@mantra-ai/core 0.1.2 → 0.1.6

package/dist/analysis/block-filter.d.ts

@@ -0,0 +1,20 @@
+import type { NormalizedDoc } from "../normalization/normalized-doc-schema";
+import type { SectionCategory } from "./types";
+export interface FilteredBlock {
+    kind: "paragraph" | "figure" | "table";
+    plain_text: string;
+    label?: string;
+    order: number;
+    section_title?: string;
+    section_category: SectionCategory;
+}
+/**
+ * Returns blocks belonging to the given section categories, sorted by `order`.
+ *
+ * Blocks carry `section_id` / `path_anchor` but not category directly, so we
+ * join through the sections array. For each block we:
+ *   1. Look up `block.section_id` in the section-ID map
+ *   2. If no match, walk `block.path_anchor` ancestors to find a categorised section
+ */
+export declare function getBlocksByCategory(doc: NormalizedDoc, categories: SectionCategory[]): FilteredBlock[];
+//# sourceMappingURL=block-filter.d.ts.map

package/dist/analysis/block-filter.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"block-filter.d.ts","sourceRoot":"","sources":["../../src/analysis/block-filter.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,wCAAwC,CAAC;AAC5E,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,SAAS,CAAC;AAE/C,MAAM,WAAW,aAAa;IAC5B,IAAI,EAAE,WAAW,GAAG,QAAQ,GAAG,OAAO,CAAC;IACvC,UAAU,EAAE,MAAM,CAAC;IACnB,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,KAAK,EAAE,MAAM,CAAC;IACd,aAAa,CAAC,EAAE,MAAM,CAAC;IACvB,gBAAgB,EAAE,eAAe,CAAC;CACnC;AAED;;;;;;;GAOG;AACH,wBAAgB,mBAAmB,CACjC,GAAG,EAAE,aAAa,EAClB,UAAU,EAAE,eAAe,EAAE,GAC5B,aAAa,EAAE,CAqCjB"}

package/dist/analysis/block-filter.js

@@ -0,0 +1,59 @@
+/**
+ * Returns blocks belonging to the given section categories, sorted by `order`.
+ *
+ * Blocks carry `section_id` / `path_anchor` but not category directly, so we
+ * join through the sections array. For each block we:
+ *   1. Look up `block.section_id` in the section-ID map
+ *   2. If no match, walk `block.path_anchor` ancestors to find a categorised section
+ */
+export function getBlocksByCategory(doc, categories) {
+    const blocks = doc.blocks ?? [];
+    const sections = doc.sections ?? [];
+    if (blocks.length === 0)
+        return [];
+    const catSet = new Set(categories);
+    // Map section anchor -> { category, title }
+    const byAnchor = new Map();
+    // Map section id -> same info (id may differ from anchor for some docs)
+    const byId = new Map();
+    for (const sec of sections) {
+        if (!sec.category)
+            continue;
+        const info = { category: sec.category, title: sec.title };
+        byAnchor.set(sec.anchor, info);
+        if (sec.id)
+            byId.set(sec.id, info);
+    }
+    const matched = [];
+    for (const block of blocks) {
+        const info = resolveCategory(block.section_id, block.path_anchor, byId, byAnchor);
+        if (!info || !catSet.has(info.category))
+            continue;
+        matched.push({
+            kind: block.kind,
+            plain_text: block.plain_text ?? "",
+            label: "label" in block ? (block.label ?? undefined) : undefined,
+            order: block.order ?? 0,
+            section_title: info.title,
+            section_category: info.category,
+        });
+    }
+    matched.sort((a, b) => a.order - b.order);
+    return matched;
+}
+function resolveCategory(sectionId, pathAnchor, byId, byAnchor) {
+    // Direct ID lookup
+    const direct = byId.get(sectionId) ?? byAnchor.get(sectionId);
+    if (direct)
+        return direct;
+    // Walk path_anchor ancestors (e.g. "body/sec-methods/sec-methods-1")
+    const parts = pathAnchor.split("/");
+    for (let i = parts.length - 1; i >= 0; i--) {
+        const ancestor = parts[i];
+        const info = byAnchor.get(ancestor) ?? byId.get(ancestor);
+        if (info)
+            return info;
+    }
+    return undefined;
+}
+//# sourceMappingURL=block-filter.js.map

