Added OutputFormats option/interface for setting output

examples/browser/basic.html

@@ -8,6 +8,7 @@
       const recognize = async ({ target: { files }  }) => {
         const { data: { text } } = await Tesseract.recognize(files[0], 'eng', {
           corePath: '../../node_modules/tesseract.js-core/tesseract-core.wasm.js',
+          workerPath: "/dist/worker.dev.js",
           logger: m => console.log(m),
         });
         console.log(text);

examples/browser/download-pdf.html

@@ -21,7 +21,7 @@
       const recognize = async ({ target: { files }  }) => {
         await worker.loadLanguage('eng');
         await worker.initialize('eng');
-        const res = await worker.recognize(files[0], {savePDF: true});
+        const res = await worker.recognize(files[0],{pdfTitle: "Example PDF"},{pdf: true});
         pdf = res.data.pdf;
         const text = res.data.text;
         const board = document.getElementById('board');

examples/browser/image-processing.html

@@ -45,8 +45,8 @@
       await worker.initialize('eng');
 
       await worker.initialize();
-      const ret = await worker.recognize(files[0], { saveImageOriginal: true, saveImageGrey: true, saveImageBinary: true, rotateAuto: true })
-      document.getElementById("imgOriginal").src = ret.data.imageOriginal;
+      const ret = await worker.recognize(files[0], {rotateAuto: true}, {imageColor: true, imageGrey: true, imageBinary: true});
+      document.getElementById("imgOriginal").src = ret.data.imageColor;
       document.getElementById("imgGrey").src = ret.data.imageGrey;
       document.getElementById("imgBinary").src = ret.data.imageBinary;
 

examples/node/download-pdf.js

@@ -12,7 +12,7 @@ console.log(`Recognizing ${image}`);
   const worker = await createWorker();
   await worker.loadLanguage('eng');
   await worker.initialize('eng');
-  const { data: { text, pdf } } = await worker.recognize(image, {savePDF: true});
+  const { data: { text, pdf } } = await worker.recognize(image, {pdfTitle: "Example PDF"}, {pdf: true});
   console.log(text);
   fs.writeFileSync('tesseract-ocr-result.pdf', Buffer.from(pdf));
   console.log('Generate PDF: tesseract-ocr-result.pdf');

examples/node/scheduler.js

@@ -0,0 +1,26 @@
+const { createWorker, createScheduler } = require('../../');
+
+const scheduler = createScheduler();
+
+// Creates worker and adds to scheduler
+const workerGen = async () => {
+  const worker = createWorker({cachePath: "."});
+  await worker.load();
+  await worker.loadLanguage('eng');
+  await worker.initialize('eng');
+  scheduler.addWorker(worker);
+}
+
+const workerN = 4;
+(async () => {
+  const resArr = Array(workerN);
+  for (let i=0; i<workerN; i++) {
+    resArr[i] = workerGen();
+  }
+  await Promise.all(resArr);
+  /** Add 4 recognition jobs */
+  const results = await Promise.all(Array(10).fill(0).map(() => (
+    scheduler.addJob('recognize', 'https://tesseract.projectnaptha.com/img/eng_bw.png').then((x) => console.log(x.data.text))
+  )))
+  await scheduler.terminate(); // It also terminates all workers.
+})();

src/constants/imageType.js

@@ -1,5 +1,5 @@
 module.exports = {
-  ORIGINAL: 0,
+  COLOR: 0,
   GREY: 1,
   BINARY: 2,
 };

src/createWorker.js

@@ -129,19 +129,11 @@ module.exports = async (_options = {}) => {
     }))
   );
 
-  const recognize = async (image, opts = {}, jobId) => (
+  const recognize = async (image, opts = {}, output = {blocks: true, text: true, hocr: true, tsv: true}, jobId) => (
     startJob(createJob({
       id: jobId,
       action: 'recognize',
-      payload: { image: await loadImage(image), options: opts },
-    }))
-  );
-
-  const threshold = async (image, opts = {}, jobId) => (
-    startJob(createJob({
-      id: jobId,
-      action: 'threshold',
-      payload: { image: await loadImage(image), options: opts },
+      payload: { image: await loadImage(image), options: opts, output },
     }))
   );
 
