Add textbook analysis

.gitignore

@@ -39,6 +39,7 @@ assets/
 
 # Uploaded catalogs (user-specific)
 catalog/uploaded_*.json
+catalog/textbook_*.json
 
 # Test & Coverage
 .pytest_cache/

api_server.py

@@ -14,7 +14,7 @@
 from queue import Queue
 from threading import Thread
 
-from fastapi import FastAPI, BackgroundTasks, HTTPException, UploadFile, File, Header
+from fastapi import FastAPI, BackgroundTasks, HTTPException, UploadFile, File, Header, Form
 from fastapi.middleware.cors import CORSMiddleware
 from fastapi.responses import FileResponse, JSONResponse, StreamingResponse
 from fastapi.staticfiles import StaticFiles
@@ -26,6 +26,7 @@
 from src import __version__
 from src.pdf_processor import PDFSlideProcessor
 from src.ADDIE_optimize import ADDIEOptimizer
+from src.textbook_reference import TextbookReferenceBuilder
 import tempfile
 import shutil
 
@@ -67,6 +68,7 @@ class OptimizeRequest(BaseModel):
     model_name: str = Field(default="gpt-4o-mini", description="OpenAI model to use")
     exp_name: str = Field(default="default", description="Experiment name for output")
     chapter_name: Optional[str] = Field(default=None, description="Specific chapter to optimize (None = all)")
+    mode: str = Field(default="regenerate", description="Improvement strategy: 'regenerate' (per-slide full rewrite) or 'refine' (localized frame-level rewrite)")
 
 class TaskStatus(BaseModel):
     task_id: str
@@ -391,6 +393,58 @@ async def upload_catalog(
     except Exception as e:
         raise HTTPException(status_code=500, detail=f"Error uploading catalog: {str(e)}")
 
+@app.post("/api/textbook/catalog")
+async def build_textbook_catalog(
+    file: UploadFile = File(...),
+    course_name: str = Form(default="")
+):
+    """
+    Build catalog-compatible reference data from an uploaded textbook file.
+
+    This proof-of-concept extracts a bounded amount of text from PDF/TXT/MD
+    input and converts it into the existing catalog schema. It does not call
+    an LLM, so it can be used before starting a paid generation task.
+    """
+    allowed_suffixes = {".pdf", ".txt", ".md"}
+    original_name = file.filename or "textbook.pdf"
+    suffix = Path(original_name).suffix.lower()
+    if suffix not in allowed_suffixes:
+        raise HTTPException(status_code=400, detail="Only PDF, TXT, and Markdown textbook files are supported")
+
+    MAX_FILE_SIZE = 50 * 1024 * 1024  # 50 MB
+    content = await file.read()
+    if len(content) > MAX_FILE_SIZE:
+        raise HTTPException(status_code=400, detail="Textbook file is too large (max 50 MB)")
+
+    try:
+        with tempfile.TemporaryDirectory() as temp_dir:
+            temp_path = Path(temp_dir) / original_name
+            temp_path.write_bytes(content)
+
+            builder = TextbookReferenceBuilder()
+            catalog_data = builder.build_catalog(temp_path, course_name=course_name)
+
+        catalog_dir = Path("catalog")
+        catalog_dir.mkdir(exist_ok=True)
+        catalog_name = f"textbook_{uuid.uuid4().hex[:8]}"
+        catalog_filename = f"{catalog_name}.json"
+        catalog_path = catalog_dir / catalog_filename
+        with open(catalog_path, "w", encoding="utf-8") as f:
+            json.dump(catalog_data, f, indent=2, ensure_ascii=False)
+
+        return {
+            "catalog_name": catalog_name,
+            "filename": catalog_filename,
+            "saved_path": str(catalog_path),
+            "catalog_data": catalog_data,
+            "message": "Textbook reference catalog generated successfully"
+        }
+
+    except ValueError as e:
+        raise HTTPException(status_code=400, detail=str(e))
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=f"Error processing textbook: {str(e)}")
+
 @app.get("/api/tasks/list")
 async def list_tasks():
     """
@@ -900,6 +954,8 @@ async def run_optimization_task(task_id: str, request: OptimizeRequest, api_key:
         sys.stdout.flush()
         print(f"Experiment: {request.exp_name}")
         sys.stdout.flush()
+        print(f"Mode: {request.mode}")
+        sys.stdout.flush()
         if request.chapter_name:
             print(f"Chapter: {request.chapter_name}")
             sys.stdout.flush()
@@ -916,6 +972,7 @@ async def run_optimization_task(task_id: str, request: OptimizeRequest, api_key:
             model_name=request.model_name,
             exp_name=request.exp_name,
             chapter_name=request.chapter_name,
+            mode=request.mode,
         )
 