package/dist/analysis/block-filter.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"block-filter.js","sourceRoot":"","sources":["../../src/analysis/block-filter.ts"],"names":[],"mappings":"AAYA;;;;;;;GAOG;AACH,MAAM,UAAU,mBAAmB,CACjC,GAAkB,EAClB,UAA6B;IAE7B,MAAM,MAAM,GAAG,GAAG,CAAC,MAAM,IAAI,EAAE,CAAC;IAChC,MAAM,QAAQ,GAAG,GAAG,CAAC,QAAQ,IAAI,EAAE,CAAC;IACpC,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAEnC,MAAM,MAAM,GAAG,IAAI,GAAG,CAAS,UAAU,CAAC,CAAC;IAE3C,4CAA4C;IAC5C,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAyD,CAAC;IAClF,wEAAwE;IACxE,MAAM,IAAI,GAAG,IAAI,GAAG,EAAyD,CAAC;IAE9E,KAAK,MAAM,GAAG,IAAI,QAAQ,EAAE,CAAC;QAC3B,IAAI,CAAC,GAAG,CAAC,QAAQ;YAAE,SAAS;QAC5B,MAAM,IAAI,GAAG,EAAE,QAAQ,EAAE,GAAG,CAAC,QAA2B,EAAE,KAAK,EAAE,GAAG,CAAC,KAAK,EAAE,CAAC;QAC7E,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;QAC/B,IAAI,GAAG,CAAC,EAAE;YAAE,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,EAAE,IAAI,CAAC,CAAC;IACrC,CAAC;IAED,MAAM,OAAO,GAAoB,EAAE,CAAC;IAEpC,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;QAC3B,MAAM,IAAI,GAAG,eAAe,CAAC,KAAK,CAAC,UAAU,EAAE,KAAK,CAAC,WAAW,EAAE,IAAI,EAAE,QAAQ,CAAC,CAAC;QAClF,IAAI,CAAC,IAAI,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,QAAQ,CAAC;YAAE,SAAS;QAElD,OAAO,CAAC,IAAI,CAAC;YACX,IAAI,EAAE,KAAK,CAAC,IAAI;YAChB,UAAU,EAAE,KAAK,CAAC,UAAU,IAAI,EAAE;YAClC,KAAK,EAAE,OAAO,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,IAAI,SAAS,CAAC,CAAC,CAAC,CAAC,SAAS;YAChE,KAAK,EAAE,KAAK,CAAC,KAAK,IAAI,CAAC;YACvB,aAAa,EAAE,IAAI,CAAC,KAAK;YACzB,gBAAgB,EAAE,IAAI,CAAC,QAAQ;SAChC,CAAC,CAAC;IACL,CAAC;IAED,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC;IAC1C,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,SAAS,eAAe,CACtB,SAAiB,EACjB,UAAkB,EAClB,IAAgE,EAChE,QAAoE;IAEpE,mBAAmB;IACnB,MAAM,MAAM,GAAG,IAAI,CAAC,GAAG,CAAC,SAAS,CAAC,IAAI,QAAQ,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;IAC9D,IAAI,MAAM;QAAE,OAAO,MAAM,CAAC;IAE1B,qEAAqE;IACrE,MAAM,KAAK,GAAG,UAAU,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IACpC,KAAK,IAAI,CAAC,GAAG,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QAC3C,MAAM,QAAQ,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC;QAC3B,MAAM,IAAI,GAAG,QAAQ,CAAC,GAAG,CAAC,QAAQ,CAAC,IAAI,IAAI,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;QAC1D,IAAI,IAAI;YAAE,OAAO,IAAI,CAAC;IACxB,CAAC;IAED,OAAO,SAAS,CAAC;AACnB,CAAC"}

package/dist/analysis/extract-findings.d.ts

@@ -0,0 +1,17 @@
+import type { NormalizedDoc } from "../normalization/normalized-doc-schema";
+import type { LLMResult } from "../ai/openai/types";
+import { type PaperDigest } from "./schemas";
+export interface ExtractFindingsOptions {
+    model?: string;
+    temperature?: number;
+    maxOutputTokens?: number;
+    maxAttempts?: number;
+}
+/**
+ * Extract structured findings from a single NormalizedDoc.
+ *
+ * Filters to methods/results/discussion/limitations/conclusions blocks,
+ * falls back to all non-back-matter blocks if too few are found.
+ */
+export declare function extractFindings(doc: NormalizedDoc, opts?: ExtractFindingsOptions): Promise<LLMResult<PaperDigest>>;
+//# sourceMappingURL=extract-findings.d.ts.map

package/dist/analysis/extract-findings.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"extract-findings.d.ts","sourceRoot":"","sources":["../../src/analysis/extract-findings.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,wCAAwC,CAAC;AAC5E,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,oBAAoB,CAAC;AAGpD,OAAO,EAAqB,KAAK,WAAW,EAAE,MAAM,WAAW,CAAC;AA+BhE,MAAM,WAAW,sBAAsB;IACrC,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,eAAe,CAAC,EAAE,MAAM,CAAC;IACzB,WAAW,CAAC,EAAE,MAAM,CAAC;CACtB;AAED;;;;;GAKG;AACH,wBAAsB,eAAe,CACnC,GAAG,EAAE,aAAa,EAClB,IAAI,CAAC,EAAE,sBAAsB,GAC5B,OAAO,CAAC,SAAS,CAAC,WAAW,CAAC,CAAC,CA+CjC"}