@@ -215,7 +207,6 @@ module.exports = async (_options = {}) => {
     initialize,
     setParameters,
     recognize,
-    threshold,
     getPDF,
     detect,
     terminate,

src/index.d.ts

@@ -23,8 +23,7 @@ declare namespace Tesseract {
     initialize(langs?: string, oem?: OEM, jobId?: string): Promise<ConfigResult>
     setParameters(params: Partial<WorkerParams>, jobId?: string): Promise<ConfigResult>
     getImage(type: imageType): string
-    recognize(image: ImageLike, options?: Partial<RecognizeOptions>, jobId?: string): Promise<RecognizeResult>
-    threshold(image: ImageLike, options?: Partial<RecognizeOptions>, jobId?: string): Promise<RecognizeResult>
+    recognize(image: ImageLike, options?: Partial<RecognizeOptions>, output?: Partial<OutputFormats>, jobId?: string): Promise<RecognizeResult>
     detect(image: ImageLike, jobId?: string): Promise<DetectResult>
     terminate(jobId?: string): Promise<ConfigResult>
     getPDF(title?: string, textonly?: boolean, jobId?: string):Promise<GetPDFResult>
@@ -54,16 +53,25 @@ declare namespace Tesseract {
     tessjs_create_unlv: string
     tessjs_create_osd: string
   }
+  interface OutputFormats {
+    text: boolean;
+    blocks: boolean;
+    hocr: boolean;
+    tsv: boolean;
+    box: boolean;
+    unlv: boolean;
+    osd: boolean;
+    pdf: boolean;
+    imageColor: boolean;
+    imageGrey: boolean;
+    imageBinary: boolean;
+  }
   interface RecognizeOptions {
     rectangle: Rectangle
-    saveImageOriginal: boolean
-    saveImageGrey: boolean
-    saveImageBinary: boolean
-    savePDF: boolean
     pdfTitle: string
     pdfTextOnly: boolean
     rotateAuto: boolean
-    rotateRadians: float
+    rotateRadians: number
   }
   interface ConfigResult {
     jobId: string
@@ -117,7 +125,7 @@ declare namespace Tesseract {
     RAW_LINE = '13'
   }
   const enum imageType {
-    ORIGINAL = 0,
+    COLOR = 0,
     GREY = 1,
     BINARY = 2
   }
@@ -218,7 +226,7 @@ declare namespace Tesseract {
     page: Page;
   }
   interface Page {
-    blocks: Block[];
+    blocks: Block[] | null;
     confidence: number;
     lines: Line[];
     oem: string;
@@ -234,7 +242,7 @@ declare namespace Tesseract {
     box: string | null;
     unlv: string | null;
     sd: string | null;
-    imageOriginal: string | null;
+    imageColor: string | null;
     imageGrey: string | null;
     imageBinary: string | null;
     rotateRadians: number | null;

src/utils/circularize.js

@@ -22,6 +22,7 @@ module.exports = (page) => {
   const words = [];
   const symbols = [];
 
+  if (page.blocks) {
     page.blocks.forEach((block) => {
       block.paragraphs.forEach((paragraph) => {
         paragraph.lines.forEach((line) => {
@@ -47,6 +48,7 @@ module.exports = (page) => {
         ...block, page,
       });
     });
+  }
 
   return {
     ...page, blocks, paragraphs, lines, words, symbols,

src/worker-script/constants/defaultOutput.js

@@ -0,0 +1,17 @@
+/*
+ * default output formats for tesseract.js
+ */
+
+module.exports = {
+  text: true,
+  blocks: true,
+  hocr: true,
+  tsv: true,
+  box: false,
+  unlv: false,
+  osd: false,
+  pdf: false,
+  imageColor: false,
+  imageGrey: false,
+  imageBinary: false
+};

src/worker-script/index.js

@@ -14,8 +14,8 @@ const dump = require('./utils/dump');
 const isWebWorker = require('../utils/getEnvironment')('type') === 'webworker';
 const setImage = require('./utils/setImage');
 const defaultParams = require('./constants/defaultParams');
+const defaultOutput = require('./constants/defaultOutput');
 const { log, setLogging } = require('../utils/log');
-const arrayBufferToBase64 = require('./utils/arrayBufferToBase64');
 const imageType = require('../constants/imageType');
 const PSM = require('../constants/PSM');
 
@@ -214,23 +214,44 @@ const getPDF = async ({ payload: { title, textonly } }, res) => {
   res.resolve(getPDFInternal(title, textonly));
 };
 