         print("\n" + "=" * 60)

frontend/app.js

@@ -96,6 +96,10 @@ const translations = {
         modeGenerate: '生成新课程',
         modeOptimize: '优化已有材料',
         pptxLabel: '同时生成 PPTX 幻灯片',
+        optimizeModeLabel: '优化方式',
+        optimizeModeRegenerate: '逐张重新生成（整章重写）',
+        optimizeModeRefine: '局部精修（仅改相关帧）',
+        optimizeModeTip: '重新生成：每张幻灯片整体重写。局部精修：只定位并改写与需求相关的帧，其余保持不变。',
         optimizeSectionTitle: '优化配置',
         optimizeSubmitButton: '<span>🔧</span><span>开始优化</span>',
         optimizeProgressTitle: '优化进度',
@@ -165,6 +169,8 @@ const translations = {
         modelOptionGpt4Turbo: 'GPT-4 Turbo',
         expNameLabel: 'Experiment Name',
         expNamePlaceholder: 'Default: default',
+        textbookFileLabel: 'Textbook Reference File (optional)',
+        textbookFileTip: 'Supports PDF/TXT/MD. The system extracts a bounded excerpt and saves a textbook catalog JSON for review.',
         copilotLabel: 'Enable Copilot Mode (Interactive Feedback)',
         catalogModeLabel: 'Catalog Mode',
         catalogOptionNone: 'Do not use',
@@ -234,9 +240,15 @@ const translations = {
         catalogListFailed: 'Failed to load catalog list',
         catalogSelectDefault: 'Select Catalog...',
         uploadCatalogFailed: 'Failed to upload catalog file',
+        uploadTextbookFailed: 'Failed to process textbook reference file',
+        textbookCatalogReady: 'Textbook catalog generated: {filename}',
         modeGenerate: 'Generate Course',
         modeOptimize: 'Optimize Materials',
         pptxLabel: 'Also generate PPTX slides',
+        optimizeModeLabel: 'Optimization Method',
+        optimizeModeRegenerate: 'Regenerate (rewrite every slide)',
+        optimizeModeRefine: 'Localized Refine (edit only relevant frames)',
+        optimizeModeTip: 'Regenerate: rewrite each slide wholesale. Localized Refine: locate and rewrite only the frames relevant to your requirements, leaving the rest untouched.',
         optimizeSectionTitle: 'Optimization Settings',
         optimizeSubmitButton: '<span>🔧</span><span>Start Optimization</span>',
         optimizeProgressTitle: 'Optimization Progress',
@@ -503,6 +515,13 @@ function setupEventListeners() {
     const catalogMode = document.getElementById('catalog-mode');
     catalogMode.addEventListener('change', handleCatalogModeChange);
 
+    // Textbook file input: when a file is selected, disable catalog mode dropdown
+    // since the textbook will generate its own catalog automatically.
+    const textbookFile = document.getElementById('textbook-file');
+    if (textbookFile) {
+        textbookFile.addEventListener('change', handleTextbookFileChange);
+    }
+
     // API Key management
     document.getElementById('save-api-key').addEventListener('click', saveApiKey);
     document.getElementById('toggle-api-key').addEventListener('click', toggleApiKeyVisibility);
@@ -648,6 +667,27 @@ function handleCatalogModeChange(e) {
     }
 }
 
+function handleTextbookFileChange() {
+    const textbookInput = document.getElementById('textbook-file');
+    const catalogMode = document.getElementById('catalog-mode');
+    const note = document.getElementById('textbook-catalog-note');
+    const hasFile = textbookInput && textbookInput.files.length > 0;
+
+    if (hasFile) {
+        catalogMode.value = 'none';
+        catalogMode.disabled = true;
+        if (note) note.style.display = 'block';
+        // Hide any open catalog sub-groups
+        ['catalog-upload-group', 'catalog-select-group', 'catalog-json-group'].forEach(id => {
+            const el = document.getElementById(id);
+            if (el) el.style.display = 'none';
+        });
+    } else {
+        catalogMode.disabled = false;
+        if (note) note.style.display = 'none';
+    }
+}
+
 async function handleFormSubmit(e) {
     e.preventDefault();
 
