@laminapdf/wasm 0.2.0-alpha.2 → 0.2.0-alpha.3

package/dist/lamina_pdf_extract.d.ts

@@ -1,7 +1,7 @@
 /* tslint:disable */
 /* eslint-disable */
 
-/** A page's text as `WasmPage.text` returns it (B9c). */
+/** A page's text as `WasmPage.text` returns it. */
 export interface PageTextJs {
     /** The text by line, in content order (runs on one baseline, joined; trimmed). */
     lines: string[];
@@ -196,6 +196,24 @@ export type AfRelationshipInit = "source" | "data" | "alternative" | "supplement
  */
 export type UndoOutcomeJs = "applied" | "partial" | "skipped";
 
+/**
+ * How the Markdown is cut into chunks.
+ */
+export interface ChunkInit {
+    /**
+     * A chunk's budget in approximate tokens (characters / 4); 512 by default.
+     */
+    maxTokens?: number;
+    /**
+     * Approximate tokens of whole blocks repeated from the chunk before; 0 by default.
+     */
+    overlap?: number;
+    /**
+     * Where chunks may start; `"headings"` by default.
+     */
+    splitOn?: SplitOnName;
+}
+
 /**
  * How viewers present the document when it opens.
  */
@@ -306,7 +324,7 @@ export interface PageLabelInit {
 }
 
 /**
- * One text run (`laminapdf::extract::TextRun`): glyphs of one font, size, direction and marked
+ * One text run (`laminapdf::extract::PageTextRun`): glyphs of one font, size, direction and marked
  * content on one baseline, in logical order.
  */
 export interface TextRunJs {
@@ -348,6 +366,44 @@ export interface TextRunJs {
     glyphs: [number, number];
 }
 
+/**
+ * Options for `WasmDocument.extract`.
+ */
+export interface ExtractInit {
+    /**
+     * What to return.
+     */
+    format: ExtractFormat;
+    /**
+     * The 0-based page indices to extract, in this order; every page by default.
+     */
+    pages?: number[];
+    /**
+     * Where the reading order comes from; `"auto"` by default.
+     */
+    readingOrder?: ReadingOrderName;
+    /**
+     * What happens to hidden text.
+     */
+    drop?: DropInit;
+    /**
+     * Which table detectors run; `"lattice"` by default.
+     */
+    tables?: TableModeName;
+    /**
+     * Join words hyphenated across lines; on by default.
+     */
+    dehyphenate?: boolean;
+    /**
+     * Write `<!-- page N -->` before each page's first block (Markdown, chunks and JSON).
+     */
+    pageMarkers?: boolean;
+    /**
+     * How chunks are cut (`chunks`, and `json` to include them).
+     */
+    chunk?: ChunkInit;
+}
+
 /**
  * Options for `WasmDocument.openBytes`.
  */
@@ -573,6 +629,11 @@ export type PageModeInit = "none" | "outlines" | "thumbnails" | "fullScreen" | "
  */
 export type SaveModeJs = "auto" | "incremental" | "fullRewrite";
 
+/**
+ * What `WasmDocument.extract` returns.
+ */
+export type ExtractFormat = "markdown" | "json" | "text" | "chunks";
+
 /**
  * What a save wrote (`laminapdf::SaveReport`).
  */
@@ -608,11 +669,41 @@ export interface SaveReportJs {
  */
 export type OutputJs = "whole" | "appendedOnly";
 
+/**
+ * What happens to hidden text and page furniture.
+ */
+export interface DropInit {
+    /**
+     * Keep each dropped run's text in its `dropped` entry (JSON only), for auditing a file.
+     */
+    keepSpans?: boolean;
+    /**
+     * Drop running headers, footers, page numbers and artifact-marked text too, listing them
+     * in `dropped`; on by default.
+     */
+    furniture?: boolean;
+}
+
 /**
  * Where a link or outline item goes.
  */
 export type LinkTargetInit = { type: "uri"; uri: string } | { type: "page"; index: number; fit?: FitInit } | { type: "named"; name: string };
 
