unpdf 1.5.0 → 1.6.0

package/README.md

@@ -272,6 +272,16 @@ In order to use this method, make sure to meet the following requirements:
 - Use the official PDF.js build (see [Official or Legacy PDF.js Build](#official-or-legacy-pdfjs-build)).
 - Install the [`@napi-rs/canvas`](https://github.com/Brooooooklyn/canvas) package if you are using Node.js. This package is required to render the PDF page as an image.
 
+> [!TIP]
+> In Node.js, `getDocumentProxy` automatically sets `disableFontFace: true` and resolves `standardFontDataUrl` from your local `pdfjs-dist` package for correct font rendering. To customize this behavior, pass your own options:
+>
+> ```ts
+> const pdf = await getDocumentProxy(buffer, {
+>   disableFontFace: false,
+>   standardFontDataUrl: 'https://unpkg.com/pdfjs-dist@latest/standard_fonts/',
+> })
+> ```
+
 **Type Declaration**
 
 ```ts

package/dist/index.cjs

@@ -1,15 +1,73 @@
 'use strict';
 
+function polyfillDOMMatrix() {
+  if (typeof globalThis.DOMMatrix !== "undefined")
+    return;
+  globalThis.DOMMatrix = class DOMMatrix {
+    a;
+    b;
+    c;
+    d;
+    e;
+    f;
+    constructor(init) {
+      if (Array.isArray(init) && init.length === 6) {
+        this.a = init[0];
+        this.b = init[1];
+        this.c = init[2];
+        this.d = init[3];
+        this.e = init[4];
+        this.f = init[5];
+      } else {
+        this.a = 1;
+        this.b = 0;
+        this.c = 0;
+        this.d = 1;
+        this.e = 0;
+        this.f = 0;
+      }
+    }
+    translateSelf(tx, ty = 0) {
+      this.e = this.a * tx + this.c * ty + this.e;
+      this.f = this.b * tx + this.d * ty + this.f;
+      return this;
+    }
+    scaleSelf(sx, sy = sx) {
+      this.a *= sx;
+      this.b *= sx;
+      this.c *= sy;
+      this.d *= sy;
+      return this;
+    }
+  };
+}
+
+function stubBrowserGlobals() {
+  polyfillDOMMatrix();
+}
+
 let resolvedModule;
 const isNode = globalThis.process?.release?.name === "node";
 const isBrowser = typeof window !== "undefined";
 async function getDocumentProxy(data, options = {}) {
   const { getDocument } = await getResolvedPDFJS();
+  let nodeDefaults = {};
+  if (isNode) {
+    try {
+      const base = undefined("pdfjs-dist/package.json");
+      nodeDefaults = {
+        disableFontFace: true,
+        standardFontDataUrl: new URL("./standard_fonts/", base).href
+      };
+    } catch {
+    }
+  }
   const pdf = await getDocument({
     data,
     isEvalSupported: false,
     // See: https://github.com/mozilla/pdf.js/issues/4244#issuecomment-1479534301
     useSystemFonts: true,
+    ...nodeDefaults,
     ...options
   }).promise;
   return pdf;
@@ -24,6 +82,7 @@ async function resolvePDFJSImport(pdfjsResolver, { reload = false } = {}) {
   if (resolvedModule && !reload) {
     return;
   }
+  stubBrowserGlobals();
   if (pdfjsResolver) {
     try {
       resolvedModule = await interopDefault(pdfjsResolver());
@@ -122,8 +181,7 @@ async function resolveCanvasModule(canvasImport) {
 function injectCanvasConstructors() {
   if (!resolvedCanvasModule)
     return;
-  if (typeof globalThis.DOMMatrix === "undefined")
-    globalThis.DOMMatrix = resolvedCanvasModule.DOMMatrix;
+  globalThis.DOMMatrix = resolvedCanvasModule.DOMMatrix;
   if (typeof globalThis.ImageData === "undefined")
     globalThis.ImageData = resolvedCanvasModule.ImageData;
   if (typeof globalThis.Path2D === "undefined")
@@ -282,6 +340,32 @@ function parseISODateString(isoDateString) {
   }
 }
 
+async function extractTextItems$1(data) {
