afpp 1.4.3 → 1.5.0
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.
- package/dist/parsePdf.js +3 -2
- package/dist/parsePdf.js.map +1 -1
- package/dist/pdf2image.js +3 -2
- package/dist/pdf2image.js.map +1 -1
- package/dist/types/utils.d.ts +3 -0
- package/dist/utils.js +27 -0
- package/dist/utils.js.map +1 -0
- package/package.json +1 -1
package/dist/parsePdf.js
CHANGED
|
@@ -4,6 +4,7 @@ exports.parsePdf = void 0;
|
|
|
4
4
|
/* eslint-disable no-underscore-dangle */
|
|
5
5
|
const promises_1 = require("node:fs/promises");
|
|
6
6
|
const canvas_1 = require("canvas");
|
|
7
|
+
const utils_1 = require("./utils");
|
|
7
8
|
const defaultParsePdfCallback = async (content) => content;
|
|
8
9
|
const parsePdfFileBuffer = async (options, callback = defaultParsePdfCallback) => import('pdfjs-dist/legacy/build/pdf.mjs').then(async (pdfjsLib) => {
|
|
9
10
|
const loadingTask = pdfjsLib.getDocument({
|
|
@@ -21,11 +22,11 @@ const parsePdfFileBuffer = async (options, callback = defaultParsePdfCallback) =
|
|
|
21
22
|
});
|
|
22
23
|
const items = textContent.items;
|
|
23
24
|
if (items.length === 0) {
|
|
24
|
-
const viewport = page.getViewport({ scale:
|
|
25
|
+
const viewport = page.getViewport({ scale: 2.0 });
|
|
25
26
|
const canvas = (0, canvas_1.createCanvas)(viewport.width, viewport.height);
|
|
26
27
|
const context = canvas.getContext('2d');
|
|
27
28
|
await page.render({ canvasContext: context, viewport }).promise;
|
|
28
|
-
const imageBuffer =
|
|
29
|
+
const imageBuffer = await (0, utils_1.toBufferAsync)(canvas);
|
|
29
30
|
pageContents[pageNum - 1] = await callback(imageBuffer);
|
|
30
31
|
}
|
|
31
32
|
else {
|
package/dist/parsePdf.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"parsePdf.js","sourceRoot":"","sources":["../src/parsePdf.ts"],"names":[],"mappings":";;;AAAA,yCAAyC;AACzC,+CAA4C;AAE5C,mCAAsC;
|
|
1
|
+
{"version":3,"file":"parsePdf.js","sourceRoot":"","sources":["../src/parsePdf.ts"],"names":[],"mappings":";;;AAAA,yCAAyC;AACzC,+CAA4C;AAE5C,mCAAsC;AAOtC,2CAAgD;AAIhD,MAAM,uBAAuB,GAAsC,KAAK,EACtE,OAAO,EAEP,EAAE,CAAC,OAAO,CAAC;AAEb,MAAM,kBAAkB,GAAG,KAAK,EAC9B,OAA+B,EAC/B,WAAgC,uBAA8C,EAC9E,EAAE,CACF,MAAM,CAAC,iCAAiC,CAAC,CAAC,IAAI,CAAC,KAAK,EAAE,QAAQ,EAAE,EAAE;IAChE,MAAM,WAAW,GAAG,QAAQ,CAAC,WAAW,CAAC;QACvC,GAAG,OAAO;QACV,SAAS,EAAE,CAAC;KACb,CAAC,CAAC;IAEH,MAAM,WAAW,GAAG,MAAM,WAAW,CAAC,OAAO,CAAC;IAE9C,MAAM,EAAE,QAAQ,EAAE,GAAG,WAAW,CAAC;IACjC,MAAM,