unpdf 0.10.1 → 0.12.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (62) hide show
  1. package/README.md +12 -8
  2. package/dist/index.cjs +15 -15
  3. package/dist/index.d.cts +32 -13
  4. package/dist/index.d.mts +32 -13
  5. package/dist/index.d.ts +32 -13
  6. package/dist/index.mjs +15 -15
  7. package/dist/pdfjs.d.mts +1 -0
  8. package/dist/pdfjs.d.ts +1 -0
  9. package/dist/pdfjs.mjs +1 -1
  10. package/dist/types/src/display/annotation_layer.d.ts +12 -5
  11. package/dist/types/src/display/annotation_storage.d.ts +10 -0
  12. package/dist/types/src/display/api.d.ts +95 -45
  13. package/dist/types/src/display/base_factory.d.ts +7 -1
  14. package/dist/types/src/display/canvas.d.ts +4 -0
  15. package/dist/types/src/display/display_utils.d.ts +6 -4
  16. package/dist/types/src/display/draw_layer.d.ts +12 -14
  17. package/dist/types/src/display/editor/alt_text.d.ts +33 -0
  18. package/dist/types/src/display/editor/annotation_editor_layer.d.ts +26 -2
  19. package/dist/types/src/display/editor/color_picker.d.ts +19 -0
  20. package/dist/types/src/display/editor/editor.d.ts +85 -16
  21. package/dist/types/src/display/editor/freetext.d.ts +6 -1
  22. package/dist/types/src/display/editor/highlight.d.ts +88 -0
  23. package/dist/types/src/display/editor/ink.d.ts +5 -4
  24. package/dist/types/src/display/editor/outliner.d.ts +53 -9
  25. package/dist/types/src/display/editor/stamp.d.ts +22 -1
  26. package/dist/types/src/display/editor/toolbar.d.ts +10 -1
  27. package/dist/types/src/display/editor/tools.d.ts +56 -6
  28. package/dist/types/src/display/fetch_stream.d.ts +3 -4
  29. package/dist/types/src/display/font_loader.d.ts +1 -5
  30. package/dist/types/src/display/network.d.ts +2 -3
  31. package/dist/types/src/display/node_stream.d.ts +5 -6
  32. package/dist/types/src/display/node_utils.d.ts +6 -5
  33. package/dist/types/src/display/optional_content_config.d.ts +6 -1
  34. package/dist/types/src/display/text_layer.d.ts +37 -106
  35. package/dist/types/src/display/transport_stream.d.ts +3 -7
  36. package/dist/types/src/display/worker_options.d.ts +20 -22
  37. package/dist/types/src/pdf.d.ts +3 -6
  38. package/dist/types/src/shared/util.d.ts +24 -22
  39. package/dist/types/web/annotation_editor_layer_builder.d.ts +10 -13
  40. package/dist/types/web/annotation_layer_builder.d.ts +8 -4
  41. package/dist/types/web/app_options.d.ts +5 -4
  42. package/dist/types/web/download_manager.d.ts +1 -2
  43. package/dist/types/web/draw_layer_builder.d.ts +23 -0
  44. package/dist/types/web/event_utils.d.ts +8 -6
  45. package/dist/types/web/genericl10n.d.ts +5 -3
  46. package/dist/types/web/interfaces.d.ts +10 -14
  47. package/dist/types/web/l10n.d.ts +7 -2
  48. package/dist/types/web/pdf_find_controller.d.ts +25 -2
  49. package/dist/types/web/pdf_history.d.ts +1 -53
  50. package/dist/types/web/pdf_link_service.d.ts +4 -107
  51. package/dist/types/web/pdf_page_view.d.ts +16 -21
  52. package/dist/types/web/pdf_scripting_manager.d.ts +6 -10
  53. package/dist/types/web/pdf_thumbnail_view.d.ts +11 -15
  54. package/dist/types/web/pdf_thumbnail_viewer.d.ts +16 -17
  55. package/dist/types/web/pdf_viewer.component.d.ts +1 -2
  56. package/dist/types/web/pdf_viewer.d.ts +67 -35
  57. package/dist/types/web/text_accessibility.d.ts +1 -1
  58. package/dist/types/web/text_highlighter.d.ts +1 -1
  59. package/dist/types/web/text_layer_builder.d.ts +17 -29
  60. package/dist/types/web/ui_utils.d.ts +7 -2
  61. package/package.json +24 -23
  62. package/dist/types/web/l10n_utils.d.ts +0 -9
package/README.md CHANGED
@@ -16,8 +16,7 @@ This library is also intended as a modern alternative to the unmaintained but st
16
16
 