+  const pdf = isPDFDocumentProxy(data) ? data : await getDocumentProxy(data);
+  const items = await Promise.all(
+    Array.from({ length: pdf.numPages }, (_, i) => getPageTextItems(pdf, i + 1))
+  );
+  return { totalPages: pdf.numPages, items };
+}
+async function getPageTextItems(document, pageNumber) {
+  const page = await document.getPage(pageNumber);
+  const content = await page.getTextContent();
+  const styles = content.styles;
+  return content.items.filter((item) => item.str != null).map((item) => {
+    const [_a, _b, c, d, e, f] = item.transform;
+    return {
+      str: item.str,
+      x: e,
+      y: f,
+      width: item.width,
+      height: item.height,
+      fontSize: Math.hypot(c, d),
+      fontFamily: styles[item.fontName]?.fontFamily ?? "",
+      dir: item.dir,
+      hasEOL: item.hasEOL
+    };
+  });
+}
 async function extractText$1(data, options = {}) {
   const { mergePages = false } = options;
   const pdf = isPDFDocumentProxy(data) ? data : await getDocumentProxy(data);
@@ -314,6 +398,10 @@ const extractText = async (...args) => {
   await resolvePDFJSImport();
   return await extractText$1(...args);
 };
+const extractTextItems = async (...args) => {
+  await resolvePDFJSImport();
+  return await extractTextItems$1(...args);
+};
 const extractImages = async (...args) => {
   await resolvePDFJSImport();
   return await extractImages$1(...args);
@@ -333,6 +421,7 @@ exports.definePDFJSModule = definePDFJSModule;
 exports.extractImages = extractImages;
 exports.extractLinks = extractLinks;
 exports.extractText = extractText;
+exports.extractTextItems = extractTextItems;
 exports.getDocumentProxy = getDocumentProxy;
 exports.getMeta = getMeta;
 exports.getResolvedPDFJS = getResolvedPDFJS;

package/dist/index.d.cts

@@ -100,6 +100,30 @@ declare function getMeta$1(data: DocumentInitParameters['data'] | PDFDocumentPro
     metadata: pdfjs_dist_types_src_display_metadata.Metadata;
 }>;
 
+interface StructuredTextItem {
+    /** Text content. */
+    str: string;
+    /** X position in PDF coordinate space (origin: bottom-left). */
+    x: number;
+    /** Y position in PDF coordinate space (origin: bottom-left). */
+    y: number;
+    /** Width in device space. */
+    width: number;
+    /** Height in device space. */
+    height: number;
+    /** Font size derived from the transformation matrix. */
+    fontSize: number;
+    /** Font family name. */
+    fontFamily: string;
+    /** Text direction: `"ltr"`, `"rtl"`, or `"ttb"`. */
+    dir: string;
+    /** Whether the text item is followed by a line break. */
+    hasEOL: boolean;
+}
+declare function extractTextItems$1(data: DocumentInitParameters['data'] | PDFDocumentProxy): Promise<{
+    totalPages: number;
+    items: StructuredTextItem[][];
+}>;
 declare function extractText$1(data: DocumentInitParameters['data'] | PDFDocumentProxy, options?: {
     mergePages?: false;
 }): Promise<{
@@ -136,6 +160,10 @@ declare function configureUnPDF(options: {
  * Applies the following defaults:
  * - `isEvalSupported: false`
  * - `useSystemFonts: true`
+ *
+ * In Node.js environments, additionally applies:
+ * - `disableFontFace: true`
+ * - `standardFontDataUrl` resolved from the local `pdfjs-dist` package
  */
 declare function getDocumentProxy(data: DocumentInitParameters['data'], options?: DocumentInitParameters): Promise<PDFDocumentProxy>;
 declare function getResolvedPDFJS(): Promise<typeof PDFJS>;
@@ -145,8 +173,10 @@ declare function resolvePDFJSImport(pdfjsResolver?: () => Promise<any>, { reload
 