package/dist/analysis/extract-findings.js

@@ -0,0 +1,79 @@
+import { generateWithSchema } from "../ai/openai/generate";
+import { getBlocksByCategory } from "./block-filter";
+import { PaperDigestSchema } from "./schemas";
+import { buildExtractionPrompt } from "./prompts";
+const EXTRACTION_CATEGORIES = [
+    "methods",
+    "results",
+    "results-discussion",
+    "discussion",
+    "limitations",
+    "conclusions",
+];
+const BACK_MATTER_CATEGORIES = [
+    "acknowledgements",
+    "data_availability",
+    "funding",
+    "competing_interests",
+    "author_contributions",
+    "abbreviations",
+    "glossary",
+    "publisher_note",
+    "provenance",
+    "trial_registration",
+    "code_availability",
+    "materials_availability",
+    "supplementary",
+    "appendices",
+    "notes",
+];
+/**
+ * Extract structured findings from a single NormalizedDoc.
+ *
+ * Filters to methods/results/discussion/limitations/conclusions blocks,
+ * falls back to all non-back-matter blocks if too few are found.
+ */
+export async function extractFindings(doc, opts) {
+    let blocks = getBlocksByCategory(doc, EXTRACTION_CATEGORIES);
+    // Fallback: use all blocks except back-matter when section categories are mostly "other"
+    if (blocks.length < 3) {
+        const allCategories = [
+            "abstract",
+            "introduction",
+            ...EXTRACTION_CATEGORIES,
+            "ethics",
+            "other",
+        ];
+        blocks = getBlocksByCategory(doc, allCategories);
+    }
+    // If still empty, use whatever blocks exist with a synthetic category
+    if (blocks.length === 0 && (doc.blocks?.length ?? 0) > 0) {
+        blocks = (doc.blocks ?? []).map((b, i) => ({
+            kind: b.kind,
+            plain_text: b.plain_text ?? "",
+            label: "label" in b ? (b.label ?? undefined) : undefined,
+            order: b.order ?? i,
+            section_category: "other",
+        }));
+    }
+    const metadata = {
+        title: doc.metadata.title,
+        doi: doc.metadata.ids?.doi,
+        keywords: doc.metadata.keywords,
+        abstract: doc.metadata.abstract?.paragraphs?.join(" "),
+    };
+    const messages = buildExtractionPrompt(metadata, blocks);
+    return generateWithSchema({
+        messages,
+        schema: PaperDigestSchema,
+        options: {
+            model: opts?.model ?? "gpt-4o",
+            temperature: opts?.temperature ?? 0.1,
+            maxOutputTokens: opts?.maxOutputTokens ?? 4096,
+        },
+        retry: {
+            maxAttempts: opts?.maxAttempts ?? 3,
+        },
+    });
+}
+//# sourceMappingURL=extract-findings.js.map

package/dist/analysis/extract-findings.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"extract-findings.js","sourceRoot":"","sources":["../../src/analysis/extract-findings.ts"],"names":[],"mappings":"AAEA,OAAO,EAAE,kBAAkB,EAAE,MAAM,uBAAuB,CAAC;AAC3D,OAAO,EAAE,mBAAmB,EAAE,MAAM,gBAAgB,CAAC;AACrD,OAAO,EAAE,iBAAiB,EAAoB,MAAM,WAAW,CAAC;AAChE,OAAO,EAAE,qBAAqB,EAAE,MAAM,WAAW,CAAC;AAGlD,MAAM,qBAAqB,GAAsB;IAC/C,SAAS;IACT,SAAS;IACT,oBAAoB;IACpB,YAAY;IACZ,aAAa;IACb,aAAa;CACd,CAAC;AAEF,MAAM,sBAAsB,GAAsB;IAChD,kBAAkB;IAClB,mBAAmB;IACnB,SAAS;IACT,qBAAqB;IACrB,sBAAsB;IACtB,eAAe;IACf,UAAU;IACV,gBAAgB;IAChB,YAAY;IACZ,oBAAoB;IACpB,mBAAmB;IACnB,wBAAwB;IACxB,eAAe;IACf,YAAY;IACZ,OAAO;CACR,CAAC;AASF;;;;;GAKG;AACH,MAAM,CAAC,KAAK,UAAU,eAAe,CACnC,GAAkB,EAClB,IAA6B;IAE7B,IAAI,MAAM,GAAG,mBAAmB,CAAC,GAAG,EAAE,qBAAqB,CAAC,CAAC;IAE7D,yFAAyF;IACzF,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACtB,MAAM,aAAa,GAAsB;YACvC,UAAU;YACV,cAAc;YACd,GAAG,qBAAqB;YACxB,QAAQ;YACR,OAAO;SACR,CAAC;QACF,MAAM,GAAG,mBAAmB,CAAC,GAAG,EAAE,aAAa,CAAC,CAAC;IACnD,CAAC;IAED,sEAAsE;IACtE,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,MAAM,EAAE,MAAM,IAAI,CAAC,CAAC,GAAG,CAAC,EAAE,CAAC;QACzD,MAAM,GAAG,CAAC,GAAG,CAAC,MAAM,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC;YACzC,IAAI,EAAE,CAAC,CAAC,IAAI;YACZ,UAAU,EAAE,CAAC,CAAC,UAAU,IAAI,EAAE;YAC9B,KAAK,EAAE,OAAO,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,IAAI,SAAS,CAAC,CAAC,CAAC,CAAC,SAAS;YACxD,KAAK,EAAE,CAAC,CAAC,KAAK,IAAI,CAAC;YACnB,gBAAgB,EAAE,OAA0B;SAC7C,CAAC,CAAC,CAAC;IACN,CAAC;IAED,MAAM,QAAQ,GAAG;QACf,KAAK,EAAE,GAAG,CAAC,QAAQ,CAAC,KAAK;QACzB,GAAG,EAAE,GAAG,CAAC,QAAQ,CAAC,GAAG,EAAE,GAAG;QAC1B,QAAQ,EAAE,GAAG,CAAC,QAAQ,CAAC,QAAQ;QAC/B,QAAQ,EAAE,GAAG,CAAC,QAAQ,CAAC,QAAQ,EAAE,UAAU,EAAE,IAAI,CAAC,GAAG,CAAC;KACvD,CAAC;IAEF,MAAM,QAAQ,GAAG,qBAAqB,CAAC,QAAQ,EAAE,MAAM,CAAC,CAAC;IAEzD,OAAO,kBAAkB,CAAc;QACrC,QAAQ;QACR,MAAM,EAAE,iBAAiB;QACzB,OAAO,EAAE;YACP,KAAK,EAAE,IAAI,EAAE,KAAK,IAAI,QAAQ;YAC9B,WAAW,EAAE,IAAI,EAAE,WAAW,IAAI,GAAG;YACrC,eAAe,EAAE,IAAI,EAAE,eAAe,IAAI,IAAI;SAC/C;QACD,KAAK,EAAE;YACL,WAAW,EAAE,IAAI,EAAE,WAAW,IAAI,CAAC;SACpC;KACF,CAAC,CAAC;AACL,CAAC"}

