tesseract.js - npm Package Compare versions

Comparing version 4.0.6 to 4.1.0

package.json

 {
   "name": "tesseract.js",
-  "version": "4.0.6",
+  "version": "4.1.0",
   "description": "Pure Javascript Multilingual OCR",
@@ -66,3 +66,2 @@ "main": "src/index.js",
     "bmp-js": "^0.1.0",
-    "file-type": "^12.4.2",
     "idb-keyval": "^6.2.0",
@@ -74,3 +73,2 @@ "is-electron": "^2.2.2",
     "regenerator-runtime": "^0.13.3",
-    "resolve-url": "^0.2.1",
     "tesseract.js-core": "^4.0.4",
@@ -77,0 +75,0 @@ "wasm-feature-detect": "^1.2.11",

src/index.d.ts

@@ -73,2 +73,3 @@ declare namespace Tesseract {
     blocks: boolean;
+    layoutBlocks: boolean;
     hocr: boolean;
@@ -149,3 +150,3 @@ tsv: boolean;
   type ImageLike = string | HTMLImageElement | HTMLCanvasElement | HTMLVideoElement
-    | CanvasRenderingContext2D | File | Blob | ImageData | Buffer;
+    | CanvasRenderingContext2D | File | Blob | ImageData | Buffer | OffscreenCanvas;
   interface Block {
@@ -152,0 +153,0 @@ paragraphs: Paragraph[];

src/utils/resolvePaths.js

 const isBrowser = require('./getEnvironment')('type') === 'browser';
-const resolveURL = isBrowser ? require('resolve-url') : s => s; // eslint-disable-line
 
+const resolveURL = isBrowser ? s => (new URL(s, window.location.href)).href : s => s; // eslint-disable-line
+
 module.exports = (options) => {
@@ -5,0 +6,0 @@ const opts = { ...options };

src/worker-script/constants/defaultOutput.js

@@ -8,2 +8,3 @@ /*
   blocks: true,
+  layoutBlocks: false,
   hocr: true,
@@ -18,2 +19,3 @@ tsv: true,
   imageBinary: false,
+  debug: false,
 };

src/worker-script/index.js

@@ -11,6 +11,5 @@ /**
 require('regenerator-runtime/runtime');
-const fileType = require('file-type');
 const isURL = require('is-url');
 const dump = require('./utils/dump');
-const isWebWorker = require('../utils/getEnvironment')('type') === 'webworker';
+const env = require('../utils/getEnvironment')('type');
 const setImage = require('./utils/setImage');
@@ -93,2 +92,4 @@ const defaultParams = require('./constants/defaultParams');
 
+    // Check for existing .traineddata file in cache
+    // This automatically fails if cacheMethod is set to 'refresh' or 'none'
     try {
@@ -103,2 +104,3 @@ const _data = await readCache(`${cachePath || '.'}/${lang}.traineddata`);
       }
+    // Attempt to fetch new .traineddata file
     } catch (e) {
@@ -110,9 +112,13 @@ newData = true;
 
-        if (isURL(langPath) || langPath.startsWith('moz-extension://') || langPath.startsWith('chrome-extension://') || langPath.startsWith('file://')) { /** When langPath is an URL */
+        // For Node.js, langPath may be a URL or local file path
+        // The is-url package is used to tell the difference
+        // For the browser version, langPath is assumed to be a URL
+        if (env !== 'node' || isURL(langPath) || langPath.startsWith('moz-extension://') || langPath.startsWith('chrome-extension://') || langPath.startsWith('file://')) { /** When langPath is an URL */
           path = langPath.replace(/\/$/, '');
         }
 
+        // langPath is a URL, fetch from server
         if (path !== null) {
           const fetchUrl = `${path}/${lang}.traineddata${gzip ? '.gz' : ''}`;
-          const resp = await (isWebWorker ? fetch : adapter.fetch)(fetchUrl);
+          const resp = await (env === 'webworker' ? fetch : adapter.fetch)(fetchUrl);
           if (!resp.ok) {
@@ -122,2 +128,5 @@ throw Error(`Network error while fetching ${fetchUrl}. Response code: ${resp.status}`);
           data = await resp.arrayBuffer();
+
+        // langPath is a local file, read .traineddata from local filesystem
+        // (adapter.readCache is a generic file read function in Node.js version)
         } else {
@@ -133,4 +142,6 @@ data = await adapter.readCache(`${langPath}/${lang}.traineddata${gzip ? '.gz' : ''}`);
 