17
17
  ## PDF.js Compatibility
18
18
 
19
- > [!NOTE]
20
- > This package is currently using PDF.js v4.0.189.
19
+ The serverless build of PDF.js provided by `unpdf` is based on PDF.js v4.6.82. If you need a different version, you can [use another PDF.js build](#use-official-or-legacy-pdfjs-build).
21
20
 
22
21
  ## Installation
23
22
 
@@ -50,7 +49,7 @@ const buffer = await fetch(
50
49
  const buffer = await readFile("./dummy.pdf");
51
50
 
52
51
  // Load PDF from buffer
53
- const pdf = await getDocumentProxy(new Uint8Array(pdf));
52
+ const pdf = await getDocumentProxy(new Uint8Array(buffer));
54
53
  // Extract text from PDF
55
54
  const { totalPages, text } = await extractText(pdf, { mergePages: true });
56
55
  ```
@@ -86,12 +85,15 @@ for (let i = 1; i <= doc.numPages; i++) {
86
85
 
87
86
  Generally speaking, you don't need to worry about the PDF.js build. `unpdf` ships with a serverless build of the latest PDF.js version. However, if you want to use the official PDF.js version or the legacy build, you can define a custom PDF.js module.
88
87
 
88
+ > [!WARNING]
89
+ > The latest PDF.js v4.6.82 uses `Promise.withResolvers`, which may not be supported in all environments, such as Node < 22. Consider to use the bundled serverless build, which includes a polyfill, or use an older version of PDF.js.
90
+
89
91
  ```ts
90
92
  // Before using any other method, define the PDF.js module
91
93
  // if you need another PDF.js build
92
94
  import { configureUnPDF } from "unpdf";
93
95
 
94
- configureUnPDF({
96
+ await configureUnPDF({
95
97
  // Use the official PDF.js build (make sure to install it first)
96
98
  pdfjs: () => import("pdfjs-dist"),
97
99
  });
@@ -138,7 +140,9 @@ function getResolvedPDFJS(): Promise<PDFJS>;
138
140
  ### `getMeta`
139
141
 
