new functionalities added for resume analyzer

shliamin · Sep 2, 2024 · 12e4e31 · 12e4e31
1 parent 0122d03
commit 12e4e31
Show file tree

Hide file tree

Showing 3 changed files with 68 additions and 20 deletions.
diff --git a/main.js b/main.js
@@ -141,6 +141,34 @@ import { analyzeCategories, detectLanguage } from './resumeAnalyzer.js';
 
 let chartInstance = null; 
 
+
+function extractTextFromPDF(pdfData) {
+    const pdfjsLib = window['pdfjsLib'];
+    if (!pdfjsLib) {
+        console.error('pdfjsLib is not loaded');
+        return;
+    }
+
+    pdfjsLib.GlobalWorkerOptions.workerSrc = 'https://cdnjs.cloudflare.com/ajax/libs/pdf.js/2.10.377/pdf.worker.min.js';
+
+    const loadingTask = pdfjsLib.getDocument({data: pdfData});
+    return loadingTask.promise.then(pdf => {
+        const maxPages = pdf.numPages;
+        const pagePromises = [];
+        for (let i = 1; i <= maxPages; i++) {
+            pagePromises.push(pdf.getPage(i).then(page => {
+                return page.getTextContent().then(textContent => {
+                    return textContent.items.map(item => item.str).join(' ');
+                });
+            }));
+        }
+        return Promise.all(pagePromises).then(pagesText => pagesText.join(' '));
+    });
+}
+
+
+
+
 // Resume upload handler
 document.getElementById('resume').addEventListener('change', function(event) {
     const resumeFile = event.target.files[0];
@@ -154,35 +182,42 @@ function analyzeResume(resume) {
 
     reader.onload = function(e) {
         const content = e.target.result;
-
-        fetch('english_keywords.txt')
-            .then(response => response.text())
-            .then(englishKeywords => {
-                fetch('german_keywords.txt')
-                    .then(response => response.text())
-                    .then(germanKeywords => {
-                        const language = detectLanguage(content, englishKeywords, germanKeywords);
-                        if (language !== 'en' && language !== 'de') {
-                            alert('The document must be in English or German.');
-                            return;
-                        }
-
-                        const categories = analyzeCategories(content, language === 'en' ? englishKeywords : germanKeywords);
-                        displayChart(categories);
-                    });
+        if (resume.type === 'application/pdf') {
+            extractTextFromPDF(content).then(text => {
+                analyzeTextContent(text);
             });
+        } else {
+            analyzeTextContent(content);
+        }
     };
 
     if (resume.type === 'application/pdf') {
-        reader.readAsBinaryString(resume);
+        reader.readAsArrayBuffer(resume);
     } else if (resume.type === 'text/plain') {
         reader.readAsText(resume);
     } else {
         alert('Unsupported file type. Please upload a PDF or TXT file.');
     }
 }
 
+function analyzeTextContent(content) {
+    fetch('english_keywords.txt')
+        .then(response => response.text())
+        .then(englishKeywords => {
+            fetch('german_keywords.txt')
+                .then(response => response.text())
+                .then(germanKeywords => {
+                    const language = detectLanguage(content, englishKeywords, germanKeywords);
+                    if (language !== 'en' && language !== 'de') {
+                        alert('The document must be in English or German.');
+                        return;
+                    }
 
+                    const categories = analyzeCategories(content, language === 'en' ? englishKeywords : germanKeywords);
+                    displayChart(categories);
+                });
+        });
+}
 
 
 

diff --git a/resumeAnalyzer.js b/resumeAnalyzer.js
@@ -29,6 +29,10 @@ export function detectLanguage(text, englishKeywords, germanKeywords) {
 }
 
 export function analyzeCategories(content, keywords) {
+    // Разбиваем строку на массив ключевых слов (это используется для анализа языка)
+    const keywordArray = keywords.split('\n').map(keyword => keyword.trim());
+    console.log('Содержимое текста:', content);
+
     const subcategoryKeywords = {
         'Experience': [
             'experience', 'work experience', 'professional experience', 'career history', 
@@ -133,9 +137,15 @@ export function analyzeCategories(content, keywords) {
             for (const subcategory in categoryInfo.subcategories) {
                 const originalValue = categoryInfo.subcategories[subcategory];
 
+                // Используем правильный массив ключевых слов для подкатегории
+                const keywordsForSubcategory = subcategoryKeywords[subcategory];
+                const found = keywordsForSubcategory.some(keyword => content.toLowerCase().includes(keyword.toLowerCase()));
 
-                const keywords = subcategoryKeywords[subcategory];
-                const found = keywords ? keywords.some(keyword => content.toLowerCase().includes(keyword.toLowerCase())) : false;
+                if (found) {
+                    console.log(`Подкатегория '${subcategory}' найдена в тексте.`);
+                } else {
+                    console.log(`Подкатегория '${subcategory}' НЕ найдена в тексте.`);
+                }
 
                 finalCategories[subcategory] = {
                     value: found ? originalValue : originalValue,
@@ -147,8 +157,11 @@ export function analyzeCategories(content, keywords) {
                 value: categoryInfo.value,
                 color: categoryInfo.color
             };
+
+            console.log(`Категория '${category}' не имеет подкатегорий.`);
         }
     }
 
+    console.log('Финальные категории:', finalCategories);
     return finalCategories;
 }
diff --git a/resume_analyzer.html b/resume_analyzer.html
@@ -285,7 +285,7 @@ <h1>Compatibility Analyzer</h1>
         </div>
     </div>
 
-
+    <script src="https://cdnjs.cloudflare.com/ajax/libs/pdf.js/2.10.377/pdf.min.js"></script>
     <script src="https://cdn.jsdelivr.net/npm/chart.js"></script>
     <script type="module" src="main.js"></script>