package/dist/analysis/generate-hypotheses.d.ts

@@ -0,0 +1,16 @@
+import type { LLMResult } from "../ai/openai/types";
+import { type HypothesisSet, type PaperDigest } from "./schemas";
+export interface GenerateHypothesesOptions {
+    model?: string;
+    temperature?: number;
+    maxOutputTokens?: number;
+    maxAttempts?: number;
+    minPerType?: number;
+}
+/**
+ * Synthesise hypotheses across multiple paper digests.
+ *
+ * Requires at least 2 digests for meaningful cross-paper reasoning.
+ */
+export declare function generateHypotheses(digests: PaperDigest[], opts?: GenerateHypothesesOptions): Promise<LLMResult<HypothesisSet>>;
+//# sourceMappingURL=generate-hypotheses.d.ts.map

package/dist/analysis/generate-hypotheses.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"generate-hypotheses.d.ts","sourceRoot":"","sources":["../../src/analysis/generate-hypotheses.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,oBAAoB,CAAC;AAEpD,OAAO,EAAuB,KAAK,aAAa,EAAE,KAAK,WAAW,EAAE,MAAM,WAAW,CAAC;AAGtF,MAAM,WAAW,yBAAyB;IACxC,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,eAAe,CAAC,EAAE,MAAM,CAAC;IACzB,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,UAAU,CAAC,EAAE,MAAM,CAAC;CACrB;AAED;;;;GAIG;AACH,wBAAsB,kBAAkB,CACtC,OAAO,EAAE,WAAW,EAAE,EACtB,IAAI,CAAC,EAAE,yBAAyB,GAC/B,OAAO,CAAC,SAAS,CAAC,aAAa,CAAC,CAAC,CAgBnC"}

package/dist/analysis/generate-hypotheses.js

@@ -0,0 +1,25 @@
+import { generateWithSchema } from "../ai/openai/generate";
+import { HypothesisSetSchema } from "./schemas";
+import { buildHypothesisPrompt } from "./prompts";
+/**
+ * Synthesise hypotheses across multiple paper digests.
+ *
+ * Requires at least 2 digests for meaningful cross-paper reasoning.
+ */
+export async function generateHypotheses(digests, opts) {
+    const minPerType = opts?.minPerType ?? 1;
+    const messages = buildHypothesisPrompt(digests, minPerType);
+    return generateWithSchema({
+        messages,
+        schema: HypothesisSetSchema,
+        options: {
+            model: opts?.model ?? "gpt-4o",
+            temperature: opts?.temperature ?? 0.4,
+            maxOutputTokens: opts?.maxOutputTokens ?? 8192,
+        },
+        retry: {
+            maxAttempts: opts?.maxAttempts ?? 3,
+        },
+    });
+}
+//# sourceMappingURL=generate-hypotheses.js.map