140
142
  ```ts
141
- function getMeta(data: BinaryData | PDFDocumentProxy): Promise<{
143
+ function getMeta(
144
+ data: DocumentInitParameters["data"] | PDFDocumentProxy,
145
+ ): Promise<{
142
146
  info: Record<string, any>;
143
147
  metadata: Record<string, any>;
144
148
  }>;
@@ -150,7 +154,7 @@ Extracts all text from a PDF. If `mergePages` is set to `true`, the text of all
150
154
 
151
155
  ```ts
152
156
  function extractText(
153
- data: BinaryData | PDFDocumentProxy,
157
+ data: DocumentInitParameters["data"] | PDFDocumentProxy,
154
158
  { mergePages }?: { mergePages?: boolean },
155
159
  ): Promise<{
156
160
  totalPages: number;
@@ -175,7 +179,7 @@ In order to use this method, you have to meet the following requirements:
175
179
  ```ts
176
180
  import { configureUnPDF, renderPageAsImage } from "unpdf";
177
181
 
178
- configureUnPDF({
182
+ await configureUnPDF({
179
183
  // Use the official PDF.js build
180
184
  pdfjs: () => import("pdfjs-dist"),
181
185
  });
@@ -194,7 +198,7 @@ await writeFile("dummy-page-1.png", Buffer.from(result));
194
198
 
195
199
  ```ts
196
200
  declare function renderPageAsImage(
197
- data: BinaryData,
201
+ data: DocumentInitParameters["data"],
198
202
  pageNumber: number,
199
203
  options?: {
200
204
  canvas?: () => Promise<typeof import("canvas")>;
package/dist/index.cjs CHANGED
@@ -31,16 +31,18 @@ async function resolvePDFJSImports(pdfjsResolver, { force = false } = {}) {
31
31
  resolvedModule = await resolvedModule.resolvePDFJS();
32
32
  }
33
33
  return;
34
- } catch {
34
+ } catch (error) {
35
+ console.error(error);
35
36
  throw new Error(
36
37
  "Resolving failed. Please check the provided configuration."
37
38
  );
38
39
  }
39
40
  }
40
41
  try {
41
- const { resolvePDFJS } = await import('unpdf/pdfjs');
42
+ const { resolvePDFJS } = await Promise.resolve().then(function () { return pdfjs; });
42
43
  resolvedModule = await resolvePDFJS();
43
- } catch {
44
+ } catch (error) {
45
+ console.error(error);
44
46
  throw new Error(
45
47
  "PDF.js is not available. Please add the package as a dependency."
46
48
  );
@@ -69,20 +71,15 @@ async function extractText$1(data, options = {}) {
69
71
  const texts = await Promise.all(
70
72
  Array.from({ length: pdf.numPages }, (_, i) => getPageText(pdf, i + 1))
71
73
  );
72
- const filteredTexts = texts.filter(Boolean);
73
74
  return {
74
75
  totalPages: pdf.numPages,
75
- text: mergePages ? filteredTexts.join("\n\n") : filteredTexts
76
+ text: mergePages ? texts.join("\n").replace(/\s+/g, " ") : texts
76
77
  };
77
78
  }
78
- async function getPageText(pdf, pageNumber) {
79
- const page = await pdf.getPage(pageNumber);
79
+ async function getPageText(document, pageNumber) {
80
+ const page = await document.getPage(pageNumber);
80
81
  const content = await page.getTextContent();
81
- const items = content.items;
82
- if (items.length === 0) {
83
- return;
84
- }
85
- return items.map(({ str }) => str).join("\n");
82
+ return content.items.filter((item) => item.str != null).map((item) => item.str + (item.hasEOL ? "\n" : "")).join("");
86
83
  }
87
84
 
88
85
  async function extractImages$1(data, pageNumber) {
@@ -109,7 +106,7 @@ async function extractImages$1(data, pageNumber) {
109
106
  }
110
107
  async function renderPageAsImage$1(data, pageNumber, options = {}) {
111
108
  const canvasFactory = await createIsomorphicCanvasFactory(options.canvas);
112
- const pdf = await getDocumentProxy(data, { canvasFactory });
109
+ const pdf = isPDFDocumentProxy(data) ? data : await getDocumentProxy(data, { canvasFactory });
113
110
  const page = await pdf.getPage(pageNumber);
114
111
  if (pageNumber < 1 || pageNumber > pdf.numPages) {
115
112
  throw new Error(
@@ -187,7 +184,6 @@ async function configureUnPDF(options) {
187
184
  }
188
185
  }
189
186
 
190
- const defineUnPDFConfig = configureUnPDF;
191
187
  const getMeta = async (...args) => {
192
188
  await resolvePDFJSImports();
193
189
  return await getMeta$1(...args);
@@ -205,8 +201,12 @@ const renderPageAsImage = async (...args) => {
205
201
  return await renderPageAsImage$1(...args);
206
202
  };
207
203
 
