd,CAAC;AAED,SAAS,WAAW,CAAC,IAAkB;IACrC,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC;IACtC,IAAI,CAAC,KAAK,IAAI,CAAC,IAAI,CAAC,GAAG;QAAE,OAAO,IAAI,CAAC;IAErC,MAAM,OAAO,GAAkB,CAAC,IAAI,CAAC,MAAM,IAAI,EAAE,CAAC;SAC/C,GAAG,CAAC,CAAC,CAAC,EAAE;QACP,MAAM,IAAI,GAAG,UAAU,CAAC,CAAC,CAAC,CAAC;QAC3B,IAAI,CAAC,IAAI;YAAE,OAAO,IAAI,CAAC;QACvB,OAAO;YACL,IAAI;YACJ,KAAK,EAAE,CAAC,CAAC,KAAK,EAAE,OAAO,CAAC,0BAA0B,EAAE,EAAE,CAAC;YACvD,WAAW,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI;SACvB,CAAC;IACnB,CAAC,CAAC;SACD,MAAM,CAAC,CAAC,CAAC,EAAoB,EAAE,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC;IAE/C,MAAM,IAAI,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,YAAY,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;WAC7C,IAAI,CAAC,SAAS,EAAE,CAAC,YAAY,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IAE9C,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,cAAc,CAAC,KAAK,iBAAiB,CAAC,EAAE,GAAG,CAAC;IACnF,MAAM,QAAQ,GAAG,qBAAqB,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;IAEtD,OAAO;QACL,GAAG,EAAE,IAAI,CAAC,GAAG;QACb,KAAK;QACL,OAAO;QACP,QAAQ;QACR,IAAI;QACJ,WAAW,EAAE,IAAI,CAAC,GAAG,IAAI,mBAAmB,IAAI,CAAC,GAAG,EAAE;QACtD,UAAU,EAAE,OAAO;QACnB,UAAU,EAAE,uBAAuB,CAAC,KAAK,EAAE,QAAQ,CAAC;QACpD,MAAM,EAAE,UAAU;QAClB,cAAc,EAAE,IAAI,CAAC,wBAAwB,CAAC;KAC/C,CAAC;AACJ,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,eAAe,CAAC,GAAW,EAAE,MAAoB;IACrE,MAAM,MAAM,GAAG,IAAI,eAAe,EAAE,CAAC;IACrC,MAAM,MAAM,GAAG,YAAY,EAAE,CAAC;IAC9B,IAAI,MAAM;QAAE,MAAM,CAAC,GAAG,CAAC,QAAQ,EAAE,MAAM,CAAC,CAAC;IACzC,MAAM,EAAE,GAAG,MAAM,CAAC,QAAQ,EAAE,CAAC;IAC7B,MAAM,GAAG,GAAG,GAAG,iBAAiB,UAAU,kBAAkB,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC;IACzF,IAAI,CAAC;QACH,MAAM,GAAG,GAAG,MAAM,KAAK,CAAC,GAAG,EAAE,EAAE,MAAM,EAAE,CAAC,CAAC;QACzC,IAAI,CAAC,GAAG,CAAC,EAAE,EAAE,CAAC;YACZ,MAAM,CAAC,IAAI,CAAC,2BAA2B,GAAG,CAAC,MAAM,QAAQ,GAAG,EAAE,CAAC,CAAC;YAChE,OAAO,IAAI,CAAC;QACd,CAAC;QACD,MAAM,IAAI,GAAG,MAAM,GAAG,CAAC,IAAI,EAAsB,CAAC;QAClD,IAAI,IAAI,CAAC,MAAM,KAAK,IAAI,IAAI,CAAC,IAAI,CAAC,OAAO;YAAE,OAAO,IAAI,CAAC;QACvD,OAAO,WAAW,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IACnC,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,MAAM,CAAC,IAAI,CAAC,8BAA8B,GAAG,KAAK,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QACtG,OAAO,IAAI,CAAC;IACd,CAAC;AACH,CAAC"}