package/dist/analysis/generate-hypotheses.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"generate-hypotheses.js","sourceRoot":"","sources":["../../src/analysis/generate-hypotheses.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,kBAAkB,EAAE,MAAM,uBAAuB,CAAC;AAC3D,OAAO,EAAE,mBAAmB,EAAwC,MAAM,WAAW,CAAC;AACtF,OAAO,EAAE,qBAAqB,EAAE,MAAM,WAAW,CAAC;AAUlD;;;;GAIG;AACH,MAAM,CAAC,KAAK,UAAU,kBAAkB,CACtC,OAAsB,EACtB,IAAgC;IAEhC,MAAM,UAAU,GAAG,IAAI,EAAE,UAAU,IAAI,CAAC,CAAC;IACzC,MAAM,QAAQ,GAAG,qBAAqB,CAAC,OAAO,EAAE,UAAU,CAAC,CAAC;IAE5D,OAAO,kBAAkB,CAAgB;QACvC,QAAQ;QACR,MAAM,EAAE,mBAAmB;QAC3B,OAAO,EAAE;YACP,KAAK,EAAE,IAAI,EAAE,KAAK,IAAI,QAAQ;YAC9B,WAAW,EAAE,IAAI,EAAE,WAAW,IAAI,GAAG;YACrC,eAAe,EAAE,IAAI,EAAE,eAAe,IAAI,IAAI;SAC/C;QACD,KAAK,EAAE;YACL,WAAW,EAAE,IAAI,EAAE,WAAW,IAAI,CAAC;SACpC;KACF,CAAC,CAAC;AACL,CAAC"}

package/dist/analysis/index.d.ts

@@ -0,0 +1,11 @@
+export { getBlocksByCategory } from "./block-filter";
+export type { FilteredBlock } from "./block-filter";
+export { PaperDigestSchema, HypothesisSetSchema, } from "./schemas";
+export type { PaperDigest, Hypothesis, HypothesisSet } from "./schemas";
+export { buildExtractionPrompt, buildHypothesisPrompt } from "./prompts";
+export { extractFindings } from "./extract-findings";
+export type { ExtractFindingsOptions } from "./extract-findings";
+export { generateHypotheses } from "./generate-hypotheses";
+export type { GenerateHypothesesOptions } from "./generate-hypotheses";
+export type { SectionCategory } from "./types";
+//# sourceMappingURL=index.d.ts.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/analysis/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,mBAAmB,EAAE,MAAM,gBAAgB,CAAC;AACrD,YAAY,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAEpD,OAAO,EACL,iBAAiB,EACjB,mBAAmB,GACpB,MAAM,WAAW,CAAC;AACnB,YAAY,EAAE,WAAW,EAAE,UAAU,EAAE,aAAa,EAAE,MAAM,WAAW,CAAC;AAExE,OAAO,EAAE,qBAAqB,EAAE,qBAAqB,EAAE,MAAM,WAAW,CAAC;AAEzE,OAAO,EAAE,eAAe,EAAE,MAAM,oBAAoB,CAAC;AACrD,YAAY,EAAE,sBAAsB,EAAE,MAAM,oBAAoB,CAAC;AAEjE,OAAO,EAAE,kBAAkB,EAAE,MAAM,uBAAuB,CAAC;AAC3D,YAAY,EAAE,yBAAyB,EAAE,MAAM,uBAAuB,CAAC;AAEvE,YAAY,EAAE,eAAe,EAAE,MAAM,SAAS,CAAC"}

package/dist/analysis/index.js

@@ -0,0 +1,6 @@
+export { getBlocksByCategory } from "./block-filter";
+export { PaperDigestSchema, HypothesisSetSchema, } from "./schemas";
+export { buildExtractionPrompt, buildHypothesisPrompt } from "./prompts";
+export { extractFindings } from "./extract-findings";
+export { generateHypotheses } from "./generate-hypotheses";
+//# sourceMappingURL=index.js.map

package/dist/analysis/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/analysis/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,mBAAmB,EAAE,MAAM,gBAAgB,CAAC;AAGrD,OAAO,EACL,iBAAiB,EACjB,mBAAmB,GACpB,MAAM,WAAW,CAAC;AAGnB,OAAO,EAAE,qBAAqB,EAAE,qBAAqB,EAAE,MAAM,WAAW,CAAC;AAEzE,OAAO,EAAE,eAAe,EAAE,MAAM,oBAAoB,CAAC;AAGrD,OAAO,EAAE,kBAAkB,EAAE,MAAM,uBAAuB,CAAC"}

package/dist/analysis/prompts.d.ts

@@ -0,0 +1,13 @@
+import type { ChatMessage } from "../ai/openai/types";
+import type { FilteredBlock } from "./block-filter";
+import type { PaperDigest } from "./schemas";
+interface PaperMetadata {
+    title?: string;
+    doi?: string;
+    keywords?: string[];
+    abstract?: string;
+}
+export declare function buildExtractionPrompt(metadata: PaperMetadata, blocks: FilteredBlock[]): ChatMessage[];
+export declare function buildHypothesisPrompt(digests: PaperDigest[], minPerType?: number): ChatMessage[];
+export {};
+//# sourceMappingURL=prompts.d.ts.map

