Merge 92cc53bf08 into a189e25a82

backend/sandbox/docker/requirements.txt

@@ -7,4 +7,5 @@ pytesseract==0.3.13
 pandas==2.3.0
 playwright>=1.40.0
 PyPDF2>=3.0.0
-bs4==0.0.2
+bs4==0.0.2
+python-pptx>=0.6.23

backend/sandbox/docker/server.py

@@ -5,9 +5,10 @@ import uvicorn
 import os
 from pathlib import Path
 
-# Import PDF router and Visual HTML Editor router
+# Import PDF router, PPTX router, and Visual HTML Editor router
 from html_to_pdf_router import router as pdf_router
 from visual_html_editor_router import router as editor_router
+from html_to_pptx_router import router as pptx_router
 
 # Ensure we're serving from the /workspace directory
 workspace_dir = "/workspace"
@@ -26,6 +27,7 @@ app.add_middleware(WorkspaceDirMiddleware)
 # Include routers
 app.include_router(pdf_router)
 app.include_router(editor_router)
+app.include_router(pptx_router)
 
 # Create output directory for generated PDFs (needed by PDF router)
 output_dir = Path("generated_pdfs")

backend/sandbox/docker/visual_html_editor_router.py

@@ -14,7 +14,7 @@ from fastapi import APIRouter, HTTPException
 from fastapi.responses import HTMLResponse
 from fastapi.staticfiles import StaticFiles
 from pydantic import BaseModel
-from bs4 import BeautifulSoup, NavigableString
+from bs4 import BeautifulSoup, NavigableString, Comment
 
 # Create router
 router = APIRouter(prefix="/api/html", tags=["visual-editor"])
@@ -78,6 +78,24 @@ async def get_editable_elements(file_path: str):
         # Find all elements that could contain text
         all_elements = soup.find_all(TEXT_ELEMENTS + ['div'])
         
+        # Filter out elements that only contain comments
+        filtered_elements = []
+        for element in all_elements:
+            # Check if element only contains comments
+            only_comments = True
+            for child in element.children:
+                if isinstance(child, Comment):
+                    continue
+                if isinstance(child, NavigableString) and not child.strip():
+                    continue
+                only_comments = False
+                break
+                
+            if not only_comments:
+                filtered_elements.append(element)
+                
+        all_elements = filtered_elements
+        
         for element in all_elements:
             # Strategy 1: Elements with ONLY text content (no child elements)
             if element.string and element.string.strip():
@@ -99,6 +117,9 @@ async def get_editable_elements(file_path: str):
                 has_mixed_content = False
                 # Process each child node
                 for child in list(element.contents):  # Use list() to avoid modification during iteration
+                    # Skip comment nodes (Comments are a subclass of NavigableString)
+                    if isinstance(child, Comment):
+                        continue
                     # Check if it's a NavigableString (raw text) with actual content
                     if (isinstance(child, NavigableString) and child.strip()):
                         
@@ -335,6 +356,24 @@ def inject_editor_functionality(html_content: str, file_path: str) -> str:
     # Find all elements that could contain text
     all_elements = soup.find_all(TEXT_ELEMENTS + ['div'])
     
+    # Filter out elements that only contain comments
+    filtered_elements = []
+    for element in all_elements:
+        # Check if element only contains comments
+        only_comments = True
+        for child in element.children:
+            if isinstance(child, Comment):
+                continue
+            if isinstance(child, NavigableString) and not child.strip():
+                continue
+            only_comments = False
+            break
+            
+        if not only_comments:
+            filtered_elements.append(element)
+            
+    all_elements = filtered_elements
+    
     for element in all_elements:
         # Strategy 1: Elements with ONLY text content (no child elements)
         if element.string and element.string.strip():
@@ -347,6 +386,9 @@ def inject_editor_functionality(html_content: str, file_path: str) -> str:
             has_mixed_content = False
             # Process each child node
             for child in list(element.contents):  # Use list() to avoid modification during iteration
+                # Skip comment nodes (Comments are a subclass of NavigableString)
+                if isinstance(child, Comment):
+                    continue
                 # Check if it's a NavigableString (raw text) with actual content
                 if (isinstance(child, NavigableString) and child.strip()):
                     

frontend/src/components/thread/tool-views/presentation-tools/FullScreenPresentationViewer.tsx

@@ -46,6 +46,7 @@ interface FullScreenPresentationViewerProps {
   sandboxUrl?: string;
   initialSlide?: number;
   onPDFDownload?: (setIsDownloadingPDF: (isDownloading: boolean) => void) => void;
+  onPPTXDownload?: (setIsDownloadingPPTX: (isDownloading: boolean) => void) => void;
 }
 
 export function FullScreenPresentationViewer({
@@ -55,6 +56,7 @@ export function FullScreenPresentationViewer({
   sandboxUrl,
   initialSlide = 1,
   onPDFDownload,
+  onPPTXDownload,
 }: FullScreenPresentationViewerProps) {
   const [metadata, setMetadata] = useState<PresentationMetadata | null>(null);
   const [currentSlide, setCurrentSlide] = useState(initialSlide);
@@ -65,6 +67,7 @@ export function FullScreenPresentationViewer({
   const [showControls, setShowControls] = useState(true);
   const [showEditor, setShowEditor] = useState(false);
   const [isDownloadingPDF, setIsDownloadingPDF] = useState(false);
+  const [isDownloadingPPTX, setIsDownloadingPPTX] = useState(false);
   