-    const type = fileType(data);
-    if (typeof type !== 'undefined' && type.mime === 'application/gzip') {
+    // Check for gzip magic numbers (1F and 8B in hex)
+    const isGzip = (data[0] === 31 && data[1] === 139) || (data[1] === 31 && data[0] === 139);
+
+    if (isGzip) {
       data = adapter.gunzip(data);
@@ -264,3 +275,3 @@ }
 
-  const nonRecOutputs = ['imageColor', 'imageGrey', 'imageBinary'];
+  const nonRecOutputs = ['imageColor', 'imageGrey', 'imageBinary', 'layoutBlocks'];
   let recOutputCount = 0;
@@ -277,3 +288,4 @@ for (const prop of Object.keys(output)) {
   }
-  return { workingOutput, recOutputCount };
+  const skipRecognition = recOutputCount === 0;
+  return { workingOutput, skipRecognition };
 };
@@ -313,3 +325,3 @@
 
-    const { workingOutput, recOutputCount } = processOutput(output);
+    const { workingOutput, skipRecognition } = processOutput(output);
 
@@ -364,5 +376,8 @@ // When the auto-rotate option is True, setImage is called with no angle,
 
-    if (recOutputCount > 0) {
+    if (!skipRecognition) {
       api.Recognize(null);
     } else {
+      if (output.layoutBlocks) {
+        api.AnalyseLayout();
+      }
       log('Skipping recognition: all output options requiring recognition are disabled.');
@@ -372,3 +387,3 @@ }
     const { pdfTextOnly } = options;
-    const result = dump(TessModule, api, workingOutput, { pdfTitle, pdfTextOnly });
+    const result = dump(TessModule, api, workingOutput, { pdfTitle, pdfTextOnly, skipRecognition });
     result.rotateRadians = rotateRadiansFinal;
@@ -375,0 +390,0 @@

src/worker-script/utils/dump.js

@@ -82,3 +82,6 @@ /**
 
-  if (output.blocks) {
+  // If output.layoutBlocks is true and options.skipRecognition is true,
+  // the user wants layout data but text recognition has not been run.
+  // In this case, fields that require text recognition are skipped.
+  if (output.blocks || output.layoutBlocks) {
     ri.Begin();
@@ -106,4 +109,4 @@ do {
           paragraphs: [],
-          text: ri.GetUTF8Text(RIL_BLOCK),
-          confidence: ri.Confidence(RIL_BLOCK),
+          text: !options.skipRecognition ? ri.GetUTF8Text(RIL_BLOCK) : null,
+          confidence: !options.skipRecognition ? ri.Confidence(RIL_BLOCK) : null,
           baseline: ri.getBaseline(RIL_BLOCK),
@@ -119,4 +122,4 @@ bbox: ri.getBoundingBox(RIL_BLOCK),
           lines: [],
-          text: ri.GetUTF8Text(RIL_PARA),
-          confidence: ri.Confidence(RIL_PARA),
+          text: !options.skipRecognition ? ri.GetUTF8Text(RIL_PARA) : null,
+          confidence: !options.skipRecognition ? ri.Confidence(RIL_PARA) : null,
           baseline: ri.getBaseline(RIL_PARA),
@@ -131,4 +134,4 @@ bbox: ri.getBoundingBox(RIL_PARA),
           words: [],
-          text: ri.GetUTF8Text(RIL_TEXTLINE),
-          confidence: ri.Confidence(RIL_TEXTLINE),
+          text: !options.skipRecognition ? ri.GetUTF8Text(RIL_TEXTLINE) : null,
+          confidence: !options.skipRecognition ? ri.Confidence(RIL_TEXTLINE) : null,
           baseline: ri.getBaseline(RIL_TEXTLINE),
@@ -146,4 +149,4 @@ bbox: ri.getBoundingBox(RIL_TEXTLINE),
 