204
+ const pdfjs = {
205
+ __proto__: null
206
+ };
207
+
208
208
  exports.configureUnPDF = configureUnPDF;
209
- exports.defineUnPDFConfig = defineUnPDFConfig;
209
+ exports.createIsomorphicCanvasFactory = createIsomorphicCanvasFactory;
210
210
  exports.extractImages = extractImages;
211
211
  exports.extractText = extractText;
212
212
  exports.getDocumentProxy = getDocumentProxy;
package/dist/index.d.cts CHANGED
@@ -1,27 +1,49 @@
1
- import { BinaryData, PDFDocumentProxy, DocumentInitParameters } from './types/src/display/api';
1
+ import { DocumentInitParameters, PDFDocumentProxy } from './types/src/display/api';
2
2
  import * as canvas from 'canvas';
3
+ import { Canvas } from 'canvas';
3
4
  import * as _PDFJS from './types/src/pdf';
4
5
 
5
- declare function getMeta$1(data: BinaryData | PDFDocumentProxy): Promise<{
6
+ declare function getMeta$1(data: DocumentInitParameters["data"] | PDFDocumentProxy): Promise<{
6
7
  info: Record<string, any>;
7
8
  metadata: Record<string, any>;
8
9
  }>;
9
10
 
10
- declare function extractText$1(data: BinaryData | PDFDocumentProxy, options?: {
11
- mergePages?: boolean;
11
+ declare function extractText$1(data: DocumentInitParameters["data"] | PDFDocumentProxy, options?: {
12
+ mergePages?: false;
12
13
  }): Promise<{
13
14
  totalPages: number;
14
- text: string | string[];
15
+ text: string[];
16
+ }>;
17
+ declare function extractText$1(data: DocumentInitParameters["data"] | PDFDocumentProxy, options: {
18
+ mergePages: true;
19
+ }): Promise<{
20
+ totalPages: number;
21
+ text: string;
15
22
  }>;
16
23
 
17
- declare function extractImages$1(data: BinaryData | PDFDocumentProxy, pageNumber: number): Promise<Uint8ClampedArray[]>;
18
- declare function renderPageAsImage$1(data: BinaryData, pageNumber: number, options?: {
24
+ declare function extractImages$1(data: DocumentInitParameters["data"] | PDFDocumentProxy, pageNumber: number): Promise<Uint8ClampedArray[]>;
25
+ declare function renderPageAsImage$1(data: DocumentInitParameters["data"] | PDFDocumentProxy, pageNumber: number, options?: {
19
26
  canvas?: () => Promise<typeof canvas>;
20
27
  /** @default 1 */
21
28
  scale?: number;
22
29
  width?: number;
23
30
  height?: number;
24
31
  }): Promise<ArrayBuffer>;
32
+ declare function createIsomorphicCanvasFactory(canvas?: () => Promise<typeof canvas>): Promise<{
33
+ _createCanvas(width: number, height: number): Canvas | HTMLCanvasElement;
34
+ create(width: number, height: number): {
35
+ canvas: Canvas | HTMLCanvasElement;
36
+ context: CanvasRenderingContext2D;
37
+ };
38
+ reset(ctx: {
39
+ canvas?: Canvas;
40
+ context?: CanvasRenderingContext2D;
41
+ }, width: number, height: number): void;
42
+ destroy(ctx: {
43
+ canvas?: Canvas;
44
+ context?: CanvasRenderingContext2D;
45
+ }): void;
46
+ }>;
25
47
 
26
48
  type PDFJS = typeof _PDFJS;
27
49
  interface UnPDFConfiguration {
@@ -37,8 +59,6 @@ interface UnPDFConfiguration {
37
59
  pdfjs?: () => Promise<any>;
38
60
  }
39
61
 
40
- declare function configureUnPDF(options: UnPDFConfiguration): Promise<void>;
41
-
42
62
  /**
43
63
  * Returns a PDFDocumentProxy instance from a given binary data.
44
64
  *
@@ -46,18 +66,17 @@ declare function configureUnPDF(options: UnPDFConfiguration): Promise<void>;
46
66
  * - `isEvalSupported: false`
47
67
  * - `useSystemFonts: true`
48
68
  */
49
- declare function getDocumentProxy(data: BinaryData, options?: DocumentInitParameters): Promise<PDFDocumentProxy>;
69
+ declare function getDocumentProxy(data: DocumentInitParameters["data"], options?: DocumentInitParameters): Promise<PDFDocumentProxy>;
50
70
  declare function getResolvedPDFJS(): Promise<PDFJS>;
51
71
  declare function resolvePDFJSImports(pdfjsResolver?: () => Promise<any>, { force }?: {
52
72
  force?: boolean | undefined;
53
73
  }): Promise<void>;
54
74
 