 declare const getMeta: typeof getMeta$1;
 declare const extractText: typeof extractText$1;
+declare const extractTextItems: typeof extractTextItems$1;
 declare const extractImages: typeof extractImages$1;
 declare const renderPageAsImage: typeof renderPageAsImage$1;
 declare const extractLinks: typeof extractLinks$1;
 
-export { configureUnPDF, createIsomorphicCanvasFactory, definePDFJSModule, extractImages, extractLinks, extractText, getDocumentProxy, getMeta, getResolvedPDFJS, renderPageAsImage, resolvePDFJSImport };
+export { configureUnPDF, createIsomorphicCanvasFactory, definePDFJSModule, extractImages, extractLinks, extractText, extractTextItems, getDocumentProxy, getMeta, getResolvedPDFJS, renderPageAsImage, resolvePDFJSImport };
+export type { StructuredTextItem };

package/dist/index.d.mts

@@ -100,6 +100,30 @@ declare function getMeta$1(data: DocumentInitParameters['data'] | PDFDocumentPro
     metadata: pdfjs_dist_types_src_display_metadata.Metadata;
 }>;
 
+interface StructuredTextItem {
+    /** Text content. */
+    str: string;
+    /** X position in PDF coordinate space (origin: bottom-left). */
+    x: number;
+    /** Y position in PDF coordinate space (origin: bottom-left). */
+    y: number;
+    /** Width in device space. */
+    width: number;
+    /** Height in device space. */
+    height: number;
+    /** Font size derived from the transformation matrix. */
+    fontSize: number;
+    /** Font family name. */
+    fontFamily: string;
+    /** Text direction: `"ltr"`, `"rtl"`, or `"ttb"`. */
+    dir: string;
+    /** Whether the text item is followed by a line break. */
+    hasEOL: boolean;
+}
+declare function extractTextItems$1(data: DocumentInitParameters['data'] | PDFDocumentProxy): Promise<{
+    totalPages: number;
+    items: StructuredTextItem[][];
+}>;
 declare function extractText$1(data: DocumentInitParameters['data'] | PDFDocumentProxy, options?: {
     mergePages?: false;
 }): Promise<{
@@ -136,6 +160,10 @@ declare function configureUnPDF(options: {
  * Applies the following defaults:
  * - `isEvalSupported: false`
  * - `useSystemFonts: true`
+ *
+ * In Node.js environments, additionally applies:
+ * - `disableFontFace: true`
+ * - `standardFontDataUrl` resolved from the local `pdfjs-dist` package
  */
 declare function getDocumentProxy(data: DocumentInitParameters['data'], options?: DocumentInitParameters): Promise<PDFDocumentProxy>;
 declare function getResolvedPDFJS(): Promise<typeof PDFJS>;
@@ -145,8 +173,10 @@ declare function resolvePDFJSImport(pdfjsResolver?: () => Promise<any>, { reload
 
 declare const getMeta: typeof getMeta$1;
 declare const extractText: typeof extractText$1;
+declare const extractTextItems: typeof extractTextItems$1;
 declare const extractImages: typeof extractImages$1;
 declare const renderPageAsImage: typeof renderPageAsImage$1;
 declare const extractLinks: typeof extractLinks$1;
 
-export { configureUnPDF, createIsomorphicCanvasFactory, definePDFJSModule, extractImages, extractLinks, extractText, getDocumentProxy, getMeta, getResolvedPDFJS, renderPageAsImage, resolvePDFJSImport };
+export { configureUnPDF, createIsomorphicCanvasFactory, definePDFJSModule, extractImages, extractLinks, extractText, extractTextItems, getDocumentProxy, getMeta, getResolvedPDFJS, renderPageAsImage, resolvePDFJSImport };
+export type { StructuredTextItem };

package/dist/index.d.ts

@@ -100,6 +100,30 @@ declare function getMeta$1(data: DocumentInitParameters['data'] | PDFDocumentPro