package/dist/analysis/prompts.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"prompts.d.ts","sourceRoot":"","sources":["../../src/analysis/prompts.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,oBAAoB,CAAC;AACtD,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AACpD,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,WAAW,CAAC;AAI7C,UAAU,aAAa;IACrB,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,GAAG,CAAC,EAAE,MAAM,CAAC;IACb,QAAQ,CAAC,EAAE,MAAM,EAAE,CAAC;IACpB,QAAQ,CAAC,EAAE,MAAM,CAAC;CACnB;AAED,wBAAgB,qBAAqB,CACnC,QAAQ,EAAE,aAAa,EACvB,MAAM,EAAE,aAAa,EAAE,GACtB,WAAW,EAAE,CAqEf;AAID,wBAAgB,qBAAqB,CACnC,OAAO,EAAE,WAAW,EAAE,EACtB,UAAU,SAAI,GACb,WAAW,EAAE,CAwEf"}

package/dist/analysis/prompts.js

@@ -0,0 +1,140 @@
+export function buildExtractionPrompt(metadata, blocks) {
+    const system = {
+        role: "system",
+        content: [
+            "You are a domain-agnostic scientific analyst. Your task is to extract structured findings from a research paper.",
+            "",
+            "Extract the following from the provided paper text:",
+            "",
+            "TECHNIQUES: Every experimental, computational, or analytical method used. Categorise each as assay, imaging, sequencing, computational, statistical, synthesis, purification, or other. State its purpose in one sentence.",
+            "",
+            "SYSTEMS: Every biological system, model organism, cell line, protein, pathway, dataset, or software tool studied or used. Categorise each as organism, cell-line, tissue, protein, pathway, dataset, software, or other.",
+            "",
+            "CONDITIONS: Every experimental condition, treatment, control, comparison group, or key parameter. Assign role: treatment, control, comparison, or parameter.",
+            "",
+            "FINDINGS: Every distinct claim or result. State each as a single declarative sentence. Classify evidence_type (experimental, computational, statistical, observational, theoretical, review-synthesis), strength (strong, moderate, weak, preliminary), whether it is quantitative, and list related technique names.",
+            "",
+            "LIMITATIONS: Every stated or clearly implied limitation. Classify as sample-size, generalizability, methodology, scope, temporal, resource, reproducibility, or other.",
+            "",
+            "OPEN QUESTIONS: Questions the authors raise or that are clearly implied by the text. Mark each as stated or implied.",
+            "",
+            "Rules:",
+            "- Extract faithfully from the text. Never hallucinate data, statistics, or claims not present.",
+            "- Each finding must be atomic: one claim per entry.",
+            "- Include quantitative details (p-values, effect sizes, percentages) when present in the text.",
+            "- The domain field should capture the broad scientific area (e.g. 'molecular biology', 'computational neuroscience', 'organic chemistry').",
+            "- For paper_id, use the DOI if available, otherwise use a slug from the title.",
+        ].join("\n"),
+    };
+    // Build section-grouped text
+    const sectionGroups = new Map();
+    for (const block of blocks) {
+        const key = block.section_title ?? block.section_category;
+        let group = sectionGroups.get(key);
+        if (!group) {
+            group = [];
+            sectionGroups.set(key, group);
+        }
+        if (block.kind === "figure" || block.kind === "table") {
+            const prefix = block.label ? `[${block.label}] ` : `[${block.kind}] `;
+            group.push(prefix + (block.plain_text || ""));
+        }
+        else {
+            group.push(block.plain_text);
+        }
+    }
+    const bodyParts = [];
+    for (const [section, texts] of sectionGroups) {
+        bodyParts.push(`## ${section}\n${texts.join("\n\n")}`);
+    }
+    const metaParts = [];
+    if (metadata.title)
+        metaParts.push(`Title: ${metadata.title}`);
+    if (metadata.doi)
+        metaParts.push(`DOI: ${metadata.doi}`);
+    if (metadata.keywords?.length)
+        metaParts.push(`Keywords: ${metadata.keywords.join(", ")}`);
+    if (metadata.abstract)