YAAY,GAAQ,KAAK,CAAC,IAAI,CAClC,EAAE,MAAM,EAAE,QAAQ,EAAE,EACpB,GAAG,EAAE,CAAC,IAAoB,CAC3B,CAAC;IACF,MAAM,YAAY,GAAmC,EAAE,CAAC;IAExD,KAAK,IAAI,OAAO,GAAG,CAAC,EAAE,OAAO,IAAI,QAAQ,EAAE,OAAO,IAAI,CAAC,EAAE,CAAC;QACxD,YAAY,CAAC,IAAI,CACf,WAAW,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,KAAK,EAAE,IAAI,EAAE,EAAE;YAC/C,MAAM,WAAW,GAAG,MAAM,IAAI,CAAC,cAAc,CAAC;gBAC5C,oBAAoB,EAAE,KAAK;aAC5B,CAAC,CAAC;YACH,MAAM,KAAK,GAAG,WAAW,CAAC,KAAmB,CAAC;YAC9C,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;gBACvB,MAAM,QAAQ,GAAG,IAAI,CAAC,WAAW,CAAC,EAAE,KAAK,EAAE,GAAG,EAAE,CAAC,CAAC;gBAClD,MAAM,MAAM,GAAG,IAAA,qBAAY,EAAC,QAAQ,CAAC,KAAK,EAAE,QAAQ,CAAC,MAAM,CAAC,CAAC;gBAC7D,MAAM,OAAO,GAAG,MAAM,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC;gBAExC,MAAM,IAAI,CAAC,MAAM,CAAC,EAAE,aAAa,EAAE,OAAO,EAAE,QAAQ,EAAE,CAAC,CAAC,OAAO,CAAC;gBAEhE,MAAM,WAAW,GAAG,MAAM,IAAA,qBAAa,EAAC,MAAM,CAAC,CAAC;gBAEhD,YAAY,CAAC,OAAO,GAAG,CAAC,CAAC,GAAG,MAAM,QAAQ,CAAC,WAAW,CAAC,CAAC;YAC1D,CAAC;iBAAM,CAAC;gBACN,MAAM,QAAQ,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;gBAC/D,YAAY,CAAC,OAAO,GAAG,CAAC,CAAC,GAAG,MAAM,QAAQ,CAAC,QAAQ,CAAC,CAAC;YACvD,CAAC;QACH,CAAC,CAAC,CACH,CAAC;IACJ,CAAC;IACD,MAAM,OAAO,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC;IAChC,OAAO,YAAY,CAAC;AACtB,CAAC,CAAC,CAAC;AAML;;;;;;;;;;;;;;;;GAgBG;AAEI,MAAM,QAAQ,GAAG,KAAK,EAC3B,KAAyC,EACzC,OAAsB,EACtB,WAAgC,uBAA8C,EAC9E,EAAE;IACF,IAAI,OAAO,QAAQ,KAAK,UAAU,EAAE,CAAC;QACnC,MAAM,IAAI,KAAK,CAAC,0BAA0B,OAAO,QAAQ,EAAE,CAAC,CAAC;IAC/D,CAAC;IACD,IAAI,OAAO,KAAK,KAAK,QAAQ,EAAE,CAAC;QAC9B,MAAM,UAAU,GAAG,MAAM,IAAA,mBAAQ,EAAC,KAAK,EAAE,EAAE,CAAC,CAAC;QAC7C,MAAM,IAAI,GAAG,IAAI,UAAU,CAAC,UAAU,CAAC,CAAC;QACxC,OAAO,kBAAkB,CAAC,EAAE,IAAI,EAAE,GAAG,OAAO,EAAE,EAAE,QAAQ,CAAC,CAAC;IAC5D,CAAC;IACD,IAAI,MAAM,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,CAAC;QAC3B,MAAM,IAAI,GAAG,IAAI,UAAU,CAAC,KAAK,CAAC,CAAC;QACnC,OAAO,kBAAkB,CAAC,EAAE,IAAI,EAAE,GAAG,OAAO,EAAE,EAAE,QAAQ,CAAC,CAAC;IAC5D,CAAC;IACD,IAAI,KAAK,YAAY,UAAU,EAAE,CAAC;QAChC,OAAO,kBAAkB,CAAC,EAAE,IAAI,EAAE,KAAK,EAAE,GAAG,OAAO,EAAE,EAAE,QAAQ,CAAC,CAAC;IACnE,CAAC;IACD,IAAI,KAAK,YAAY,GAAG,EAAE,CAAC;QACzB,OAAO,kBAAkB,CAAC,EAAE,GAAG,EAAE,KAAK,EAAE,GAAG,OAAO,EAAE,EAAE,QAAQ,CAAC,CAAC;IAClE,CAAC;IACD,MAAM,IAAI,KAAK,CAAC,wBAAwB,OAAO,KAAK,EAAE,CAAC,CAAC;AAC1D,CAAC,CAAC;AAxBW,QAAA,QAAQ,YAwBnB"}
|
package/dist/pdf2image.js
CHANGED
|
@@ -4,6 +4,7 @@ exports.pdf2image = void 0;
|
|
|
4
4
|
/* eslint-disable no-underscore-dangle */
