@mantra-ai/core 0.1.0

package/dist/normalization/normalized-doc-schema.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"normalized-doc-schema.d.ts","sourceRoot":"","sources":["../../src/normalization/normalized-doc-schema.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AAExB,eAAO,MAAM,mBAAmB,iaA6BsC,CAAC;AAEvE,eAAO,MAAM,mBAAmB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;EA6Z7B,CAAC;AAEJ,MAAM,MAAM,aAAa,GAAG,CAAC,CAAC,KAAK,CAAC,OAAO,mBAAmB,CAAC,CAAC"}

package/dist/normalization/normalized-doc-schema.js

@@ -0,0 +1,410 @@
+import { z } from "zod";
+export const SectionCategoryEnum = z
+    .enum([
+    "abstract",
+    "introduction",
+    "methods",
+    "results",
+    "discussion",
+    "conclusions",
+    "results-discussion",
+    "limitations",
+    "ethics",
+    "acknowledgements",
+    // Annex and back-matter categories
+    "data_availability",
+    "funding",
+    "competing_interests",
+    "author_contributions",
+    "abbreviations",
+    "glossary",
+    "publisher_note",
+    "provenance",
+    "trial_registration",
+    "code_availability",
+    "materials_availability",
+    "supplementary",
+    "appendices",
+    "notes",
+    "other",
+])
+    .describe("Canonical section categories in the normalized document");
+export const NormalizedDocSchema = z
+    .object({
+    schemaVersion: z
+        .literal("v1")
+        .optional()
+        .describe("Normalized schema version identifier"),
+    metadata: z
+        .object({
+        ids: z
+            .object({
+            pmcid: z.string().optional(),
+            doi: z.string().optional(),
+            publisher_id: z.string().optional(),
+            art_access_id: z.string().optional(),
+        })
+            .partial()
+            .optional(),
+        title: z.string().optional(),
+        journal: z
+            .object({
+            title: z.string().optional(),
+            ids: z
+                .array(z
+                .object({
+                type: z.string().optional(),
+                value: z.string(),
+            })
+                .strict())
+                .optional(),
+            issn: z
+                .array(z
+                .object({
+                pub_type: z.string().optional(),
+                value: z.string(),
+            })
+                .strict())
+                .optional(),
+            publisher: z
+                .object({
+                name: z.string().optional(),
+                location: z.string().optional(),
+            })
+                .strict()
+                .optional(),
+        })
+            .strict()
+            .optional(),
+        pub_dates: z
+            .array(z
+            .object({
+            pub_type: z.string().optional(),
+            day: z.string().optional(),
+            month: z.string().optional(),
+            season: z.string().optional(),
+            year: z.string().optional(),
+        })
+            .strict())
+            .optional(),
+        volume: z.string().optional(),
+        issue: z.string().optional(),
+        elocation_id: z.string().optional(),
+        license: z
+            .object({
+            type: z.string().optional(),
+            href: z.string().optional(),
+            text: z.string().optional(),
+        })
+            .strict()
+            .optional(),
+        self_uri: z.string().optional(),
+        abstract: z
+            .object({
+            paragraphs: z.array(z.string()).optional(),
+            xml: z.array(z.string()).optional(),
+            xml_minimal: z.array(z.string()).optional(),
+            title: z.string().optional(),
+        })
+            .strict()
+            .optional(),
+        contributors: z
+            .array(z
+            .object({
+            role: z.string().optional(),
+            given: z.string().optional(),