+        metaParts.push(`Abstract: ${metadata.abstract}`);
+    const user = {
+        role: "user",
+        content: [
+            "# Paper Metadata",
+            metaParts.join("\n"),
+            "",
+            "# Paper Content",
+            bodyParts.join("\n\n"),
+        ].join("\n"),
+    };
+    return [system, user];
+}
+// ── Hypothesis prompt (cross-paper) ─────────────────────────────────
+export function buildHypothesisPrompt(digests, minPerType = 1) {
+    const system = {
+        role: "system",
+        content: [
+            "You are a scientific hypothesis generator. Given structured digests from multiple research papers, synthesise novel, testable hypotheses.",
+            "",
+            "Generate hypotheses in three categories:",
+            "",
+            "MECHANISTIC: Propose causal mechanisms that connect findings across papers. These should explain how observed phenomena relate or predict interactions between systems/pathways studied in different papers.",
+            "",
+            "METHODOLOGICAL: Propose novel combinations of techniques from different papers that could yield new insights, or identify methodological improvements suggested by comparing approaches across papers.",
+            "",
+            "GAP: Identify knowledge gaps revealed by contradictions, unexplored intersections, or limitations shared across papers, and formulate hypotheses that address them.",
+            "",
+            `Generate at least ${minPerType} hypothesis per category (${minPerType * 3} total minimum).`,
+            "",
+            "Rules:",
+            "- Each hypothesis must cite at least 2 papers by their paper_id.",
+            "- Each hypothesis must be testable and falsifiable.",
+            "- Each hypothesis must be novel: it should go beyond what any single paper concludes.",
+            "- Do not propose experiments requiring technology that does not currently exist.",
+            "- Provide a concrete testability assessment: the experimental approach, feasibility (straightforward, moderate, challenging), and key techniques needed.",
+            "- Assign confidence (high, moderate, low) based on strength of supporting evidence.",
+            "- Use H1, H2, H3... as hypothesis IDs.",
+            "- The domain_summary should describe the shared scientific landscape across all papers in 1-2 sentences.",
+            "- Identify cross_cutting_themes: recurring concepts, methods, or systems that appear in multiple papers.",
+        ].join("\n"),
+    };
+    // Serialize digests
+    const digestTexts = digests.map((d, i) => {
+        const parts = [
+            `### Paper ${i + 1}: ${d.title}`,
+            `ID: ${d.paper_id}`,
+            `Domain: ${d.domain}`,
+        ];
+        if (d.techniques.length > 0) {
+            parts.push(`Techniques: ${d.techniques.map((t) => `${t.name} (${t.category}: ${t.purpose})`).join("; ")}`);
+        }
+        if (d.systems.length > 0) {
+            parts.push(`Systems: ${d.systems.map((s) => `${s.name} (${s.type})`).join("; ")}`);
+        }
+        if (d.conditions.length > 0) {
+            parts.push(`Conditions: ${d.conditions.map((c) => `${c.description} [${c.role}]`).join("; ")}`);
+        }
+        if (d.findings.length > 0) {
+            parts.push("Findings:");
+            for (const f of d.findings) {
+                parts.push(`  - ${f.claim} [${f.evidence_type}, ${f.strength}${f.quantitative ? ", quantitative" : ""}]`);
+            }
+        }
+        if (d.limitations.length > 0) {
+            parts.push(`Limitations: ${d.limitations.map((l) => `${l.description} (${l.type})`).join("; ")}`);
+        }
+        if (d.open_questions.length > 0) {
+            parts.push(`Open questions: ${d.open_questions.map((q) => `${q.question} [${q.source}]`).join("; ")}`);
+        }
+        return parts.join("\n");
+    });
+    const user = {
+        role: "user",
+        content: [
+            `# Paper Digests (${digests.length} papers)`,
+            "",
+            digestTexts.join("\n\n"),
+        ].join("\n"),
+    };
+    return [system, user];
+}
+//# sourceMappingURL=prompts.js.map