|
|
5
5
|
const promises_1 = require("node:fs/promises");
|
|
6
6
|
const canvas_1 = require("canvas");
|
|
7
|
+
const utils_1 = require("./utils");
|
|
7
8
|
const parsePdfFileBuffer = async (options) => import('pdfjs-dist/legacy/build/pdf.mjs').then(async (pdfjsLib) => {
|
|
8
9
|
const loadingTask = pdfjsLib.getDocument({
|
|
9
10
|
...options,
|
|
@@ -15,11 +16,11 @@ const parsePdfFileBuffer = async (options) => import('pdfjs-dist/legacy/build/pd
|
|
|
15
16
|
const pagePromises = [];
|
|
16
17
|
for (let pageNum = 1; pageNum <= numPages; pageNum += 1) {
|
|
17
18
|
pagePromises.push(pdfDocument.getPage(pageNum).then(async (page) => {
|
|
18
|
-
const viewport = page.getViewport({ scale:
|
|
19
|
+
const viewport = page.getViewport({ scale: 2.0 });
|
|
19
20
|
const canvas = (0, canvas_1.createCanvas)(viewport.width, viewport.height);
|
|
20
21
|
const context = canvas.getContext('2d');
|
|
21
22
|
await page.render({ canvasContext: context, viewport }).promise;
|
|
22
|
-
const imageBuffer =
|
|
23
|
+
const imageBuffer = await (0, utils_1.toBufferAsync)(canvas);
|
|
23
24
|
pageContents[pageNum - 1] = imageBuffer;
|
|
24
25
|
}));
|
|
25
26
|
}
|
package/dist/pdf2image.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"pdf2image.js","sourceRoot":"","sources":["../src/pdf2image.ts"],"names":[],"mappings":";;;AAAA,yCAAyC;AACzC,+CAA4C;AAE5C,mCAAsC;AAItC,MAAM,kBAAkB,GAAG,KAAK,EAAE,OAA+B,EAAE,EAAE,CACnE,MAAM,CAAC,iCAAiC,CAAC,CAAC,IAAI,CAAC,KAAK,EAAE,QAAQ,EAAE,EAAE;IAChE,MAAM,WAAW,GAAG,QAAQ,CAAC,WAAW,CAAC;QACvC,GAAG,OAAO;QACV,SAAS,EAAE,CAAC,EAAE,wBAAwB;KACvC,CAAC,CAAC;IAEH,MAAM,WAAW,GAAG,MAAM,WAAW,CAAC,OAAO,CAAC;IAE9C,MAAM,EAAE,QAAQ,EAAE,GAAG,WAAW,CAAC;IACjC,MAAM,YAAY,GAAa,IAAI,KAAK,CAAS,QAAQ,CAAC,CAAC,IAAI,CAC7D,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAChB,CAAC;IACF,MAAM,YAAY,GAAmC,EAAE,CAAC;IAExD,KAAK,IAAI,OAAO,GAAG,CAAC,EAAE,OAAO,IAAI,QAAQ,EAAE,OAAO,IAAI,CAAC,EAAE,CAAC;QACxD,YAAY,CAAC,IAAI,CACf,WAAW,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,KAAK,EAAE,IAAI,EAAE,EAAE;YAC/C,MAAM,QAAQ,GAAG,IAAI,CAAC,WAAW,CAAC,EAAE,KAAK,EAAE,GAAG,EAAE,CAAC,CAAC;YAClD,MAAM,MAAM,GAAG,IAAA,qBAAY,EAAC,QAAQ,CAAC,KAAK,EAAE,QAAQ,CAAC,MAAM,CAAC,CAAC;YAC7D,MAAM,OAAO,GAAG,MAAM,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC;YAExC,MAAM,IAAI,CAAC,MAAM,CAAC,EAAE,aAAa,EAAE,OAAO,EAAE,QAAQ,EAAE,CAAC,CAAC,OAAO,CAAC;
|
|
1
|
+