+            surname: z.string().optional(),
+            full: z.string().optional(),
+            xref_aff: z.array(z.string()).optional(),
+            corresp: z.boolean().optional(),
+            corresp_id: z.string().optional(),
+            email: z.string().optional(),
+        })
+            .strict())
+            .optional(),
+        affiliations: z
+            .array(z
+            .object({
+            id: z.string().optional(),
+            label: z.string().optional(),
+            text: z.string().optional(),
+            country: z.string().optional(),
+        })
+            .strict())
+            .optional(),
+        keywords: z.array(z.string()).optional(),
+    })
+        .strict()
+        .describe("Front-matter metadata (ids, journal, contributors, etc.)"),
+    blocks: z
+        .array(z.discriminatedUnion("kind", [
+        z
+            .object({
+            kind: z.literal("paragraph"),
+            order: z.number().int().min(0).optional(),
+            index_in_section: z.number().int().min(0).optional(),
+            block_id: z.string().optional(),
+            plain_text: z
+                .string()
+                .describe("Plain text content of the paragraph"),
+            raw_xml: z
+                .string()
+                .optional()
+                .describe("Raw XML of the paragraph, when available"),
+            minimal_xml: z
+                .string()
+                .optional()
+                .describe("Minimized XML of the paragraph, when available"),
+            xrefs: z
+                .array(z
+                .object({
+                type: z
+                    .enum([
+                    "figure-ref",
+                    "table-ref",
+                    "citation",
+                    "supplementary-ref",
+                    "extlink",
+                ])
+                    .describe("Cross-reference type"),
+                target_id: z
+                    .string()
+                    .optional()
+                    .describe("Target identifier (rid/target/href)"),
+                label: z
+                    .string()
+                    .optional()
+                    .describe("Displayed label text"),
+                span: z
+                    .tuple([
+                    z.number().int().min(0),
+                    z.number().int().min(0),
+                ])
+                    .optional()
+                    .describe("[start,end) offsets in plain_text"),
+            })
+                .strict())
+                .optional()
+                .describe("Inline cross-references extracted from minimal XML"),
+            section_id: z
+                .string()
+                .describe("Innermost section id for the block"),
+            path_anchor: z
+                .string()
+                .describe("Canonical path of anchors from root to leaf"),
+        })
+            .strict(),
+        z
+            .object({
+            kind: z.literal("figure"),
+            order: z.number().int().min(0).optional(),
+            index_in_section: z.number().int().min(0).optional(),
+            block_id: z.string().optional(),
+            id: z.string().optional(),
+            label: z.string().optional(),
+            title: z.string().optional(),
+            slug: z.string().optional(),
+            plain_text: z
+                .string()
+                .optional()
+                .describe("Figure caption as plain text"),
+            raw_xml: z.string().optional().describe("Figure caption raw XML"),
+            minimal_xml: z
+                .string()
+                .optional()
+                .describe("Figure caption minimal XML"),
+            xrefs: z
+                .array(z
+                .object({
+                type: z.enum([