55
- /** @deprecated Use `configureUnPDF` instead */
56
- declare const defineUnPDFConfig: typeof configureUnPDF;
75
+ declare function configureUnPDF(options: UnPDFConfiguration): Promise<void>;
57
76
 
58
77
  declare const getMeta: typeof getMeta$1;
59
78
  declare const extractText: typeof extractText$1;
60
79
  declare const extractImages: typeof extractImages$1;
61
80
  declare const renderPageAsImage: typeof renderPageAsImage$1;
62
81
 
63
- export { configureUnPDF, defineUnPDFConfig, extractImages, extractText, getDocumentProxy, getMeta, getResolvedPDFJS, renderPageAsImage, resolvePDFJSImports };
82
+ export { configureUnPDF, createIsomorphicCanvasFactory, extractImages, extractText, getDocumentProxy, getMeta, getResolvedPDFJS, renderPageAsImage, resolvePDFJSImports };
package/dist/index.d.mts CHANGED
@@ -1,27 +1,49 @@
1
- import { BinaryData, PDFDocumentProxy, DocumentInitParameters } from './types/src/display/api';
1
+ import { DocumentInitParameters, PDFDocumentProxy } from './types/src/display/api';
2
2
  import * as canvas from 'canvas';
3
+ import { Canvas } from 'canvas';
3
4
  import * as _PDFJS from './types/src/pdf';
4
5
 
5
- declare function getMeta$1(data: BinaryData | PDFDocumentProxy): Promise<{
6
+ declare function getMeta$1(data: DocumentInitParameters["data"] | PDFDocumentProxy): Promise<{
6
7
  info: Record<string, any>;
7
8
  metadata: Record<string, any>;
8
9
  }>;
9
10
 
10
- declare function extractText$1(data: BinaryData | PDFDocumentProxy, options?: {
11
- mergePages?: boolean;
11
+ declare function extractText$1(data: DocumentInitParameters["data"] | PDFDocumentProxy, options?: {
12
+ mergePages?: false;
12
13
  }): Promise<{
13
14
  totalPages: number;
14
- text: string | string[];
15
+ text: string[];
16
+ }>;
17
+ declare function extractText$1(data: DocumentInitParameters["data"] | PDFDocumentProxy, options: {
18
+ mergePages: true;
19
+ }): Promise<{
20
+ totalPages: number;
21
+ text: string;
15
22
  }>;
16
23
 
17
- declare function extractImages$1(data: BinaryData | PDFDocumentProxy, pageNumber: number): Promise<Uint8ClampedArray[]>;
18
- declare function renderPageAsImage$1(data: BinaryData, pageNumber: number, options?: {
24
+ declare function extractImages$1(data: DocumentInitParameters["data"] | PDFDocumentProxy, pageNumber: number): Promise<Uint8ClampedArray[]>;
25
+ declare function renderPageAsImage$1(data: DocumentInitParameters["data"] | PDFDocumentProxy, pageNumber: number, options?: {
19
26
  canvas?: () => Promise<typeof canvas>;
20
27
  /** @default 1 */
21
28
  scale?: number;
22
29
  width?: number;
23
30
  height?: number;
24
31
  }): Promise<ArrayBuffer>;
32
+ declare function createIsomorphicCanvasFactory(canvas?: () => Promise<typeof canvas>): Promise<{
33
+ _createCanvas(width: number, height: number): Canvas | HTMLCanvasElement;
34
+ create(width: number, height: number): {
35
+ canvas: Canvas | HTMLCanvasElement;
36
+ context: CanvasRenderingContext2D;
37
+ };
38
+ reset(ctx: {
39
+ canvas?: Canvas;
40
+ context?: CanvasRenderingContext2D;
41
+ }, width: number, height: number): void;
42
+ destroy(ctx: {
43
+ canvas?: Canvas;
44
+ context?: CanvasRenderingContext2D;
45
+ }): void;
46
+ }>;
25
47
 
