monocr 0.1.2 → 0.1.3

package/bin/monocr.js

@@ -89,7 +89,7 @@ program
         try {
             const { MonOCR } = require('../src/index');
             const ocr = new MonOCR();
-            await ocr.modelManager.downloadModels();
+            await ocr.modelManager.downloadModel();
         } catch (err) {
             console.error('Error:', err.message);
             process.exit(1);

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "monocr",
-  "version": "0.1.2",
+  "version": "0.1.3",
   "description": "Cross-platform Mon (mnw) language OCR using ONNX Runtime. Supports Node.js.",
   "main": "src/index.js",
   "bin": {

package/src/charset.txt

@@ -0,0 +1 @@
+ !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~ကခဂဃငစဆဇဈဉညဋဌဍဎဏတထဒဓနပဖဗဘမယရလဝသဟဠအဢဣဤဥဦဧဨဩဪါာိီုူေဲဳဴဵံ့း္်ျြွှဿ၀၁၂၃၄၅၆၇၈၉၊။၌၍၎၏ၐၑၓၚၛၜၝၞၟၠၡၢၣၤၥၨၪၰၱၲၳၴၵၷၸၹၺၻၼၾၿႀႄႅႆႇႈႉႊႏ႐႒႓႔႕႘႙ႜႝ႟

package/src/model-manager.js

@@ -9,12 +9,10 @@ class ModelManager {
         // Default cache directory in user's home
         this.cacheDir = path.join(os.homedir(), '.monocr', 'models');
         
-        // HuggingFace model URLs
+        // HuggingFace model URL
         this.baseUrl = 'https://huggingface.co/janakhpon/monocr/resolve/main';
-        this.modelFiles = {
-            model: 'onnx/monocr.onnx',
-            charset: 'charset.txt'
-        };
+        this.modelFileName = 'monocr.onnx';
+        this.hfModelPath = 'onnx/monocr.onnx';
     }
 
     /**
@@ -27,19 +25,17 @@ class ModelManager {
     }
 
     /**
-     * Get local path for a model file
+     * Get local path for the model
      */
-    getLocalPath(fileKey) {
-        return path.join(this.cacheDir, path.basename(this.modelFiles[fileKey]));
+    getModelPath() {
+        return path.join(this.cacheDir, this.modelFileName);
     }
 
     /**
-     * Check if model files exist locally
+     * Check if model exists locally
      */
-    hasModels() {
-        const modelPath = this.getLocalPath('model');
-        const charsetPath = this.getLocalPath('charset');
-        return fs.existsSync(modelPath) && fs.existsSync(charsetPath);
+    hasModel() {
+        return fs.existsSync(this.getModelPath());
     }
 
     /**
@@ -49,43 +45,44 @@ class ModelManager {
         return new Promise((resolve, reject) => {
             const file = fs.createWriteStream(destPath);
             
-            https.get(url, { headers: { 'User-Agent': 'monocr-npm' } }, (response) => {
-                if (response.statusCode === 302 || response.statusCode === 301) {
-                    // Follow redirect
-                    https.get(response.headers.location, (redirectResponse) => {
-                        const totalSize = parseInt(redirectResponse.headers['content-length'], 10);
+            const request = (requestUrl) => {
+                https.get(requestUrl, { headers: { 'User-Agent': 'monocr-npm' } }, (response) => {
+                    if ([301, 302, 307, 308].includes(response.statusCode)) {
+                        let redirectUrl = response.headers.location;
+                        if (!redirectUrl.startsWith('http')) {
+                            const originalUrl = new URL(requestUrl);
+                            redirectUrl = `${originalUrl.protocol}//${originalUrl.host}${redirectUrl}`;
+                        }
+                        request(redirectUrl);
+                    } else if (response.statusCode === 200) {
+                        const totalSize = parseInt(response.headers['content-length'], 10);
                         let downloadedSize = 0;
                         
-                        redirectResponse.on('data', (chunk) => {
+                        response.on('data', (chunk) => {
                             downloadedSize += chunk.length;
-                            const progress = ((downloadedSize / totalSize) * 100).toFixed(1);
-                            process.stdout.write(`\r  Progress: ${progress}% (${(downloadedSize / 1024 / 1024).toFixed(2)} MB)`);
+                            if (totalSize) {
+                                const progress = ((downloadedSize / totalSize) * 100).toFixed(1);
+                                process.stdout.write(`\r  Downloading model: ${progress}% (${(downloadedSize / 1024 / 1024).toFixed(2)} MB)`);
+                            }
                         });
-                        
-                        redirectResponse.pipe(file);
+
+                        response.pipe(file);
                         