-const getImage = (type) => {
-  api.WriteImage(type, '/image.png');
-  const pngBuffer = TessModule.FS.readFile('/image.png');
-  const pngStr = `data:image/png;base64,${arrayBufferToBase64(pngBuffer.buffer)}`;
-  TessModule.FS.unlink('/image.png');
-  return pngStr;
-};
+// Combines default output with user-specified options and
+// counts (1) total output formats requested and (2) outputs that require OCR
+const processOutput = (output) => {
+  const workingOutput = JSON.parse(JSON.stringify(defaultOutput));
+  // Output formats were set using `setParameters` in previous versions
+  // These settings are copied over for compatability
+  if (params.tessjs_create_box === "1") workingOutput.box = true;
+  if (params.tessjs_create_hocr === "1") workingOutput.hocr = true;
+  if (params.tessjs_create_osd === "1") workingOutput.osd = true;
+  if (params.tessjs_create_tsv === "1") workingOutput.tsv = true;
+  if (params.tessjs_create_unlv === "1") workingOutput.unlv = true;
+
+  const nonRecOutputs = ["imageColor", "imageGrey", "imageBinary"];
+  let recOutputCount = 0;
+  for (const prop in output) {
+    workingOutput[prop] = output[prop];
+  }
+  for (const prop in workingOutput) {
+    if (workingOutput[prop]) {
+      if (!nonRecOutputs.includes(prop)) {
+        recOutputCount++;
+      }
+    }
+  }
+  return {workingOutput, recOutputCount}
+}
 
 const recognize = async ({
   payload: {
     image, options: {
-      rectangle: rec, saveImageOriginal, saveImageGrey, saveImageBinary, savePDF, pdfTitle,
+      rectangle: rec, pdfTitle,
       pdfTextOnly, rotateAuto, rotateRadians,
-    },
+    }, output
   },
 }, res) => {
   try {
+    const {workingOutput, recOutputCount} = processOutput(output);
+
     // When the auto-rotate option is True, setImage is called with no angle,
     // then the angle is calculated by Tesseract and then setImage is re-called.
     // Otherwise, setImage is called once using the user-provided rotateRadiansFinal value.
@@ -274,96 +295,14 @@ const recognize = async ({
     if (typeof rec === 'object') {
       api.SetRectangle(rec.left, rec.top, rec.width, rec.height);
     }
-    api.Recognize(null);
-    const result = dump(TessModule, api, params);
-    if (saveImageOriginal) {
-      result.imageOriginal = getImage(imageType.ORIGINAL);
-    } else {
-      result.imageOriginal = null;
-    }
-    if (saveImageGrey) {
-      result.imageGrey = getImage(imageType.GREY);
-    } else {
-      result.imageGrey = null;
-    }
-    if (saveImageBinary) {
-      result.imageBinary = getImage(imageType.BINARY);
-    } else {
-      result.imageBinary = null;
-    }
-    if (savePDF) {
-      result.pdf = getPDFInternal(pdfTitle ?? 'Tesseract OCR Result', pdfTextOnly ?? false);
-    } else {
-      result.pdf = null;
-    }
-    result.rotateRadians = rotateRadiansFinal;
-    res.resolve(result);
-    TessModule._free(ptr);
-  } catch (err) {
-    res.reject(err.toString());
-  }
-};
-
-// `threshold` is similar to `recognize` except it skips the recognition step
-// Useful for getting rotated/binarized images without running recognition
-const threshold = async ({
-  payload: {
-    image, options: {
-      rectangle: rec, saveImageOriginal, saveImageGrey, saveImageBinary, rotateAuto, rotateRadians,
-    },
-  },
-}, res) => {
-  try {
-    let ptr;
-    let rotateRadiansFinal;
-    if (rotateAuto) {
-      const psmInit = api.GetPageSegMode();
-      let psmEdit = false;
-      if (![PSM.AUTO, PSM.AUTO_ONLY, PSM.OSD].includes(psmInit)) {
-        psmEdit = true;
-        api.SetVariable('tessedit_pageseg_mode', String(PSM.AUTO));
-      }
-
-      ptr = setImage(TessModule, api, image);
-      api.FindLines();
-      const rotateRadiansCalc = api.GetAngle();
-
-      // Restore user-provided PSM setting
-      if (psmEdit) {
-        api.SetVariable('tessedit_pageseg_mode', String(psmInit));
-      }
 
-      // Small angles (<0.005 radians/~0.3 degrees) are ignored to save on runtime
-      if (Math.abs(rotateRadiansCalc) >= 0.005) {
-        rotateRadiansFinal = rotateRadiansCalc;
-        ptr = setImage(TessModule, api, image, rotateRadiansFinal);
+    if (recOutputCount > 0) {
+      api.Recognize(null);
     } else {
-        rotateRadiansFinal = 0;
-      }
-    } else {
-      rotateRadiansFinal = rotateRadians || 0;
-      ptr = setImage(TessModule, api, image, rotateRadiansFinal);
+      log(`Skipping recognition: all output options requiring recognition are disabled.`);
     }
     