+                    "figure-ref",
+                    "table-ref",
+                    "citation",
+                    "supplementary-ref",
+                    "extlink",
+                ]),
+                target_id: z.string().optional(),
+                label: z.string().optional(),
+                span: z
+                    .tuple([
+                    z.number().int().min(0),
+                    z.number().int().min(0),
+                ])
+                    .optional(),
+            })
+                .strict())
+                .optional(),
+            link: z
+                .object({
+                xml: z
+                    .string()
+                    .url()
+                    .optional()
+                    .describe("Full XML URL for the article"),
+                image_large_jpg: z
+                    .string()
+                    .url()
+                    .optional()
+                    .describe("Large JPG rendition for this figure, e.g., F1.large.jpg"),
+            })
+                .strict()
+                .optional()
+                .describe("Useful links related to this figure"),
+            section_id: z
+                .string()
+                .describe("Innermost section id for the block"),
+            path_anchor: z
+                .string()
+                .describe("Canonical path of anchors from root to leaf"),
+        })
+            .strict(),
+        z
+            .object({
+            kind: z.literal("table"),
+            order: z.number().int().min(0).optional(),
+            index_in_section: z.number().int().min(0).optional(),
+            block_id: z.string().optional(),
+            id: z.string().optional(),
+            label: z.string().optional(),
+            title: z.string().optional(),
+            slug: z.string().optional(),
+            plain_text: z
+                .string()
+                .optional()
+                .describe("Table caption as plain text"),
+            raw_xml: z.string().optional().describe("Table caption raw XML"),
+            minimal_xml: z
+                .string()
+                .optional()
+                .describe("Table caption minimal XML"),
+            xrefs: z
+                .array(z
+                .object({
+                type: z.enum([
+                    "figure-ref",
+                    "table-ref",
+                    "citation",
+                    "supplementary-ref",
+                    "extlink",
+                ]),
+                target_id: z.string().optional(),
+                label: z.string().optional(),
+                span: z
+                    .tuple([
+                    z.number().int().min(0),
+                    z.number().int().min(0),
+                ])
+                    .optional(),
+            })
+                .strict())
+                .optional(),
+            link: z
+                .object({
+                xml: z
+                    .string()
+                    .url()
+                    .optional()
+                    .describe("Full XML URL for the article"),
+                image_medium_gif: z
+                    .string()
+                    .url()
+                    .optional()
+                    .describe("Medium GIF rendition for this table, e.g., T1.medium.gif"),
+            })
+                .strict()
+                .optional()
+                .describe("Useful links related to this table"),
+            section_id: z
+                .string()
+                .describe("Innermost section id for the block"),
+            path_anchor: z
+                .string()
+                .describe("Canonical path of anchors from root to leaf"),
+        })
+            .strict(),
+    ]))
+        .optional()
+        .describe("Unified stream of all blocks (body + back-matter)"),