{"version":3,"file":"pdf2image.js","sourceRoot":"","sources":["../src/pdf2image.ts"],"names":[],"mappings":";;;AAAA,yCAAyC;AACzC,+CAA4C;AAE5C,mCAAsC;AAItC,2CAAgD;AAEhD,MAAM,kBAAkB,GAAG,KAAK,EAAE,OAA+B,EAAE,EAAE,CACnE,MAAM,CAAC,iCAAiC,CAAC,CAAC,IAAI,CAAC,KAAK,EAAE,QAAQ,EAAE,EAAE;IAChE,MAAM,WAAW,GAAG,QAAQ,CAAC,WAAW,CAAC;QACvC,GAAG,OAAO;QACV,SAAS,EAAE,CAAC,EAAE,wBAAwB;KACvC,CAAC,CAAC;IAEH,MAAM,WAAW,GAAG,MAAM,WAAW,CAAC,OAAO,CAAC;IAE9C,MAAM,EAAE,QAAQ,EAAE,GAAG,WAAW,CAAC;IACjC,MAAM,YAAY,GAAa,IAAI,KAAK,CAAS,QAAQ,CAAC,CAAC,IAAI,CAC7D,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAChB,CAAC;IACF,MAAM,YAAY,GAAmC,EAAE,CAAC;IAExD,KAAK,IAAI,OAAO,GAAG,CAAC,EAAE,OAAO,IAAI,QAAQ,EAAE,OAAO,IAAI,CAAC,EAAE,CAAC;QACxD,YAAY,CAAC,IAAI,CACf,WAAW,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,KAAK,EAAE,IAAI,EAAE,EAAE;YAC/C,MAAM,QAAQ,GAAG,IAAI,CAAC,WAAW,CAAC,EAAE,KAAK,EAAE,GAAG,EAAE,CAAC,CAAC;YAClD,MAAM,MAAM,GAAG,IAAA,qBAAY,EAAC,QAAQ,CAAC,KAAK,EAAE,QAAQ,CAAC,MAAM,CAAC,CAAC;YAC7D,MAAM,OAAO,GAAG,MAAM,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC;YAExC,MAAM,IAAI,CAAC,MAAM,CAAC,EAAE,aAAa,EAAE,OAAO,EAAE,QAAQ,EAAE,CAAC,CAAC,OAAO,CAAC;YAEhE,MAAM,WAAW,GAAG,MAAM,IAAA,qBAAa,EAAC,MAAM,CAAC,CAAC;YAChD,YAAY,CAAC,OAAO,GAAG,CAAC,CAAC,GAAG,WAAW,CAAC;QAC1C,CAAC,CAAC,CACH,CAAC;IACJ,CAAC;IACD,MAAM,OAAO,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC;IAChC,OAAO,YAAY,CAAC;AACtB,CAAC,CAAC,CAAC;AAML;;;;;;;;;;;;;;;GAeG;AACI,MAAM,SAAS,GAAG,KAAK,EAC5B,KAAyC,EACzC,OAAsB,EACtB,EAAE;IACF,IAAI,OAAO,KAAK,KAAK,QAAQ,EAAE,CAAC;QAC9B,MAAM,UAAU,GAAG,MAAM,IAAA,mBAAQ,EAAC,KAAK,EAAE,EAAE,CAAC,CAAC;QAC7C,MAAM,IAAI,GAAG,IAAI,UAAU,CAAC,UAAU,CAAC,CAAC;QACxC,OAAO,kBAAkB,CAAC,EAAE,IAAI,EAAE,GAAG,OAAO,EAAE,CAAC,CAAC;IAClD,CAAC;IACD,IAAI,MAAM,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,CAAC;QAC3B,MAAM,IAAI,GAAG,IAAI,UAAU,CAAC,KAAK,CAAC,CAAC;QACnC,OAAO,kBAAkB,CAAC,EAAE,IAAI,EAAE,GAAG,OAAO,EAAE,CAAC,CAAC;IAClD,CAAC;IACD,IAAI,KAAK,YAAY,UAAU,EAAE,CAAC;QAChC,OAAO,kBAAkB,CAAC,EAAE,IAAI,EAAE,KAAK,EAAE,GAAG,OAAO,EAAE,CAAC,CAAC;IACzD,CAAC;IACD,IAAI,KAAK,YAAY,GAAG,EAAE,CAAC;QACzB,OAAO,kBAAkB,CAAC,EAAE,GAAG,EAAE,KAAK,EAAE,GAAG,OAAO,EAAE,CAAC,CAAC;IACxD,CAAC;IACD,MAAM,IAAI,KAAK,CAAC,wBAAwB,OAAO,KAAK,EAAE,CAAC,CAAC;AAC1D,CAAC,CAAC;AApBW,QAAA,SAAS,aAoBpB"}
|
package/dist/utils.js
ADDED
|
@@ -0,0 +1,27 @@
|
|
|
1
|
+
"use strict";
|
|
2
|
+
Object.defineProperty(exports, "__esModule", { value: true });
|
|
3
|
+
exports.toBufferAsync = toBufferAsync;
|
|
4
|
+