-    if (typeof rec === 'object') {
-      api.SetRectangle(rec.left, rec.top, rec.width, rec.height);
-    }
-    const result = {};
-    if (saveImageOriginal) {
-      result.imageOriginal = getImage(imageType.ORIGINAL);
-    } else {
-      result.imageOriginal = null;
-    }
-    if (saveImageGrey) {
-      result.imageGrey = getImage(imageType.GREY);
-    } else {
-      result.imageGrey = null;
-    }
-    if (saveImageBinary) {
-      result.imageBinary = getImage(imageType.BINARY);
-    } else {
-      result.imageBinary = null;
-    }
+    const result = dump(TessModule, api, workingOutput, {pdfTitle, pdfTextOnly});
     result.rotateRadians = rotateRadiansFinal;
     res.resolve(result);
     TessModule._free(ptr);
@@ -372,6 +311,7 @@ const threshold = async ({
   }
 };
 
+
 const detect = async ({ payload: { image } }, res) => {
   try {
     const ptr = setImage(TessModule, api, image);
@@ -451,7 +391,6 @@ exports.dispatchHandlers = (packet, send) => {
     initialize,
     setParameters,
     recognize,
-    threshold,
     getPDF,
     detect,
     terminate,

src/worker-script/utils/dump.js

@@ -7,6 +7,8 @@
  * @author Guillermo Webster <gui@mit.edu>
  * @author Jerome Wu <jeromewus@gmail.com>
  */
+const arrayBufferToBase64 = require('./arrayBufferToBase64');
+const imageType = require('../../constants/imageType');
 
 /**
  * deindent
@@ -37,13 +39,7 @@ const deindent = (html) => {
  * @function dump recognition result to a JSON object
  * @access public
  */
-module.exports = (TessModule, api, {
-  tessjs_create_hocr,
-  tessjs_create_tsv,
-  tessjs_create_box,
-  tessjs_create_unlv,
-  tessjs_create_osd,
-}) => {
+module.exports = (TessModule, api, output, options) => {
   const ri = api.GetIterator();
   const {
     RIL_BLOCK,
@@ -65,6 +61,26 @@ module.exports = (TessModule, api, {
       .map((e) => e.slice(prefix.length + 1))[0]
   );
 
+  const getImage = (type) => {
+    api.WriteImage(type, '/image.png');
+    const pngBuffer = TessModule.FS.readFile('/image.png');
+    const pngStr = `data:image/png;base64,${arrayBufferToBase64(pngBuffer.buffer)}`;
+    TessModule.FS.unlink('/image.png');
+    return pngStr;
+  };
+
+  const getPDFInternal = (title, textonly) => {
+    const pdfRenderer = new TessModule.TessPDFRenderer('tesseract-ocr', '/', textonly);
+    pdfRenderer.BeginDocument(title);
+    pdfRenderer.AddImage(api);
+    pdfRenderer.EndDocument();
+    TessModule._free(pdfRenderer);
+
+    return TessModule.FS.readFile('/tesseract-ocr.pdf');
+  };
+
+  if (output.blocks) {
+
     ri.Begin();
     do {
       if (ri.IsAtBeginningOf(RIL_BLOCK)) {
@@ -185,15 +201,21 @@ module.exports = (TessModule, api, {
     } while (ri.Next(RIL_SYMBOL));
     TessModule.destroy(ri);
 
+  }
+
   return {
-    text: api.GetUTF8Text(),
-    hocr: tessjs_create_hocr === '1' ? deindent(api.GetHOCRText()) : null,
-    tsv: tessjs_create_tsv === '1' ? api.GetTSVText() : null,
-    box: tessjs_create_box === '1' ? api.GetBoxText() : null,
-    unlv: tessjs_create_unlv === '1' ? api.GetUNLVText() : null,
-    osd: tessjs_create_osd === '1' ? api.GetOsdText() : null,
+    text: output.text ? api.GetUTF8Text() : null,
+    hocr: output.hocr ? deindent(api.GetHOCRText()) : null,
+    tsv: output.tsv ? api.GetTSVText() : null,
+    box: output.box ? api.GetBoxText() : null,
+    unlv: output.unlv ? api.GetUNLVText() : null,
+    osd: output.osd ? api.GetOsdText() : null,
+    pdf: output.pdf ? getPDFInternal(options.pdfTitle ?? 'Tesseract OCR Result', options.pdfTextOnly ?? false) : null,
+    imageColor: output.imageColor ? getImage(imageType.COLOR) : null,
+    imageGrey: output.imageColor ? getImage(imageType.GREY) : null,
+    imageBinary: output.imageColor ? getImage(imageType.BINARY) : null,
     confidence: api.MeanTextConf(),
-    blocks,
+    blocks: output.blocks ? blocks : null,
     psm: enumToString(api.GetPageSegMode(), 'PSM'),
     oem: enumToString(api.oem(), 'OEM'),
     version: api.Version(),