+    sections: z
+        .array(z
+        .object({
+        id: z.string().optional(),
+        title: z.string().optional(),
+        normalized_title: z.string().optional(),
+        category: SectionCategoryEnum.optional(),
+        level: z.number().int().min(1),
+        order: z.number().int().min(0),
+        anchor: z.string(),
+        path_anchor: z.string().optional(),
+        path_anchors: z.array(z.string()).optional(),
+        top_normalized_title: z.string().optional(),
+    })
+        .strict())
+        .optional()
+        .describe("Section metadata present in the document (no text content)"),
+    citations: z
+        .array(z
+        .object({
+        id: z.string().optional(),
+        label: z.string().optional(),
+        raw: z.string().optional(),
+        text: z.string().optional(),
+        xml: z.string().optional(),
+        xml_minimal: z.string().optional(),
+        articleTitle: z.string().optional(),
+        sourceTitle: z.string().optional(),
+        year: z.number().int().optional(),
+        volume: z.string().optional(),
+        issue: z.string().optional(),
+        fpage: z.string().optional(),
+        lpage: z.string().optional(),
+        doi: z.string().optional(),
+        pmid: z.string().optional(),
+        pmcid: z.string().optional(),
+        authors: z.array(z.unknown()).optional(),
+        pubType: z.string().optional(),
+    })
+        .strict())
+        .optional()
+        .describe("Reference list entries (citations)"),
+    assets: z
+        .object({
+        figures: z
+            .array(z
+            .object({
+            id: z.string().optional(),
+            label: z.string().optional(),
+            title: z.string().optional(),
+            caption_text: z.string().optional(),
+            hash: z.string().regex(/^sha256:/),
+        })
+            .strict())
+            .optional(),
+        tables: z
+            .array(z
+            .object({
+            id: z.string().optional(),
+            label: z.string().optional(),
+            title: z.string().optional(),
+            caption_text: z.string().optional(),
+            hash: z.string().regex(/^sha256:/),
+        })
+            .strict())
+            .optional(),
+        supplements: z
+            .array(z
+            .object({
+            href: z.string().optional(),
+            type: z.string().optional(),
+            hash: z.string().regex(/^sha256:/),
+        })
+            .strict())
+            .optional(),
+    })
+        .strict()
+        .optional()
+        .describe("Consolidated assets: figures, tables, and supplementary files"),
+})
+    .strict()
+    .describe("Normalized document: body + selected back-matter; front-matter in metadata");
+//# sourceMappingURL=normalized-doc-schema.js.map

package/dist/normalization/normalized-doc-schema.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"normalized-doc-schema.js","sourceRoot":"","sources":["../../src/normalization/normalized-doc-schema.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AAExB,MAAM,CAAC,MAAM,mBAAmB,GAAG,CAAC;KACjC,IAAI,CAAC;IACJ,UAAU;IACV,cAAc;IACd,SAAS;IACT,SAAS;IACT,YAAY;IACZ,aAAa;IACb,oBAAoB;IACpB,aAAa;IACb,QAAQ;IACR,kBAAkB;IAClB,mCAAmC;IACnC,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;IACP,OAAO;CACR,CAAC;KACD,QAAQ,CAAC,yDAAyD,CAAC,CAAC;AAEvE,MAAM,CAAC,MAAM,mBAAmB,GAAG,CAAC;KACjC,MAAM,CAAC;IACN,aAAa,EAAE,CAAC;SACb,OAAO,CAAC,IAAI,CAAC;SACb,QAAQ,EAAE;SACV,QAAQ,CAAC,sCAAsC,CAAC;IACnD,QAAQ,EAAE,CAAC;SACR,MAAM,CAAC;QACN,GAAG,EAAE,CAAC;aACH,MAAM,CAAC;YACN,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC5B,GAAG,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC1B,YAAY,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YACnC,aAAa,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;SACrC