     metadata: pdfjs_dist_types_src_display_metadata.Metadata;
 }>;
 
+interface StructuredTextItem {
+    /** Text content. */
+    str: string;
+    /** X position in PDF coordinate space (origin: bottom-left). */
+    x: number;
+    /** Y position in PDF coordinate space (origin: bottom-left). */
+    y: number;
+    /** Width in device space. */
+    width: number;
+    /** Height in device space. */
+    height: number;
+    /** Font size derived from the transformation matrix. */
+    fontSize: number;
+    /** Font family name. */
+    fontFamily: string;
+    /** Text direction: `"ltr"`, `"rtl"`, or `"ttb"`. */
+    dir: string;
+    /** Whether the text item is followed by a line break. */
+    hasEOL: boolean;
+}
+declare function extractTextItems$1(data: DocumentInitParameters['data'] | PDFDocumentProxy): Promise<{
+    totalPages: number;
+    items: StructuredTextItem[][];
+}>;
 declare function extractText$1(data: DocumentInitParameters['data'] | PDFDocumentProxy, options?: {
     mergePages?: false;
 }): Promise<{
@@ -136,6 +160,10 @@ declare function configureUnPDF(options: {
  * Applies the following defaults:
  * - `isEvalSupported: false`
  * - `useSystemFonts: true`
+ *
+ * In Node.js environments, additionally applies:
+ * - `disableFontFace: true`
+ * - `standardFontDataUrl` resolved from the local `pdfjs-dist` package
  */
 declare function getDocumentProxy(data: DocumentInitParameters['data'], options?: DocumentInitParameters): Promise<PDFDocumentProxy>;
 declare function getResolvedPDFJS(): Promise<typeof PDFJS>;
@@ -145,8 +173,10 @@ declare function resolvePDFJSImport(pdfjsResolver?: () => Promise<any>, { reload
 
 declare const getMeta: typeof getMeta$1;
 declare const extractText: typeof extractText$1;
+declare const extractTextItems: typeof extractTextItems$1;
 declare const extractImages: typeof extractImages$1;
 declare const renderPageAsImage: typeof renderPageAsImage$1;
 declare const extractLinks: typeof extractLinks$1;
 
-export { configureUnPDF, createIsomorphicCanvasFactory, definePDFJSModule, extractImages, extractLinks, extractText, getDocumentProxy, getMeta, getResolvedPDFJS, renderPageAsImage, resolvePDFJSImport };
+export { configureUnPDF, createIsomorphicCanvasFactory, definePDFJSModule, extractImages, extractLinks, extractText, extractTextItems, getDocumentProxy, getMeta, getResolvedPDFJS, renderPageAsImage, resolvePDFJSImport };
+export type { StructuredTextItem };

package/dist/index.mjs

@@ -1,13 +1,71 @@
+function polyfillDOMMatrix() {
+  if (typeof globalThis.DOMMatrix !== "undefined")
+    return;
+  globalThis.DOMMatrix = class DOMMatrix {
+    a;
+    b;
+    c;
+    d;
+    e;
+    f;
+    constructor(init) {
+      if (Array.isArray(init) && init.length === 6) {
+        this.a = init[0];
+        this.b = init[1];
+        this.c = init[2];
+        this.d = init[3];
+        this.e = init[4];
+        this.f = init[5];
+      } else {
+        this.a = 1;
+        this.b = 0;
+        this.c = 0;
+        this.d = 1;
+        this.e = 0;
+        this.f = 0;
+      }
+    }
+    translateSelf(tx, ty = 0) {
+      this.e = this.a * tx + this.c * ty + this.e;
+      this.f = this.b * tx + this.d * ty + this.f;
+      return this;
+    }
+    scaleSelf(sx, sy = sx) {
+      this.a *= sx;
+      this.b *= sx;
+      this.c *= sy;
+      this.d *= sy;
+      return this;
+    }
+  };
+}
+
+function stubBrowserGlobals() {
+  polyfillDOMMatrix();
+}
+
 let resolvedModule;
 const isNode = globalThis.process?.release?.name === "node";