exports.imageDataToGrayscale = imageDataToGrayscale;
|
|
5
|
+
const util_1 = require("util");
|
|
6
|
+
function toBufferAsync(canvas) {
|
|
7
|
+
return (0, util_1.promisify)(canvas.toBuffer.bind(canvas))();
|
|
8
|
+
}
|
|
9
|
+
// Experimental method to switch context to grayscale, usefull for next proccesing, for example OCR
|
|
10
|
+
function imageDataToGrayscale(canvas, context) {
|
|
11
|
+
// Convert image to grayscale manually
|
|
12
|
+
const imageData = context.getImageData(0, 0, canvas.width, canvas.height);
|
|
13
|
+
const pixels = imageData.data;
|
|
14
|
+
for (let i = 0; i < pixels.length; i += 4) {
|
|
15
|
+
const r = pixels[i] || 0;
|
|
16
|
+
const g = pixels[i + 1] || 0;
|
|
17
|
+
const b = pixels[i + 2] || 0;
|
|
18
|
+
// Grayscale formula: (0.3 * R) + (0.59 * G) + (0.11 * B)
|
|
19
|
+
const grayscale = 0.3 * r + 0.59 * g + 0.11 * b;
|
|
20
|
+
pixels[i] = grayscale; // Red channel
|
|
21
|
+
pixels[i + 1] = grayscale; // Green channel
|
|
22
|
+
pixels[i + 2] = grayscale; // Blue channel
|
|
23
|
+
// Alpha channel remains unchanged (pixels[i + 3])
|
|
24
|
+
}
|
|
25
|
+
context.putImageData(imageData, 0, 0);
|
|
26
|
+
}
|
|
27
|
+
//# sourceMappingURL=utils.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"utils.js","sourceRoot":"","sources":["../src/utils.ts"],"names":[],"mappings":";;AAIA,sCAEC;AAGD,oDAuBC;AAhCD,+BAAiC;AAIjC,SAAgB,aAAa,CAAC,MAAc;IAC1C,OAAO,IAAA,gBAAS,EAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,EAAqB,CAAC;AACtE,CAAC;AAED,mGAAmG;AACnG,SAAgB,oBAAoB,CAClC,MAAc,EACd,OAAiC;IAEjC,sCAAsC;IACtC,MAAM,SAAS,GAAG,OAAO,CAAC,YAAY,CAAC,CAAC,EAAE,CAAC,EAAE,MAAM,CAAC,KAAK,EAAE,MAAM,CAAC,MAAM,CAAC,CAAC;IAC1E,MAAM,MAAM,GAAG,SAAS,CAAC,IAAI,CAAC;IAE9B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;QAC1C,MAAM,CAAC,GAAG,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;QACzB,MAAM,CAAC,GAAG,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC;QAC7B,MAAM,CAAC,GAAG,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC;QAE7B,yDAAyD;QACzD,MAAM,SAAS,GAAG,GAAG,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,CAAC;QAEhD,MAAM,CAAC,CAAC,CAAC,GAAG,SAAS,CAAC,CAAC,cAAc;QACrC,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,SAAS,CAAC,CAAC,gBAAgB;QAC3C,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,SAAS,CAAC,CAAC,eAAe;QAC1C,kDAAkD;IACpD,CAAC;IAED,OAAO,CAAC,YAAY,CAAC,SAAS,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC;AACxC,CAAC"}
|