,CAAC;aACD,OAAO,EAAE;aACT,QAAQ,EAAE;QACb,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QAC5B,OAAO,EAAE,CAAC;aACP,MAAM,CAAC;YACN,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC5B,GAAG,EAAE,CAAC;iBACH,KAAK,CACJ,CAAC;iBACE,MAAM,CAAC;gBACN,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;gBAC3B,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE;aAClB,CAAC;iBACD,MAAM,EAAE,CACZ;iBACA,QAAQ,EAAE;YACb,IAAI,EAAE,CAAC;iBACJ,KAAK,CACJ,CAAC;iBACE,MAAM,CAAC;gBACN,QAAQ,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;gBAC/B,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE;aAClB,CAAC;iBACD,MAAM,EAAE,CACZ;iBACA,QAAQ,EAAE;YACb,SAAS,EAAE,CAAC;iBACT,MAAM,CAAC;gBACN,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;gBAC3B,QAAQ,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;aAChC,CAAC;iBACD,MAAM,EAAE;iBACR,QAAQ,EAAE;SACd,CAAC;aACD,MAAM,EAAE;aACR,QAAQ,EAAE;QACb,SAAS,EAAE,CAAC;aACT,KAAK,CACJ,CAAC;aACE,MAAM,CAAC;YACN,QAAQ,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC/B,GAAG,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC1B,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC5B,MAAM,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC7B,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;SAC5B,CAAC;aACD,MAAM,EAAE,CACZ;aACA,QAAQ,EAAE;QACb,MAAM,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QAC7B,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QAC5B,YAAY,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QACnC,OAAO,EAAE,CAAC;aACP,MAAM,CAAC;YACN,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC3B,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC3B,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;SAC5B,CAAC;aACD,MAAM,EAAE;aACR,QAAQ,EAAE;QACb,QAAQ,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QAC/B,QAAQ,EAAE,CAAC;aACR,MAAM,CAAC;YACN,UAAU,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC,QAAQ,EAAE;YAC1C,GAAG,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC,QAAQ,EAAE;YACnC,WAAW,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC,QAAQ,EAAE;YAC3C,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;SAC7B,CAAC;aACD,MAAM,EAAE;aACR,QAAQ,EAAE;QACb,YAAY,EAAE,CAAC;aACZ,KAAK,CACJ,CAAC;aACE,MAAM,CAAC;YACN,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC3B,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC5B,OAAO,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC9B,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC3B,QAAQ,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC,QAAQ,EAAE;YACxC,OAAO,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC,QAAQ,EAAE;YAC/B,UAAU,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YACjC,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;SAC7B,CAAC;aACD,MAAM,EAAE,CACZ;aACA,QAAQ,EAAE;QACb,YAAY,EAAE,CAAC;aACZ,KAAK,CACJ,CAAC;aACE,MAAM,CAAC;YACN,EAAE,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YACzB,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC5B,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC3B,OAAO,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;SAC/B,CAAC;aACD,MAAM,EAAE,CACZ;aACA,QAAQ,EAAE;QACb,QAAQ,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC,QAAQ,EAAE;KACzC,CAAC;SACD,MAAM,EAAE;SACR,QAAQ,CAAC,0DAA0D,CAAC;IACvE,MAAM,EAAE,CAAC;SACN,KAAK,CACJ,CAAC,CAAC,kBAAkB,CAAC,MAAM,EAAE;QAC3B,CAAC;aACE,MAAM,CAAC;YACN,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,WAAW,CAAC;YAC5B,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,EAAE;YACzC,gBAAgB,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,EAAE;YACpD,QAAQ,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC/B,UAAU,EAAE,CAAC;iBACV,MAAM,EAAE;iBACR,QAAQ,CAAC,qCAAqC,CAAC;YAClD,OAAO,EAAE,CAAC;iBACP,MAAM,EAAE;iBACR,QAAQ,EAAE;iBACV,QAAQ,CAAC,0CAA0C,CAAC;YACvD,WAAW,EAAE,CAAC;iBACX,MAAM,EAAE;iBACR,QAAQ,EAAE;iBACV,QAAQ,CAAC,gDAAgD,CAAC;YAC7D,KAAK,EAAE,CAAC;iBACL,KAAK,CACJ,CAAC;iBACE,MAAM,CAAC;gBACN,IAAI,EAAE,CAAC;qBACJ,IAAI,CAAC;oBACJ,YAAY;oBACZ,WAAW;oBACX,UAAU;oBACV,mBAAmB;oBACnB,SAAS;iBACV,CAAC;qBACD,QAAQ,CAAC,sBAAsB,CAAC;gBACnC,SAAS,EAAE,CAAC;qBACT,MAAM,EAAE;qBACR,QAAQ,EAAE;qBACV,QAAQ,CAAC,qCAAqC,CAAC;gBAClD,KAAK,EAAE,CAAC;qBACL,MAAM,EAAE;qBACR,QAAQ,EAAE;qBACV,QAAQ,CAAC,sBAAsB,CAAC;gBACnC,IAAI,EAAE,CAAC;qBACJ,KAAK,CAAC;oBACL,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;oBACvB,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;iBACxB,CAAC;qBACD,QAAQ,EAAE;qBACV,QAAQ,CAAC,mCAAmC,CAAC;aACjD,CAAC;iBACD,MAAM,EAAE,CACZ;iBACA,QAAQ,EAAE;iBACV,QAAQ,CAAC,oDAAoD,CAAC;YACjE,UAAU,EAAE,CAAC;iBACV,MAAM,EAAE;iBACR,QAAQ,CAAC,oCAAoC,CAAC;YACjD,WAAW,EAAE,CAAC;iBACX,MAAM,EAAE;iBACR,QAAQ,CAAC,6CAA6C,CAAC;SAC3D,CAAC;aACD,MAAM,EAAE;QACX,CAAC;aACE,MAAM,CAAC;YACN,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,QAAQ,CAAC;YACzB,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,EAAE;YACzC,gBAAgB,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,EAAE;YACpD,QAAQ,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC/B,EAAE,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YACzB,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC5B,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC5B,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC3B,UAAU,EAAE,CAAC;iBACV,MAAM,EAAE;iBACR,QAAQ,EAAE;iBACV,QAAQ,CAAC,8BAA8B,CAAC;YAC3C,OAAO,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE,CAAC,QAAQ,CAAC,wBAAwB,CAAC;YACjE,WAAW,EAAE,CAAC;iBACX,MAAM,EAAE;iBACR,QAAQ,EAAE;iBACV,QAAQ,CAAC,4BAA4B,CAAC;YACzC,KAAK,EAAE,CAAC;iBACL,KAAK,CACJ,CAAC;iBACE,MAAM,CAAC;gBACN,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC;oBACX,YAAY;oBACZ,WAAW;oBACX,UAAU;oBACV,mBAAmB;oBACnB,SAAS;iBACV,CAAC;gBACF,SAAS,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;gBAChC,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;gBAC5B,IAAI,EAAE,CAAC;qBACJ,KAAK,CAAC;oBACL,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;oBACvB,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;iBACxB,CAAC;qBACD,QAAQ,EAAE;aACd,CAAC;iBACD,MAAM,EAAE,CACZ;iBACA,QAAQ,EAAE;YACb,IAAI,EAAE,CAAC;iBACJ,MAAM,CAAC;gBACN,GAAG,EAAE,CAAC;qBACH,MAAM,EAAE;qBACR,GAAG,EAAE;qBACL,QAAQ,EAAE;qBACV,QAAQ,CAAC,8BAA8B,CAAC;gBAC3C,eAAe,EAAE,CAAC;qBACf,MAAM,EAAE;qBACR,GAAG,EAAE;qBACL,QAAQ,EAAE;qBACV,QAAQ,CACP,yDAAyD,CAC1D;aACJ,CAAC;iBACD,MAAM,EAAE;iBACR,QAAQ,EAAE;iBACV,QAAQ,CAAC,qCAAqC,CAAC;YAClD,UAAU,EAAE,CAAC;iBACV,MAAM,EAAE;iBACR,QAAQ,CAAC,oCAAoC,CAAC;YACjD,WAAW,EAAE,CAAC;iBACX,MAAM,EAAE;iBACR,QAAQ,CAAC,6CAA6C,CAAC;SAC3D,CAAC;aACD,MAAM,EAAE;QACX,CAAC;aACE,MAAM,CAAC;YACN,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,OAAO,CAAC;YACxB,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,EAAE;YACzC,gBAAgB,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,EAAE;YACpD,QAAQ,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC/B,EAAE,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YACzB,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC5B,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC5B,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC3B,UAAU,EAAE,CAAC;iBACV,MAAM,EAAE;iBACR,QAAQ,EAAE;iBACV,QAAQ,CAAC,6BAA6B,CAAC;YAC1C,OAAO,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE,CAAC,QAAQ,CAAC,uBAAuB,CAAC;YAChE,WAAW,EAAE,CAAC;iBACX,MAAM,EAAE;iBACR,QAAQ,EAAE;iBACV,QAAQ,CAAC,2BAA2B,CAAC;YACxC,KAAK,EAAE,CAAC;iBACL,KAAK,CACJ,CAAC;iBACE,MAAM,CAAC;gBACN,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC;oBACX,YAAY;oBACZ,WAAW;oBACX,UAAU;oBACV,mBAAmB;oBACnB,SAAS;iBACV,CAAC;gBACF,SAAS,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;gBAChC,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;gBAC5B,IAAI,EAAE,CAAC;qBACJ,KAAK,CAAC;oBACL,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;oBACvB,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;iBACxB,CAAC;qBACD,QAAQ,EAAE;aACd,CAAC;iBACD,MAAM,EAAE,CACZ;iBACA,QAAQ,EAAE;YACb,IAAI,EAAE,CAAC;iBACJ,MAAM,CAAC;gBACN,GAAG,EAAE,CAAC;qBACH,MAAM,EAAE;qBACR,GAAG,EAAE;qBACL,QAAQ,EAAE;qBACV,QAAQ,CAAC,8BAA8B,CAAC;gBAC3C,gBAAgB,EAAE,CAAC;qBAChB,MAAM,EAAE;qBACR,GAAG,EAAE;qBACL,QAAQ,EAAE;qBACV,QAAQ,CACP,0DAA0D,CAC3D;aACJ,CAAC;iBACD,MAAM,EAAE;iBACR,QAAQ,EAAE;iBACV,QAAQ,CAAC,oCAAoC,CAAC;YACjD,UAAU,EAAE,CAAC;iBACV,MAAM,EAAE;iBACR,QAAQ,CAAC,oCAAoC,CAAC;YACjD,WAAW,EAAE,CAAC;iBACX,MAAM,EAAE;iBACR,QAAQ,CAAC,6CAA6C,CAAC;SAC3D,CAAC;aACD,MAAM,EAAE;KACZ,CAAC,CACH;SACA,QAAQ,EAAE;SACV,QAAQ,CAAC,mDAAmD,CAAC;IAChE,QAAQ,EAAE,CAAC;SACR,KAAK,CACJ,CAAC;SACE,MAAM,CAAC;QACN,EAAE,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QACzB,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QAC5B,gBAAgB,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QACvC,QAAQ,EAAE,mBAAmB,CAAC,QAAQ,EAAE;QACxC,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;QAC9B,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;QAC9B,MAAM,EAAE,CAAC,CAAC,MAAM,EAAE;QAClB,WAAW,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QAClC,YAAY,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC,QAAQ,EAAE;QAC5C,oBAAoB,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;KAC5C,CAAC;SACD,MAAM,EAAE,CACZ;SACA,QAAQ,EAAE;SACV,QAAQ,CAAC,4DAA4D,CAAC;IACzE,SAAS,EAAE,CAAC;SACT,KAAK,CACJ,CAAC;SACE,MAAM,CAAC;QACN,EAAE,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QACzB,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QAC5B,GAAG,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QAC1B,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QAC3B,GAAG,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QAC1B,WAAW,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QAClC,YAAY,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QACnC,WAAW,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QAClC,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,QAAQ,EAAE;QACjC,MAAM,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QAC7B,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QAC5B,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QAC5B,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QAC5B,GAAG,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QAC1B,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QAC3B,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;QAC5B,OAAO,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC,CAAC,QAAQ,EAAE;QACxC,OAAO,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;KAC/B,CAAC;SACD,MAAM,EAAE,CACZ;SACA,QAAQ,EAAE;SACV,QAAQ,CAAC,oCAAoC,CAAC;IACjD,MAAM,EAAE,CAAC;SACN,MAAM,CAAC;QACN,OAAO,EAAE,CAAC;aACP,KAAK,CACJ,CAAC;aACE,MAAM,CAAC;YACN,EAAE,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YACzB,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC5B,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC5B,YAAY,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YACnC,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,KAAK,CAAC,UAAU,CAAC;SACnC,CAAC;aACD,MAAM,EAAE,CACZ;aACA,QAAQ,EAAE;QACb,MAAM,EAAE,CAAC;aACN,KAAK,CACJ,CAAC;aACE,MAAM,CAAC;YACN,EAAE,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YACzB,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC5B,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC5B,YAAY,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YACnC,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,KAAK,CAAC,UAAU,CAAC;SACnC,CAAC;aACD,MAAM,EAAE,CACZ;aACA,QAAQ,EAAE;QACb,WAAW,EAAE,CAAC;aACX,KAAK,CACJ,CAAC;aACE,MAAM,CAAC;YACN,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC3B,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;YAC3B,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,KAAK,CAAC,UAAU,CAAC;SACnC,CAAC;aACD,MAAM,EAAE,CACZ;aACA,QAAQ,EAAE;KACd,CAAC;SACD,MAAM,EAAE;SACR,QAAQ,EAAE;SACV,QAAQ,CACP,+DAA+D,CAChE;CACJ,CAAC;KACD,MAAM,EAAE;KACR,QAAQ,CACP,4EAA4E,CAC7E,CAAC"}