26
48
  type PDFJS = typeof _PDFJS;
27
49
  interface UnPDFConfiguration {
@@ -37,8 +59,6 @@ interface UnPDFConfiguration {
37
59
  pdfjs?: () => Promise<any>;
38
60
  }
39
61
 
40
- declare function configureUnPDF(options: UnPDFConfiguration): Promise<void>;
41
-
42
62
  /**
43
63
  * Returns a PDFDocumentProxy instance from a given binary data.
44
64
  *
@@ -46,18 +66,17 @@ declare function configureUnPDF(options: UnPDFConfiguration): Promise<void>;
46
66
  * - `isEvalSupported: false`
47
67
  * - `useSystemFonts: true`
48
68
  */
49
- declare function getDocumentProxy(data: BinaryData, options?: DocumentInitParameters): Promise<PDFDocumentProxy>;
69
+ declare function getDocumentProxy(data: DocumentInitParameters["data"], options?: DocumentInitParameters): Promise<PDFDocumentProxy>;
50
70
  declare function getResolvedPDFJS(): Promise<PDFJS>;
51
71
  declare function resolvePDFJSImports(pdfjsResolver?: () => Promise<any>, { force }?: {
52
72
  force?: boolean | undefined;
53
73
  }): Promise<void>;
54
74
 
55
- /** @deprecated Use `configureUnPDF` instead */
56
- declare const defineUnPDFConfig: typeof configureUnPDF;
75
+ declare function configureUnPDF(options: UnPDFConfiguration): Promise<void>;
57
76
 
58
77
  declare const getMeta: typeof getMeta$1;
59
78
  declare const extractText: typeof extractText$1;
60
79
  declare const extractImages: typeof extractImages$1;
61
80
  declare const renderPageAsImage: typeof renderPageAsImage$1;
62
81
 
63
- export { configureUnPDF, defineUnPDFConfig, extractImages, extractText, getDocumentProxy, getMeta, getResolvedPDFJS, renderPageAsImage, resolvePDFJSImports };
82
+ export { configureUnPDF, createIsomorphicCanvasFactory, extractImages, extractText, getDocumentProxy, getMeta, getResolvedPDFJS, renderPageAsImage, resolvePDFJSImports };
package/dist/index.d.ts CHANGED
@@ -1,27 +1,49 @@
1
- import { BinaryData, PDFDocumentProxy, DocumentInitParameters } from './types/src/display/api';
1
+ import { DocumentInitParameters, PDFDocumentProxy } from './types/src/display/api';
2
2
  import * as canvas from 'canvas';
3
+ import { Canvas } from 'canvas';
3
4
  import * as _PDFJS from './types/src/pdf';
4
5
 
5
- declare function getMeta$1(data: BinaryData | PDFDocumentProxy): Promise<{
6
+ declare function getMeta$1(data: DocumentInitParameters["data"] | PDFDocumentProxy): Promise<{
6
7
  info: Record<string, any>;
7
8
  metadata: Record<string, any>;
8
9
  }>;
9
10
 
10
- declare function extractText$1(data: BinaryData | PDFDocumentProxy, options?: {
11
- mergePages?: boolean;
11
+ declare function extractText$1(data: DocumentInitParameters["data"] | PDFDocumentProxy, options?: {
12
+ mergePages?: false;
12
13
  }): Promise<{
13
14
  totalPages: number;
14
- text: string | string[];
15
+ text: string[];
16
+ }>;
17
+ declare function extractText$1(data: DocumentInitParameters["data"] | PDFDocumentProxy, options: {
18
+ mergePages: true;
19
+ }): Promise<{
20
+ totalPages: number;
21
+ text: string;
15
22
  }>;
16
23
 