                         file.on('finish', () => {
                             file.close();
                             process.stdout.write('\n');
                             resolve();
                         });
-                    }).on('error', (err) => {
-                        fs.unlink(destPath, () => {});
-                        reject(err);
-                    });
-                } else if (response.statusCode === 200) {
-                    response.pipe(file);
-                    file.on('finish', () => {
-                        file.close();
-                        resolve();
-                    });
-                } else {
-                    reject(new Error(`Failed to download: ${response.statusCode}`));
-                }
-            }).on('error', (err) => {
-                fs.unlink(destPath, () => {});
-                reject(err);
-            });
+                    } else {
+                        reject(new Error(`Failed to download: ${response.statusCode}`));
+                    }
+                }).on('error', (err) => {
+                    fs.unlink(destPath, () => {});
+                    reject(err);
+                });
+            };
+
+            request(url);
             
             file.on('error', (err) => {
                 fs.unlink(destPath, () => {});
@@ -95,44 +92,32 @@ class ModelManager {
     }
 
     /**
-     * Download all model files
+     * Download model file
      */
-    async downloadModels() {
+    async downloadModel() {
         this.ensureCacheDir();
         
-        console.log('Downloading monocr models from HuggingFace...');
-        console.log(`Cache directory: ${this.cacheDir}\n`);
-        
-        // Download model
-        const modelUrl = `${this.baseUrl}/${this.modelFiles.model}`;
-        const modelPath = this.getLocalPath('model');
-        console.log('Downloading monocr.onnx...');
-        await this.downloadFile(modelUrl, modelPath);
-        console.log('✓ Model downloaded\n');
+        console.log('Downloading monocr model from HuggingFace...');
+        console.log(`Cache directory: ${this.cacheDir}`);
         
-        // Download charset
-        const charsetUrl = `${this.baseUrl}/${this.modelFiles.charset}`;
-        const charsetPath = this.getLocalPath('charset');
-        console.log('Downloading charset.txt...');
-        await this.downloadFile(charsetUrl, charsetPath);
-        console.log('✓ Charset downloaded\n');
+        const modelUrl = `${this.baseUrl}/${this.hfModelPath}`;
+        const destPath = this.getModelPath();
         
-        console.log('All models downloaded successfully!');
+        await this.downloadFile(modelUrl, destPath);
+        console.log('✓ Model downloaded successfully!');
     }
 
     /**
-     * Get model paths, downloading if needed
+     * Get model path, downloading if needed
      */
-    async getModelPaths() {
-        if (!this.hasModels()) {
-            await this.downloadModels();
+    async ensureModel() {
+        if (!this.hasModel()) {
+            await this.downloadModel();
         }
-        
-        return {
-            modelPath: this.getLocalPath('model'),
-            charsetPath: this.getLocalPath('charset')
-        };
+        return this.getModelPath();
     }
 }
 
 module.exports = ModelManager;
+
+module.exports = ModelManager;

package/src/monocr.js

@@ -1,6 +1,7 @@
 const ort = require('onnxruntime-node');
 const sharp = require('sharp');
 const fs = require('fs');
+const path = require('path');
 const LineSegmenter = require('./segmenter');
 const ModelManager = require('./model-manager');
 
@@ -21,17 +22,18 @@ class MonOCR {
     async init() {
         if (this.session) return;
         
-        // If paths not provided, use auto-download
-        if (!this.modelPath || !this.charsetPath) {
-            const paths = await this.modelManager.getModelPaths();
-            this.modelPath = this.modelPath || paths.modelPath;
-            this.charsetPath = this.charsetPath || paths.charsetPath;
+        // Ensure model exists
+        if (!this.modelPath) {
+            this.modelPath = await this.modelManager.ensureModel();
         }
         
-        this.session = await ort.InferenceSession.create(this.modelPath);
-        if (this.charsetPath) {
-            this.charset = fs.readFileSync(this.charsetPath, 'utf-8').trim();
+        // Use bundled charset if not provided
+        if (!this.charsetPath) {
+            this.charsetPath = path.join(__dirname, 'charset.txt');
         }
+        
+        this.session = await ort.InferenceSession.create(this.modelPath);
+        this.charset = fs.readFileSync(this.charsetPath, 'utf-8').trim();
     }
 
     /**