@@ -675,32 +715,48 @@ async function handleFormSubmit(e) {
             generate_pptx: document.getElementById('pptx-mode').checked
         };
 
-        // Handle catalog
-        const catalogMode = document.getElementById('catalog-mode').value;
-        if (catalogMode === 'default') {
-            formData.catalog = 'default_catalog';
-        } else if (catalogMode === 'select') {
-            const selected = document.getElementById('catalog-select').value;
-            if (selected) {
-                formData.catalog = selected;
-            }
-        } else if (catalogMode === 'upload') {
-            // Handle file upload or JSON input
-            const fileInput = document.getElementById('catalog-file');
-            const jsonInput = document.getElementById('catalog-json').value;
-
-            if (fileInput.files.length > 0) {
-                // Upload file first
-                const uploadResponse = await uploadCatalogFile(fileInput.files[0]);
-                formData.catalog = uploadResponse.filename.replace('.json', '');
-            } else if (jsonInput.trim()) {
-                // Use JSON input directly
-                try {
-                    formData.catalog_data = JSON.parse(jsonInput);
-                } catch (err) {
-                    alert(t('invalidCatalogJson'));
-                    setSubmitButtonLoading(false);
-                    return;
+        // Handle optional textbook reference upload first.
+        // When a textbook file is provided, it takes priority — the catalog mode
+        // dropdown is ignored because the textbook generates its own catalog.
+        const textbookInput = document.getElementById('textbook-file');
+        if (textbookInput && textbookInput.files.length > 0) {
+            const textbookResponse = await uploadTextbookCatalog(
+                textbookInput.files[0],
+                formData.course_name
+            );
+
+            formData.catalog = textbookResponse.catalog_name;
+
+            console.log(t('textbookCatalogReady', { filename: textbookResponse.filename }));
+            await loadCatalogs();
+        } else {
+            // Handle catalog mode (only when no textbook is uploaded)
+            const catalogMode = document.getElementById('catalog-mode').value;
+            if (catalogMode === 'default') {
+                formData.catalog = 'default_catalog';
+            } else if (catalogMode === 'select') {
+                const selected = document.getElementById('catalog-select').value;
+                if (selected) {
+                    formData.catalog = selected;
+                }
+            } else if (catalogMode === 'upload') {
+                // Handle file upload or JSON input
+                const fileInput = document.getElementById('catalog-file');
+                const jsonInput = document.getElementById('catalog-json').value;
+
+                if (fileInput.files.length > 0) {
+                    // Upload file first
+                    const uploadResponse = await uploadCatalogFile(fileInput.files[0]);
+                    formData.catalog = uploadResponse.filename.replace('.json', '');
+                } else if (jsonInput.trim()) {
+                    // Use JSON input directly
+                    try {
+                        formData.catalog_data = JSON.parse(jsonInput);
+                    } catch (err) {
+                        alert(t('invalidCatalogJson'));
+                        setSubmitButtonLoading(false);
+                        return;
+                    }
                 }
             }
         }
@@ -794,6 +850,24 @@ async function uploadCatalogFile(file) {
     return await response.json();
 }
 
+async function uploadTextbookCatalog(file, courseName) {
+    const formData = new FormData();
+    formData.append('file', file);
+    formData.append('course_name', courseName || '');
+
+    const response = await fetch(`${API_BASE_URL}/api/textbook/catalog`, {
+        method: 'POST',
+        body: formData
+    });
+
+    if (!response.ok) {
+        const errorText = await response.text();
+        throw new Error(`${t('uploadTextbookFailed')}: ${errorText}`);
+    }
+
+    return await response.json();
+}
+
 function startStatusPolling() {
     if (statusCheckInterval) {
         clearInterval(statusCheckInterval);
@@ -1414,6 +1488,8 @@ async function handleOptimizeSubmit(e) {
     const modelName = document.getElementById('optimize-model-name').value;
     const expName = document.getElementById('optimize-exp-name').value.trim() || 'default';
     const chapterName = document.getElementById('chapter-name').value.trim() || null;
+    const modeSelect = document.getElementById('optimize-mode-select');
+    const mode = modeSelect ? modeSelect.value : 'regenerate';
 