package/dist/analysis/prompts.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"prompts.js","sourceRoot":"","sources":["../../src/analysis/prompts.ts"],"names":[],"mappings":"AAaA,MAAM,UAAU,qBAAqB,CACnC,QAAuB,EACvB,MAAuB;IAEvB,MAAM,MAAM,GAAgB;QAC1B,IAAI,EAAE,QAAQ;QACd,OAAO,EAAE;YACP,kHAAkH;YAClH,EAAE;YACF,qDAAqD;YACrD,EAAE;YACF,4NAA4N;YAC5N,EAAE;YACF,0NAA0N;YAC1N,EAAE;YACF,8JAA8J;YAC9J,EAAE;YACF,uTAAuT;YACvT,EAAE;YACF,wKAAwK;YACxK,EAAE;YACF,sHAAsH;YACtH,EAAE;YACF,QAAQ;YACR,gGAAgG;YAChG,qDAAqD;YACrD,gGAAgG;YAChG,4IAA4I;YAC5I,gFAAgF;SACjF,CAAC,IAAI,CAAC,IAAI,CAAC;KACb,CAAC;IAEF,6BAA6B;IAC7B,MAAM,aAAa,GAAG,IAAI,GAAG,EAAoB,CAAC;IAClD,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;QAC3B,MAAM,GAAG,GAAG,KAAK,CAAC,aAAa,IAAI,KAAK,CAAC,gBAAgB,CAAC;QAC1D,IAAI,KAAK,GAAG,aAAa,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QACnC,IAAI,CAAC,KAAK,EAAE,CAAC;YACX,KAAK,GAAG,EAAE,CAAC;YACX,aAAa,CAAC,GAAG,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;QAChC,CAAC;QACD,IAAI,KAAK,CAAC,IAAI,KAAK,QAAQ,IAAI,KAAK,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;YACtD,MAAM,MAAM,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,KAAK,IAAI,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,IAAI,IAAI,CAAC;YACtE,KAAK,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,KAAK,CAAC,UAAU,IAAI,EAAE,CAAC,CAAC,CAAC;QAChD,CAAC;aAAM,CAAC;YACN,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC;QAC/B,CAAC;IACH,CAAC;IAED,MAAM,SAAS,GAAa,EAAE,CAAC;IAC/B,KAAK,MAAM,CAAC,OAAO,EAAE,KAAK,CAAC,IAAI,aAAa,EAAE,CAAC;QAC7C,SAAS,CAAC,IAAI,CAAC,MAAM,OAAO,KAAK,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC;IACzD,CAAC;IAED,MAAM,SAAS,GAAa,EAAE,CAAC;IAC/B,IAAI,QAAQ,CAAC,KAAK;QAAE,SAAS,CAAC,IAAI,CAAC,UAAU,QAAQ,CAAC,KAAK,EAAE,CAAC,CAAC;IAC/D,IAAI,QAAQ,CAAC,GAAG;QAAE,SAAS,CAAC,IAAI,CAAC,QAAQ,QAAQ,CAAC,GAAG,EAAE,CAAC,CAAC;IACzD,IAAI,QAAQ,CAAC,QAAQ,EAAE,MAAM;QAAE,SAAS,CAAC,IAAI,CAAC,aAAa,QAAQ,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IAC3F,IAAI,QAAQ,CAAC,QAAQ;QAAE,SAAS,CAAC,IAAI,CAAC,aAAa,QAAQ,CAAC,QAAQ,EAAE,CAAC,CAAC;IAExE,MAAM,IAAI,GAAgB;QACxB,IAAI,EAAE,MAAM;QACZ,OAAO,EAAE;YACP,kBAAkB;YAClB,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC;YACpB,EAAE;YACF,iBAAiB;YACjB,SAAS,CAAC,IAAI,CAAC,MAAM,CAAC;SACvB,CAAC,IAAI,CAAC,IAAI,CAAC;KACb,CAAC;IAEF,OAAO,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;AACxB,CAAC;AAED,uEAAuE;AAEvE,MAAM,UAAU,qBAAqB,CACnC,OAAsB,EACtB,UAAU,GAAG,CAAC;IAEd,MAAM,MAAM,GAAgB;QAC1B,IAAI,EAAE,QAAQ;QACd,OAAO,EAAE;YACP,2IAA2I;YAC3I,EAAE;YACF,0CAA0C;YAC1C,EAAE;YACF,8MAA8M;YAC9M,EAAE;YACF,wMAAwM;YACxM,EAAE;YACF,qKAAqK;YACrK,EAAE;YACF,qBAAqB,UAAU,6BAA6B,UAAU,GAAG,CAAC,kBAAkB;YAC5F,EAAE;YACF,QAAQ;YACR,kEAAkE;YAClE,qDAAqD;YACrD,uFAAuF;YACvF,kFAAkF;YAClF,0JAA0J;YAC1J,qFAAqF;YACrF,wCAAwC;YACxC,0GAA0G;YAC1G,0GAA0G;SAC3G,CAAC,IAAI,CAAC,IAAI,CAAC;KACb,CAAC;IAEF,oBAAoB;IACpB,MAAM,WAAW,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE;QACvC,MAAM,KAAK,GAAG;YACZ,aAAa,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,KAAK,EAAE;YAChC,OAAO,CAAC,CAAC,QAAQ,EAAE;YACnB,WAAW,CAAC,CAAC,MAAM,EAAE;SACtB,CAAC;QAEF,IAAI,CAAC,CAAC,UAAU,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAC5B,KAAK,CAAC,IAAI,CAAC,eAAe,CAAC,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,GAAG,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,QAAQ,KAAK,CAAC,CAAC,OAAO,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAC7G,CAAC;QACD,IAAI,CAAC,CAAC,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACzB,KAAK,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,GAAG,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QACrF,CAAC;QACD,IAAI,CAAC,CAAC,UAAU,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAC5B,KAAK,CAAC,IAAI,CAAC,eAAe,CAAC,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,GAAG,CAAC,CAAC,WAAW,KAAK,CAAC,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAClG,CAAC;QACD,IAAI,CAAC,CAAC,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAC1B,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC;YACxB,KAAK,MAAM,CAAC,IAAI,CAAC,CAAC,QAAQ,EAAE,CAAC;gBAC3B,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,KAAK,KAAK,CAAC,CAAC,aAAa,KAAK,CAAC,CAAC,QAAQ,GAAG,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,gBAAgB,CAAC,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;YAC5G,CAAC;QACH,CAAC;QACD,IAAI,CAAC,CAAC,WAAW,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAC7B,KAAK,CAAC,IAAI,CAAC,gBAAgB,CAAC,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,GAAG,CAAC,CAAC,WAAW,KAAK,CAAC,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QACpG,CAAC;QACD,IAAI,CAAC,CAAC,cAAc,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAChC,KAAK,CAAC,IAAI,CAAC,mBAAmB,CAAC,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,GAAG,CAAC,CAAC,QAAQ,KAAK,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QACzG,CAAC;QAED,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAC1B,CAAC,CAAC,CAAC;IAEH,MAAM,IAAI,GAAgB;QACxB,IAAI,EAAE,MAAM;QACZ,OAAO,EAAE;YACP,oBAAoB,OAAO,CAAC,MAAM,UAAU;YAC5C,EAAE;YACF,WAAW,CAAC,IAAI,CAAC,MAAM,CAAC;SACzB,CAAC,IAAI,CAAC,IAAI,CAAC;KACb,CAAC;IAEF,OAAO,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;AACxB,CAAC"}