package/dist/normalization/pdf/index.d.ts

@@ -0,0 +1,4 @@
+export { normalizePdfToDoc, buildPdfRawEnvelope } from "./normalize";
+export type { NormalizePdfInput, NormalizedPdf } from "./normalize";
+export { PDF_EXTRACTION_PROMPT, PDF_EXTRACTION_SYSTEM } from "./prompt";
+//# sourceMappingURL=index.d.ts.map

package/dist/normalization/pdf/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../src/normalization/pdf/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,iBAAiB,EAAE,mBAAmB,EAAE,MAAM,aAAa,CAAC;AACrE,YAAY,EAAE,iBAAiB,EAAE,aAAa,EAAE,MAAM,aAAa,CAAC;AACpE,OAAO,EAAE,qBAAqB,EAAE,qBAAqB,EAAE,MAAM,UAAU,CAAC"}

package/dist/normalization/pdf/index.js

@@ -0,0 +1,3 @@
+export { normalizePdfToDoc, buildPdfRawEnvelope } from "./normalize";
+export { PDF_EXTRACTION_PROMPT, PDF_EXTRACTION_SYSTEM } from "./prompt";
+//# sourceMappingURL=index.js.map

package/dist/normalization/pdf/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../../src/normalization/pdf/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,iBAAiB,EAAE,mBAAmB,EAAE,MAAM,aAAa,CAAC;AAErE,OAAO,EAAE,qBAAqB,EAAE,qBAAqB,EAAE,MAAM,UAAU,CAAC"}

package/dist/normalization/pdf/normalize.d.ts

@@ -0,0 +1,31 @@
+import { type NormalizedDoc } from "../normalized-doc-schema";
+import type { NormalizedRawEnvelope } from "../jats/utils/types";
+import type { CompleteIds, LoggerLike } from "../../works/types";
+export type NormalizePdfInput = {
+    pdfBuffer: Buffer;
+    completeIds: CompleteIds;
+    logger?: LoggerLike;
+    cid?: string;
+};
+export type NormalizedPdf = {
+    isPdf: true;
+    versionLabel?: string;
+    normalizedJson: NormalizedDoc;
+    metadata: NormalizedDoc["metadata"] | null;
+    schemaVersion: string | null;
+    title: string | undefined;
+    abstractText?: string;
+    licenseFromMeta?: string | null;
+    license?: string | null;
+    source: "pdf-gemini";
+    sourceUrl?: string | null;
+    rawXml: null;
+    pdfSha256: string;
+    blocks?: NormalizedDoc["blocks"];
+    sections?: NormalizedDoc["sections"];
+    citations?: NormalizedDoc["citations"];
+    assets?: NormalizedDoc["assets"];
+};
+export declare function normalizePdfToDoc(input: NormalizePdfInput): Promise<NormalizedPdf>;
+export declare function buildPdfRawEnvelope(pdfBuffer: Buffer, pdfSha256: string): NormalizedRawEnvelope;
+//# sourceMappingURL=normalize.d.ts.map

package/dist/normalization/pdf/normalize.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"normalize.d.ts","sourceRoot":"","sources":["../../../src/normalization/pdf/normalize.ts"],"names":[],"mappings":"AAGA,OAAO,EAEL,KAAK,aAAa,EACnB,MAAM,0BAA0B,CAAC;AAClC,OAAO,KAAK,EAAE,qBAAqB,EAAE,MAAM,qBAAqB,CAAC;AACjE,OAAO,KAAK,EAAE,WAAW,EAAE,UAAU,EAAE,MAAM,mBAAmB,CAAC;AASjE,MAAM,MAAM,iBAAiB,GAAG;IAC9B,SAAS,EAAE,MAAM,CAAC;IAClB,WAAW,EAAE,WAAW,CAAC;IACzB,MAAM,CAAC,EAAE,UAAU,CAAC;IACpB,GAAG,CAAC,EAAE,MAAM,CAAC;CACd,CAAC;AAEF,MAAM,MAAM,aAAa,GAAG;IAC1B,KAAK,EAAE,IAAI,CAAC;IACZ,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,cAAc,EAAE,aAAa,CAAC;IAC9B,QAAQ,EAAE,aAAa,CAAC,UAAU,CAAC,GAAG,IAAI,CAAC;IAC3C,aAAa,EAAE,MAAM,GAAG,IAAI,CAAC;IAC7B,KAAK,EAAE,MAAM,GAAG,SAAS,CAAC;IAC1B,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,eAAe,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IAChC,OAAO,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IACxB,MAAM,EAAE,YAAY,CAAC;IACrB,SAAS,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IAC1B,MAAM,EAAE,IAAI,CAAC;IACb,SAAS,EAAE,MAAM,CAAC;IAClB,MAAM,CAAC,EAAE,aAAa,CAAC,QAAQ,CAAC,CAAC;IACjC,QAAQ,CAAC,EAAE,aAAa,CAAC,UAAU,CAAC,CAAC;IACrC,SAAS,CAAC,EAAE,aAAa,CAAC,WAAW,CAAC,CAAC;IACvC,MAAM,CAAC,EAAE,aAAa,CAAC,QAAQ,CAAC,CAAC;CAClC,CAAC;AAwJF,wBAAsB,iBAAiB,CACrC,KAAK,EAAE,iBAAiB,GACvB,OAAO,CAAC,aAAa,CAAC,CAoKxB;AAED,wBAAgB,mBAAmB,CACjC,SAAS,EAAE,MAAM,EACjB,SAAS,EAAE,MAAM,GAChB,qBAAqB,CAgBvB"}