     const submitBtn = document.getElementById('optimize-submit-btn');
     submitBtn.disabled = true;
@@ -1439,7 +1515,8 @@ async function handleOptimizeSubmit(e) {
                 user_requirements: userRequirements,
                 model_name: modelName,
                 exp_name: expName,
-                chapter_name: chapterName
+                chapter_name: chapterName,
+                mode: mode
             })
         });
 

frontend/index.html

@@ -78,6 +78,13 @@ <h2 data-i18n="configSectionTitle">课程配置</h2>
                         <input type="text" id="exp-name" name="exp_name" 
                                placeholder="默认：default" value="default" data-i18n-placeholder="expNamePlaceholder">
                     </div>
+
+                    <div class="form-group">
+                        <label for="textbook-file" data-i18n="textbookFileLabel">上传教材参考文件（可选）</label>
+                        <input type="file" id="textbook-file" accept=".pdf,.txt,.md">
+                        <small data-i18n="textbookFileTip">支持 PDF/TXT/MD；系统会先提取前若干页并保存为 textbook catalog JSON 供检查。</small>
+                    </div>
+
 
                     <div class="form-group checkbox-group">
                         <label>
@@ -101,6 +108,9 @@ <h2 data-i18n="configSectionTitle">课程配置</h2>
                             <option value="upload" data-i18n="catalogOptionUpload">上传 Catalog 文件</option>
                             <option value="select" data-i18n="catalogOptionSelect">选择已有 Catalog</option>
                         </select>
+                        <small id="textbook-catalog-note" style="display:none;color:#856404;background:#fff3cd;padding:4px 8px;border-radius:4px;margin-top:4px;display:none;">
+                            Textbook file uploaded above — the extracted catalog will be used automatically. Catalog mode selection is ignored.
+                        </small>
                     </div>
 
                     <div class="form-group" id="catalog-upload-group" style="display: none;">
@@ -179,6 +189,15 @@ <h2 data-i18n="optimizeSectionTitle">优化配置</h2>
                         </select>
                     </div>
 
+                    <div class="form-group">
+                        <label for="optimize-mode-select" data-i18n="optimizeModeLabel">优化方式</label>
+                        <select id="optimize-mode-select" name="mode">
+                            <option value="regenerate" selected data-i18n="optimizeModeRegenerate">逐张重新生成（整章重写）</option>
+                            <option value="refine" data-i18n="optimizeModeRefine">局部精修（仅改相关帧）</option>
+                        </select>
+                        <small data-i18n="optimizeModeTip">重新生成：每张幻灯片整体重写。局部精修：只定位并改写与需求相关的帧，其余保持不变。</small>
+                    </div>
+
                     <div class="form-group">
                         <label for="optimize-exp-name" data-i18n="expNameLabel">实验名称</label>
                         <input type="text" id="optimize-exp-name" name="exp_name"

run.py

@@ -116,7 +116,8 @@ def run_instructional_design(course_name: str, copilot = None, catalog = None, m
 
 
 def run_optimization(storage_id: str, user_requirements: str, model_name: str = "gpt-4o-mini",
-                     exp_name: str = "optimize", chapter_name: str = None):
+                     exp_name: str = "optimize", chapter_name: str = None,
+                     mode: str = "regenerate"):
     """
     Run the optimization workflow on existing slide materials.
 
@@ -126,6 +127,8 @@ def run_optimization(storage_id: str, user_requirements: str, model_name: str =
         model_name: Name of the LLM model to use
         exp_name: Experiment name for output directory
         chapter_name: Specific chapter to optimize (None = all chapters)
+        mode: "regenerate" (per-slide full rewrite) or "refine" (localized
+            frame-level rewrite via SlideRefiner)
     """
     # Ensure the OPENAI_API_KEY is set
     if not os.environ.get("OPENAI_API_KEY"):
@@ -146,6 +149,7 @@ def run_optimization(storage_id: str, user_requirements: str, model_name: str =
         output_dir=output_dir,
         exp_name=exp_name,
         chapter_name=chapter_name,
+        mode=mode,
     )