17
- declare function extractImages$1(data: BinaryData | PDFDocumentProxy, pageNumber: number): Promise<Uint8ClampedArray[]>;
18
- declare function renderPageAsImage$1(data: BinaryData, pageNumber: number, options?: {
24
+ declare function extractImages$1(data: DocumentInitParameters["data"] | PDFDocumentProxy, pageNumber: number): Promise<Uint8ClampedArray[]>;
25
+ declare function renderPageAsImage$1(data: DocumentInitParameters["data"] | PDFDocumentProxy, pageNumber: number, options?: {
19
26
  canvas?: () => Promise<typeof canvas>;
20
27
  /** @default 1 */
21
28
  scale?: number;
22
29
  width?: number;
23
30
  height?: number;
24
31
  }): Promise<ArrayBuffer>;
32
+ declare function createIsomorphicCanvasFactory(canvas?: () => Promise<typeof canvas>): Promise<{
33
+ _createCanvas(width: number, height: number): Canvas | HTMLCanvasElement;
34
+ create(width: number, height: number): {
35
+ canvas: Canvas | HTMLCanvasElement;
36
+ context: CanvasRenderingContext2D;
37
+ };
38
+ reset(ctx: {
39
+ canvas?: Canvas;
40
+ context?: CanvasRenderingContext2D;
41
+ }, width: number, height: number): void;
42
+ destroy(ctx: {
43
+ canvas?: Canvas;
44
+ context?: CanvasRenderingContext2D;
45
+ }): void;
46
+ }>;
25
47
 
26
48
  type PDFJS = typeof _PDFJS;
27
49
  interface UnPDFConfiguration {
@@ -37,8 +59,6 @@ interface UnPDFConfiguration {
37
59
  pdfjs?: () => Promise<any>;
38
60
  }
39
61
 
40
- declare function configureUnPDF(options: UnPDFConfiguration): Promise<void>;
41
-
42
62
  /**
43
63
  * Returns a PDFDocumentProxy instance from a given binary data.
44
64
  *
@@ -46,18 +66,17 @@ declare function configureUnPDF(options: UnPDFConfiguration): Promise<void>;
46
66
  * - `isEvalSupported: false`
47
67
  * - `useSystemFonts: true`
48
68
  */
49
- declare function getDocumentProxy(data: BinaryData, options?: DocumentInitParameters): Promise<PDFDocumentProxy>;
69
+ declare function getDocumentProxy(data: DocumentInitParameters["data"], options?: DocumentInitParameters): Promise<PDFDocumentProxy>;
50
70
  declare function getResolvedPDFJS(): Promise<PDFJS>;
51
71
  declare function resolvePDFJSImports(pdfjsResolver?: () => Promise<any>, { force }?: {
52
72
  force?: boolean | undefined;
53
73
  }): Promise<void>;
54
74
 
55
- /** @deprecated Use `configureUnPDF` instead */
56
- declare const defineUnPDFConfig: typeof configureUnPDF;
75
+ declare function configureUnPDF(options: UnPDFConfiguration): Promise<void>;
57
76
 
58
77
  declare const getMeta: typeof getMeta$1;
59
78
  declare const extractText: typeof extractText$1;
60
79
  declare const extractImages: typeof extractImages$1;
61
80
  declare const renderPageAsImage: typeof renderPageAsImage$1;
62
81
 
