Add system prompt, glossary, presets for Ollama/WebLLM, image translation support

translators/word_translator.py

@@ -9,7 +9,11 @@ from docx.table import Table, _Cell
 from docx.oxml.text.paragraph import CT_P
 from docx.oxml.table import CT_Tbl
 from docx.section import Section
+from docx.shared import Inches, Pt
+from docx.oxml.ns import qn
 from services.translation_service import translation_service
+import tempfile
+import os
 
 
 class WordTranslator:
@@ -39,11 +43,66 @@ class WordTranslator:
         for section in document.sections:
             self._translate_section(section, target_language)
         
+        # Translate images if enabled
+        if getattr(self.translation_service, 'translate_images', False):
+            self._translate_images(document, target_language, input_path)
+        
         # Save the translated document
         document.save(output_path)
         
         return output_path
     
+    def _translate_images(self, document: Document, target_language: str, input_path: Path):
+        """
+        Extract text from images and add translations as captions
+        """
+        from services.translation_service import OllamaTranslationProvider
+        
+        # Only works with Ollama vision
+        if not isinstance(self.translation_service.provider, OllamaTranslationProvider):
+            return
+        
+        try:
+            import zipfile
+            import base64
+            
+            # Extract images from docx (it's a zip file)
+            with zipfile.ZipFile(input_path, 'r') as zip_ref:
+                image_files = [f for f in zip_ref.namelist() if f.startswith('word/media/')]
+                
+                for idx, image_file in enumerate(image_files):
+                    try:
+                        # Extract image
+                        image_data = zip_ref.read(image_file)
+                        
+                        # Create temp file
+                        ext = os.path.splitext(image_file)[1]
+                        with tempfile.NamedTemporaryFile(suffix=ext, delete=False) as tmp:
+                            tmp.write(image_data)
+                            tmp_path = tmp.name
+                        
+                        # Translate image with vision
+                        translated_text = self.translation_service.provider.translate_image(tmp_path, target_language)
+                        
+                        # Clean up temp file
+                        os.unlink(tmp_path)
+                        
+                        if translated_text and translated_text.strip():
+                            # Add translated text as a new paragraph after image
+                            # We'll add it at the end with a note
+                            p = document.add_paragraph()
+                            p.add_run(f"[Image {idx + 1} translation: ").bold = True
+                            p.add_run(translated_text)
+                            p.add_run("]").bold = True
+                            
+                            print(f"Translated image {idx + 1}: {translated_text[:50]}...")
+                    except Exception as e:
+                        print(f"Error translating image {image_file}: {e}")
+                        continue
+                        
+        except Exception as e:
+            print(f"Error processing images: {e}")
+    
     def _translate_document_body(self, document: Document, target_language: str):
         """
         Translate all elements in the document body