Removed various depreciated functions, including default parameters to fix #975 (#981)

Balearica · web-flow · commit 15ed5973ebe7 · 2024-12-24T00:42:32.000-08:00
diff --git a/src/createWorker.js b/src/createWorker.js
@@ -104,10 +104,6 @@ module.exports = async (langs = 'eng', oem = OEM.LSTM_ONLY, _options = {}, confi
     }))
   );
 
-  const loadLanguage = () => (
-    console.warn('`loadLanguage` is depreciated and should be removed from code (workers now come with language pre-loaded)')
-  );
-
   const loadLanguageInternal = (_langs, jobId) => startJob(createJob({
     id: jobId,
     action: 'loadLanguage',
@@ -125,10 +121,6 @@ module.exports = async (langs = 'eng', oem = OEM.LSTM_ONLY, _options = {}, confi
     },
   }));
 
-  const initialize = () => (
-    console.warn('`initialize` is depreciated and should be removed from code (workers now come pre-initialized)')
-  );
-
   const initializeInternal = (_langs, _oem, _config, jobId) => (
     startJob(createJob({
       id: jobId,
@@ -182,15 +174,6 @@ module.exports = async (langs = 'eng', oem = OEM.LSTM_ONLY, _options = {}, confi
     }))
   );
 
-  const getPDF = (title = 'Tesseract OCR Result', textonly = false, jobId) => {
-    console.log('`getPDF` function is depreciated. `recognize` option `savePDF` should be used instead.');
-    return startJob(createJob({
-      id: jobId,
-      action: 'getPDF',
-      payload: { title, textonly },
-    }));
-  };
-
   const detect = async (image, jobId) => {
     if (lstmOnlyCore) throw Error('`worker.detect` requires Legacy model, which was not loaded.');
 
@@ -224,8 +207,6 @@ module.exports = async (langs = 'eng', oem = OEM.LSTM_ONLY, _options = {}, confi
       let d = data;
       if (action === 'recognize') {
         d = circularize(data);
-      } else if (action === 'getPDF') {
-        d = Array.from({ ...data, length: Object.keys(data).length });
       }
       promises[promiseId].resolve({ jobId, data: d });
       delete promises[promiseId];
@@ -251,12 +232,9 @@ module.exports = async (langs = 'eng', oem = OEM.LSTM_ONLY, _options = {}, confi
     readText,
     removeFile,
     FS,
-    loadLanguage,
-    initialize,
     reinitialize,
     setParameters,
     recognize,
-    getPDF,
     detect,
     terminate,
   };
diff --git a/src/index.d.ts b/src/index.d.ts
@@ -26,7 +26,6 @@ declare namespace Tesseract {
     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>
   }
 
   interface Lang {
@@ -71,11 +70,6 @@ declare namespace Tesseract {
     tessedit_char_blacklist: string
     preserve_interword_spaces: string
     user_defined_dpi: string
-    tessjs_create_hocr: string
-    tessjs_create_tsv: string
-    tessjs_create_box: string
-    tessjs_create_unlv: string
-    tessjs_create_osd: string
     [propName: string]: any
   }
   interface OutputFormats {
@@ -108,10 +102,6 @@ declare namespace Tesseract {
     jobId: string
     data: Page
   }
-  interface GetPDFResult {
-    jobId: string
-    data: number[]
-  }
   interface DetectResult {
     jobId: string
     data: DetectData
diff --git a/src/worker-script/constants/defaultParams.js b/src/worker-script/constants/defaultParams.js
diff --git a/src/worker-script/index.js b/src/worker-script/index.js
@@ -12,7 +12,6 @@ const isURL = require('is-url');
 const dump = require('./utils/dump');
 const env = require('../utils/getEnvironment')('type');
 const setImage = require('./utils/setImage');
-const defaultParams = require('./constants/defaultParams');
 const defaultOutput = require('./constants/defaultOutput');
 const { log, setLogging } = require('../utils/log');
 const PSM = require('../constants/PSM');
@@ -27,7 +26,7 @@ let TessModule;
 let api = null;
 let latestJob;
 let adapter = {};
-let params = defaultParams;
+let params = {};
 let loadLanguageLangsWorker;
 let loadLanguageOptionsWorker;
 let dataFromCache = false;
@@ -305,8 +304,6 @@ const initialize = async ({
       res.reject('initialization failed');
     }
 
-    params = defaultParams;
-    await setParameters({ payload: { params } });
     res.progress({
       workerId, status: statusText, progress: 1,
     });
@@ -316,31 +313,10 @@ const initialize = async ({
   }
 };
 
-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');
-};
-
-const getPDF = async ({ payload: { title, textonly } }, res) => {
-  res.resolve(getPDFInternal(title, textonly));
-};
-
 // 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', 'layoutBlocks', 'debug'];
   let recOutputCount = 0;
@@ -546,7 +522,6 @@ exports.dispatchHandlers = (packet, send) => {
     initialize,
     setParameters,
     recognize,
-    getPDF,
     detect,
     terminate,
   })[packet.action](packet, res)