63
- export { configureUnPDF, defineUnPDFConfig, extractImages, extractText, getDocumentProxy, getMeta, getResolvedPDFJS, renderPageAsImage, resolvePDFJSImports };
82
+ export { configureUnPDF, createIsomorphicCanvasFactory, extractImages, extractText, getDocumentProxy, getMeta, getResolvedPDFJS, renderPageAsImage, resolvePDFJSImports };
package/dist/index.mjs CHANGED
@@ -29,16 +29,18 @@ async function resolvePDFJSImports(pdfjsResolver, { force = false } = {}) {
29
29
  resolvedModule = await resolvedModule.resolvePDFJS();
30
30
  }
31
31
  return;
32
- } catch {
32
+ } catch (error) {
33
+ console.error(error);
33
34
  throw new Error(
34
35
  "Resolving failed. Please check the provided configuration."
35
36
  );
36
37
  }
37
38
  }
38
39
  try {
39
- const { resolvePDFJS } = await import('unpdf/pdfjs');
40
+ const { resolvePDFJS } = await Promise.resolve().then(function () { return pdfjs; });
40
41
  resolvedModule = await resolvePDFJS();
41
- } catch {
42
+ } catch (error) {
43
+ console.error(error);
42
44
  throw new Error(
43
45
  "PDF.js is not available. Please add the package as a dependency."
44
46
  );
@@ -67,20 +69,15 @@ async function extractText$1(data, options = {}) {
67
69
  const texts = await Promise.all(
68
70
  Array.from({ length: pdf.numPages }, (_, i) => getPageText(pdf, i + 1))
69
71
  );
70
- const filteredTexts = texts.filter(Boolean);
71
72
  return {
72
73
  totalPages: pdf.numPages,
73
- text: mergePages ? filteredTexts.join("\n\n") : filteredTexts
74
+ text: mergePages ? texts.join("\n").replace(/\s+/g, " ") : texts
74
75
  };
75
76
  }
76
- async function getPageText(pdf, pageNumber) {
77
- const page = await pdf.getPage(pageNumber);
77
+ async function getPageText(document, pageNumber) {
78
+ const page = await document.getPage(pageNumber);
78
79
  const content = await page.getTextContent();
79
- const items = content.items;
80
- if (items.length === 0) {
81
- return;
82
- }
83
- return items.map(({ str }) => str).join("\n");
80
+ return content.items.filter((item) => item.str != null).map((item) => item.str + (item.hasEOL ? "\n" : "")).join("");
84
81
  }
85
82
 
86
83
  async function extractImages$1(data, pageNumber) {
@@ -107,7 +104,7 @@ async function extractImages$1(data, pageNumber) {
107
104
  }
108
105
  async function renderPageAsImage$1(data, pageNumber, options = {}) {
109
106
  const canvasFactory = await createIsomorphicCanvasFactory(options.canvas);
110
- const pdf = await getDocumentProxy(data, { canvasFactory });
107
+ const pdf = isPDFDocumentProxy(data) ? data : await getDocumentProxy(data, { canvasFactory });
111
108
  const page = await pdf.getPage(pageNumber);
112
109
  if (pageNumber < 1 || pageNumber > pdf.numPages) {
113
110
  throw new Error(
@@ -185,7 +182,6 @@ async function configureUnPDF(options) {
185
182
  }
186
183
  }
187
184
 
188
- const defineUnPDFConfig = configureUnPDF;
189
185
  const getMeta = async (...args) => {
190
186
  await resolvePDFJSImports();
191
187
  return await getMeta$1(...args);
@@ -203,4 +199,8 @@ const renderPageAsImage = async (...args) => {
203
199
  return await renderPageAsImage$1(...args);
204
200
  };
205
201
 
206
- export { configureUnPDF, defineUnPDFConfig, extractImages, extractText, getDocumentProxy, getMeta, getResolvedPDFJS, renderPageAsImage, resolvePDFJSImports };
202
+ const pdfjs = {
203
+ __proto__: null
204
+ };
205
+
206
+ export { configureUnPDF, createIsomorphicCanvasFactory, extractImages, extractText, getDocumentProxy, getMeta, getResolvedPDFJS, renderPageAsImage, resolvePDFJSImports };
package/dist/pdfjs.d.mts CHANGED
@@ -1,3 +1,4 @@
1
1
  import * as PDFJS from './types/src/pdf'
2
2
  declare function resolvePDFJS(): Promise<typeof PDFJS>
3
3
  export { resolvePDFJS }
4
+ export * from './types/src/pdf'
package/dist/pdfjs.d.ts CHANGED
@@ -1,3 +1,4 @@
1
1
  import * as PDFJS from './types/src/pdf'
2
2
  declare function resolvePDFJS(): Promise<typeof PDFJS>
3
3
  export { resolvePDFJS }
4
+ export * from './types/src/pdf'