+/**
+ * Where chunks may start.
+ */
+export type SplitOnName = "headings" | "blocks" | "pages";
+
+/**
+ * Where the reading order comes from.
+ */
+export type ReadingOrderName = "auto" | "tagged" | "layout" | "content-stream";
+
+/**
+ * Which table detectors run.
+ */
+export type TableModeName = "off" | "lattice" | "lattice-and-stream";
+
 /**
  * `Limits` with every field optional (ADR 0003 §3; names as in Rust, camelCase).
  */
@@ -782,6 +873,15 @@ export class WasmDocument {
      * error (nothing is applied).
      */
     apply(ops: Op[]): EditedJs;
+    /**
+     * The document's text for retrieval and LLMs, as one string: Markdown, plain text, the
+     * `laminapdf.extract/1` JSON, or the chunks as a JSON array (the extract and text modules).
+     *
+     * # Errors
+     * `page-index-out-of-range` for a page in `pages` past the end; `data-pending` and the
+     * reading errors; `unknown-option` / `invalid-argument` for bad options.
+     */
+    extract(options: ExtractInit): string;
     /**
      * Whether there are changes no save has written yet.
      */
@@ -935,7 +1035,8 @@ export class WasmPage {
      * The page's current 0-based index.
      *
      * # Errors
-     * `unknown-page` when the page is gone; reading errors.
+     * `unknown-page` when the page is gone; `disposed` when its document is freed; reading
+     * errors.
      */
     index(): number;
     /**
@@ -946,12 +1047,12 @@ export class WasmPage {
      */
     info(): PageInfoJs;
     /**
-     * The page's text: lines, runs, glyph quads and glyph text (B9c; the extract and text
-     * modules).
+     * The page's text: lines, runs, glyph quads and glyph text (the extract and text modules).
      *
      * # Errors
-     * `unknown-page` when the page is gone; `data-pending` and the reading errors;
-     * `unknown-option` / `invalid-argument` for bad options.
+     * `unknown-page` when the page is gone; `disposed` when its document is freed;
+     * `data-pending` and the reading errors; `unknown-option` / `invalid-argument` for bad
+     * options.
      */
     text(options: TextOptionsInit | undefined): PageTextJs;
     /**
@@ -1155,7 +1256,8 @@ export class WasmSaveJob {
      * `{ state: "done", bytes, report }`.
      *
      * # Errors
-     * `cancelled`, `document-changed`, `job-finished`, or the save's own errors.
+     * `cancelled`, `document-changed`, `job-finished`, `disposed` when the document is freed,
+     * or the save's own errors.
      */
     step(budget: number): ({ state: "progress" } & ProgressJs) | { state: "done"; bytes: Uint8Array<ArrayBuffer>; report: SaveReportJs };
 }
@@ -1214,6 +1316,14 @@ export class WasmWarningCursor {
     remaining(): number;
 }
 
+export function __wbg_reset_state(): void;
+
+/**
+ * Whether the most recent allocation failed: after a trap, `true` means the module ran out of
+ * memory.
+ */
+export function allocationFailed(): boolean;
+
 /**
  * Whether this module is the text module (fonts, `text`, `textBox`).
  */
@@ -1236,6 +1346,7 @@ export interface InitOutput {
     readonly __wbg_wasmsavejob_free: (a: number, b: number) => void;
     readonly __wbg_wasmsparsesource_free: (a: number, b: number) => void;
     readonly __wbg_wasmwarningcursor_free: (a: number, b: number) => void;
+    readonly allocationFailed: () => number;
     readonly hasText: () => number;
     readonly version: () => [number, number];
     readonly wasmbuilder_attach: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: any) => [number, number];
@@ -1250,6 +1361,7 @@ export interface InitOutput {
     readonly wasmbuilder_viewerPreferences: (a: number, b: any) => [number, number];
     readonly wasmdocument_apply: (a: number, b: any) => [number, number, number];
     readonly wasmdocument_applyLog: (a: number, b: number, c: number) => [number, number, number];
+    readonly wasmdocument_extract: (a: number, b: any) => [number, number, number, number];
     readonly wasmdocument_hasUnsavedChanges: (a: number) => number;
     readonly wasmdocument_info: (a: number) => [number, number, number];
     readonly wasmdocument_openBytes: (a: number, b: number, c: any) => [number, number, number];