 const isBrowser = typeof window !== "undefined";
 async function getDocumentProxy(data, options = {}) {
   const { getDocument } = await getResolvedPDFJS();
+  let nodeDefaults = {};
+  if (isNode) {
+    try {
+      const base = import.meta.resolve("pdfjs-dist/package.json");
+      nodeDefaults = {
+        disableFontFace: true,
+        standardFontDataUrl: new URL("./standard_fonts/", base).href
+      };
+    } catch {
+    }
+  }
   const pdf = await getDocument({
     data,
     isEvalSupported: false,
     // See: https://github.com/mozilla/pdf.js/issues/4244#issuecomment-1479534301
     useSystemFonts: true,
+    ...nodeDefaults,
     ...options
   }).promise;
   return pdf;
@@ -22,6 +80,7 @@ async function resolvePDFJSImport(pdfjsResolver, { reload = false } = {}) {
   if (resolvedModule && !reload) {
     return;
   }
+  stubBrowserGlobals();
   if (pdfjsResolver) {
     try {
       resolvedModule = await interopDefault(pdfjsResolver());
@@ -120,8 +179,7 @@ async function resolveCanvasModule(canvasImport) {
 function injectCanvasConstructors() {
   if (!resolvedCanvasModule)
     return;
-  if (typeof globalThis.DOMMatrix === "undefined")
-    globalThis.DOMMatrix = resolvedCanvasModule.DOMMatrix;
+  globalThis.DOMMatrix = resolvedCanvasModule.DOMMatrix;
   if (typeof globalThis.ImageData === "undefined")
     globalThis.ImageData = resolvedCanvasModule.ImageData;
   if (typeof globalThis.Path2D === "undefined")
@@ -280,6 +338,32 @@ function parseISODateString(isoDateString) {
   }
 }
 
+async function extractTextItems$1(data) {
+  const pdf = isPDFDocumentProxy(data) ? data : await getDocumentProxy(data);
+  const items = await Promise.all(
+    Array.from({ length: pdf.numPages }, (_, i) => getPageTextItems(pdf, i + 1))
+  );
+  return { totalPages: pdf.numPages, items };
+}
+async function getPageTextItems(document, pageNumber) {
+  const page = await document.getPage(pageNumber);
+  const content = await page.getTextContent();
+  const styles = content.styles;
+  return content.items.filter((item) => item.str != null).map((item) => {
+    const [_a, _b, c, d, e, f] = item.transform;
+    return {
+      str: item.str,
+      x: e,
+      y: f,
+      width: item.width,
+      height: item.height,
+      fontSize: Math.hypot(c, d),
+      fontFamily: styles[item.fontName]?.fontFamily ?? "",
+      dir: item.dir,
+      hasEOL: item.hasEOL
+    };
+  });
+}
 async function extractText$1(data, options = {}) {
   const { mergePages = false } = options;
   const pdf = isPDFDocumentProxy(data) ? data : await getDocumentProxy(data);
@@ -312,6 +396,10 @@ const extractText = async (...args) => {
   await resolvePDFJSImport();
   return await extractText$1(...args);
 };
+const extractTextItems = async (...args) => {
+  await resolvePDFJSImport();
+  return await extractTextItems$1(...args);
+};
 const extractImages = async (...args) => {
   await resolvePDFJSImport();
   return await extractImages$1(...args);
@@ -325,4 +413,4 @@ const extractLinks = async (...args) => {
   return await extractLinks$1(...args);
 };
 
-export { configureUnPDF, createIsomorphicCanvasFactory, definePDFJSModule, extractImages, extractLinks, extractText, getDocumentProxy, getMeta, getResolvedPDFJS, renderPageAsImage, resolvePDFJSImport };
+export { configureUnPDF, createIsomorphicCanvasFactory, definePDFJSModule, extractImages, extractLinks, extractText, extractTextItems, getDocumentProxy, getMeta, getResolvedPDFJS, renderPageAsImage, resolvePDFJSImport };