   // Create a stable refresh timestamp when metadata changes (like PresentationViewer)
   const refreshTimestamp = useMemo(() => Date.now(), [metadata]);
@@ -418,7 +421,7 @@ export function FullScreenPresentationViewer({
                   className="h-8 w-8 p-0"
                   title="Export presentation"
                 >
-                  {isDownloadingPDF ? (
+                  {(isDownloadingPDF || isDownloadingPPTX) ? (
                     <Loader2 className="h-3.5 w-3.5 animate-spin" />
                   ) : (
                     <Download className="h-3.5 w-3.5" />
@@ -430,7 +433,7 @@ export function FullScreenPresentationViewer({
                   <FileText className="h-4 w-4 mr-2" />
                   PDF
                 </DropdownMenuItem>
-                <DropdownMenuItem className="cursor-pointer">
+                <DropdownMenuItem className="cursor-pointer" onClick={() => onPPTXDownload(setIsDownloadingPPTX)}>
                   <Presentation className="h-4 w-4 mr-2" />
                   PPTX
                 </DropdownMenuItem>

frontend/src/components/thread/tool-views/presentation-tools/PresentationViewer.tsx

@@ -79,6 +79,7 @@ export function PresentationViewer({
   const [isFullScreenOpen, setIsFullScreenOpen] = useState(false);
   const [fullScreenInitialSlide, setFullScreenInitialSlide] = useState<number | null>(null);
   const [isDownloadingPDF, setIsDownloadingPDF] = useState(false);
+  const [isDownloadingPPTX, setIsDownloadingPPTX] = useState(false);
 
   // Extract presentation info from tool data
   const { toolResult } = extractToolData(toolContent);
@@ -505,6 +506,36 @@ export function PresentationViewer({
     }
   };
 
+  const downloadPresentationAsPPTX = async (sandboxUrl: string, presentationName: string): Promise<void> => {
+    try {
+      const response = await fetch(`${sandboxUrl}/presentation/convert-to-pptx`, {
+        method: 'POST',
+        headers: {
+          'Content-Type': 'application/json',
+        },
+        body: JSON.stringify({
+          presentation_path: `/workspace/presentations/${presentationName}`,
+          download: true
+        })
+      });
+      
+      if (!response.ok) {
+        throw new Error('Failed to download PPTX');
+      }
+      
+      const blob = await response.blob();
+      const url = window.URL.createObjectURL(blob);
+      const a = document.createElement('a');
+      a.href = url;
+      a.download = `${presentationName}.pptx`;
+      a.click();
+      window.URL.revokeObjectURL(url);
+    } catch (error) {
+      console.error('Error downloading PPTX:', error);
+      toast.error(`Error: ${error instanceof Error ? error.message : 'Unknown error'}`);
+    }
+  };
+
 
   const handlePDFDownload = useCallback(async (setIsDownloadingPDF: (isDownloading: boolean) => void) => {
     
@@ -520,6 +551,20 @@ export function PresentationViewer({
     }
   }, [project?.sandbox?.sandbox_url, extractedPresentationName]);
 
+  const handlePPTXDownload = useCallback(async (setIsDownloadingPPTX: (isDownloading: boolean) => void) => {
+    
+    if (!project?.sandbox?.sandbox_url || !extractedPresentationName) return;
+
+    setIsDownloadingPPTX(true);
+    try{
+      await downloadPresentationAsPPTX(project.sandbox.sandbox_url, extractedPresentationName);
+    } catch (error) {
+      console.error('Error downloading PPTX:', error);
+    } finally {
+      setIsDownloadingPPTX(false);
+    }
+  }, [project?.sandbox?.sandbox_url, extractedPresentationName]);
+
   return (
     <Card className="gap-0 flex border shadow-none border-t border-b-0 border-x-0 p-0 rounded-none flex-col h-full overflow-hidden bg-card">
       <CardHeader className="h-14 bg-zinc-50/80 dark:bg-zinc-900/80 backdrop-blur-sm border-b p-2 px-4 space-y-2">
@@ -560,7 +605,7 @@ export function PresentationViewer({
                       className="h-8 w-8 p-0"
                       title="Export presentation"
                     >
-                      {isDownloadingPDF ? (
+                      {(isDownloadingPDF || isDownloadingPPTX) ? (
                         <Loader2 className="h-3.5 w-3.5 animate-spin" />
                       ) : (
                         <Download className="h-3.5 w-3.5" />
@@ -571,16 +616,15 @@ export function PresentationViewer({
                     <DropdownMenuItem 
                       onClick={() => handlePDFDownload(setIsDownloadingPDF)}
                       className="cursor-pointer"
+                      disabled={isDownloadingPPTX}
                     >
                       <FileText className="h-4 w-4 mr-2" />
                       PDF
                     </DropdownMenuItem>
                     <DropdownMenuItem 
-                      onClick={() => {
+                      onClick={() => handlePPTXDownload(setIsDownloadingPPTX)}
-                        // TODO: Implement PPTX export
-                        console.log('Export as PPTX');
-                      }}
                       className="cursor-pointer"
+                      disabled={isDownloadingPDF}
                     >
                       <Presentation className="h-4 w-4 mr-2" />
                       PPTX
@@ -591,6 +635,7 @@ export function PresentationViewer({
                         console.log('Export to Google Slides');
                       }}
                       className="cursor-pointer"
+                      disabled={isDownloadingPDF || isDownloadingPPTX}
                     >
                       <ExternalLink className="h-4 w-4 mr-2" />
                       Google Slides
@@ -781,6 +826,7 @@ export function PresentationViewer({
         sandboxUrl={project?.sandbox?.sandbox_url}
         initialSlide={fullScreenInitialSlide || visibleSlide || currentSlideNumber || slides[0]?.number || 1}
         onPDFDownload={handlePDFDownload}
+        onPPTXDownload={handlePPTXDownload}
       />
     </Card>
   );