-          text: ri.GetUTF8Text(RIL_WORD),
-          confidence: ri.Confidence(RIL_WORD),
+          text: !options.skipRecognition ? ri.GetUTF8Text(RIL_WORD) : null,
+          confidence: !options.skipRecognition ? ri.Confidence(RIL_WORD) : null,
           baseline: ri.getBaseline(RIL_WORD),
@@ -170,4 +173,4 @@ bbox: ri.getBoundingBox(RIL_WORD),
           word.choices.push({
-            text: wc.GetUTF8Text(),
-            confidence: wc.Confidence(),
+            text: !options.skipRecognition ? wc.GetUTF8Text() : null,
+            confidence: !options.skipRecognition ? wc.Confidence() : null,
           });
@@ -188,4 +191,4 @@ } while (wc.Next());
           image: null,
-          text: ri.GetUTF8Text(RIL_SYMBOL),
-          confidence: ri.Confidence(RIL_SYMBOL),
+          text: !options.skipRecognition ? ri.GetUTF8Text(RIL_SYMBOL) : null,
+          confidence: !options.skipRecognition ? ri.Confidence(RIL_SYMBOL) : null,
           baseline: ri.getBaseline(RIL_SYMBOL),
@@ -201,4 +204,4 @@ bbox: ri.getBoundingBox(RIL_SYMBOL),
           symbol.choices.push({
-            text: ci.GetUTF8Text(),
-            confidence: ci.Confidence(),
+            text: !options.skipRecognition ? ci.GetUTF8Text() : null,
+            confidence: !options.skipRecognition ? ci.Confidence() : null,
           });
@@ -223,4 +226,5 @@ } while (ci.Next());
     imageBinary: output.imageBinary ? getImage(imageType.BINARY) : null,
-    confidence: api.MeanTextConf(),
-    blocks: output.blocks ? blocks : null,
+    confidence: !options.skipRecognition ? api.MeanTextConf() : null,
+    blocks: output.blocks && !options.skipRecognition ? blocks : null,
+    layoutBlocks: output.layoutBlocks && options.skipRecognition ? blocks : null,
     psm: enumToString(api.GetPageSegMode(), 'PSM'),
@@ -227,0 +231,0 @@ oem: enumToString(api.oem(), 'OEM'),

src/worker-script/utils/setImage.js

 const bmp = require('bmp-js');
-const fileType = require('file-type');
 
@@ -12,3 +11,4 @@ /**
 module.exports = (TessModule, api, image, angle = 0) => {
-  const type = fileType(image);
+  // Check for bmp magic numbers (42 and 4D in hex)
+  const isBmp = (image[0] === 66 && image[1] === 77) || (image[1] === 66 && image[0] === 77);
 
@@ -22,3 +22,3 @@ const exif = image.slice(0, 500).toString().match(/\x01\x12\x00\x03\x00\x00\x00\x01\x00(.)/)?.[1]?.charCodeAt(0) || 1;
   //  */
-  if (type && type.mime === 'image/bmp') {
+  if (isBmp) {
     // Not sure what this line actually does, but removing breaks the function
@@ -25,0 +25,0 @@ const buf = Buffer.from(Array.from({ ...image, length: Object.keys(image).length }));

src/worker/browser/defaultOptions.js

@@ -1,2 +0,2 @@
-const resolveURL = require('resolve-url');
+const resolveURL = (s) => (new URL(s, window.location.href)).href;
 const { version } = require('../../../package.json');
@@ -3,0 +3,0 @@ const defaultOptions = require('../../constants/defaultOptions');

src/worker/browser/loadImage.js

@@ -1,3 +0,1 @@
-const resolveURL = require('resolve-url');
-
 /**
@@ -43,6 +41,6 @@ * readFromBlobOrFile
     } else {
-      const resp = await fetch(resolveURL(image));
+      const resp = await fetch(image);
       data = await resp.arrayBuffer();
     }
-  } else if (image instanceof HTMLElement) {
+  } else if (typeof HTMLElement !== 'undefined' && image instanceof HTMLElement) {
     if (image.tagName === 'IMG') {
@@ -62,2 +60,5 @@ data = await loadImage(image.src);
     }
+  } else if (typeof OffscreenCanvas !== 'undefined' && image instanceof OffscreenCanvas) {
+    const blob = await image.convertToBlob();
+    data = await readFromBlobOrFile(blob);
   } else if (image instanceof File || image instanceof Blob) {
@@ -64,0 +65,0 @@ data = await readFromBlobOrFile(image);

New alerts

License Policy Violation

License

This package is not allowed per your license policy. Review the package's license to ensure compliance.

Found 1 instance in 1 package

Fixed alerts

License Policy Violation

License

This package is not allowed per your license policy. Review the package's license to ensure compliance.

Found 1 instance in 1 package

Improved metrics

Dependency count

10

decreased by-16.67%

Worsened metrics

Total package byte prevSize

1434549

decreased by-5.16%

Lines of code

3331

decreased by-1.65%

Dependency changes

• - Removedfile-type@^12.4.2

• - Removedresolve-url@^0.2.1

• - Removedfile-type@12.4.2(transitive)

• - Removedresolve-url@0.2.1(transitive)