Converter with notes

file-management/pdf-with-notes.py

@@ -1,5 +1,6 @@
 import os
-import re
+import shutil
+import fitz  # PyMuPDF
 from pptx import Presentation
 from fpdf import FPDF
 from fpdf.enums import XPos, YPos
@@ -33,30 +34,35 @@ def clean_text_for_fpdf(text):
     
     return text.encode('latin-1', 'ignore').decode('latin-1')
 
-def find_slide_image(images_folder, slide_num):
+def create_handout_pdf(pptx_path, slides_pdf_path, output_pdf):
-    """Smartly searches for an image file matching the slide number."""
+    temp_images_folder = "temp_slide_images"
-    if not os.path.exists(images_folder):
-        return None
-        
-    for filename in os.listdir(images_folder):
-        # Check if it's an image file
-        if filename.lower().endswith(('.png', '.jpg', '.jpeg')):
-            # Extract all numbers from the filename
-            numbers = re.findall(r'\d+', filename)
-            # If the last number in the filename matches our slide number, we found it!
-            if numbers and int(numbers[-1]) == slide_num:
-                return os.path.join(images_folder, filename)
-    return None
-
-def create_handout_pdf(pptx_path, images_folder, output_pdf):
-    print(f"Reading presentation: {pptx_path}...")
     
+    # 1. Check if the exported PDF exists
+    if not os.path.exists(slides_pdf_path):
+        print(f"⚠️ ERROR: Could not find {slides_pdf_path}.")
+        print("Please open OnlyOffice and Save As -> PDF into your Assets folder.")
+        return
+
+    # 2. Slice the PDF into images automatically
+    print(f"Slicing slides from {slides_pdf_path}...")
+    if not os.path.exists(temp_images_folder):
+        os.makedirs(temp_images_folder)
+        
+    pdf_doc = fitz.open(slides_pdf_path)
+    for i in range(len(pdf_doc)):
+        page = pdf_doc[i]
+        pix = page.get_pixmap(dpi=150) # 150 DPI is perfect for standard PDF viewing
+        pix.save(os.path.join(temp_images_folder, f"Slide{i+1}.png"))
+        
+    # 3. Read the PPTX to get the notes
+    print(f"Reading presentation notes from {pptx_path}...")
     try:
         prs = Presentation(pptx_path)
     except Exception as e:
         print(f"Error loading PPTX: {e}")
         return
 
+    # 4. Build the final PDF
     pdf = NotesPDF()
     pdf.set_auto_page_break(auto=True, margin=15)
 
@@ -64,20 +70,19 @@ def create_handout_pdf(pptx_path, images_folder, output_pdf):
         slide_num = i + 1
         pdf.add_page()
         
-        # 1. Smartly find and place the Slide Image
+        # Place the automatically generated image
-        img_path = find_slide_image(images_folder, slide_num)
+        img_path = os.path.join(temp_images_folder, f"Slide{slide_num}.png")
             
-        if img_path:
+        if os.path.exists(img_path):
-            # A4 width is ~210mm. 170mm width leaves 20mm margins on each side.
             pdf.image(img_path, x=20, w=170)
             pdf.ln(5) 
         else:
             pdf.set_font("Helvetica", "B", 12)
             pdf.set_text_color(255, 0, 0)
-            pdf.cell(0, 10, f"[Slide {slide_num} Image Missing - Check 'images' folder]", new_x=XPos.LMARGIN, new_y=YPos.NEXT)
+            pdf.cell(0, 10, f"[Slide {slide_num} Image Missing]", new_x=XPos.LMARGIN, new_y=YPos.NEXT)
             pdf.ln(5)
         
-        # 2. Extract and Place the Notes
+        # Place the Notes
         pdf.set_font("Helvetica", "B", 14)
         pdf.set_text_color(0, 0, 0)
         pdf.cell(0, 10, f"Notes - Slide {slide_num}:", new_x=XPos.LMARGIN, new_y=YPos.NEXT)
@@ -94,19 +99,21 @@ def create_handout_pdf(pptx_path, images_folder, output_pdf):
         notes_text = clean_text_for_fpdf(notes_text)
         pdf.multi_cell(0, 6, notes_text)
 
-    print(f"Compiling PDF...")
+    print(f"Compiling final PDF...")
     pdf.output(output_pdf)
+    
+    # 5. Clean up the temporary images folder
+    if os.path.exists(temp_images_folder):
+        shutil.rmtree(temp_images_folder)
+        
     print(f"✅ Success! Your file is ready: {output_pdf}")
 
 if __name__ == "__main__":
+    # Pointing exactly to your Downloads/Assets folder
     PPTX_FILE = "/Users/guillemhernandezsola/Downloads/Assets/presentation.pptx"
-    # Make sure this folder exists in the same directory where you run the script!
+    SLIDES_PDF = "/Users/guillemhernandezsola/Downloads/Assets/slides.pdf"
-    IMAGES_DIR = "/Users/guillemhernandezsola/Downloads/Assets/images"       
+    
+    # The final file will be saved in your current code directory
     OUTPUT_FILE = "/Users/guillemhernandezsola/Downloads/Assets/Final_Presentation_Notes.pdf"
     
-    if not os.path.exists(IMAGES_DIR):
+    create_handout_pdf(PPTX_FILE, SLIDES_PDF, OUTPUT_FILE)
-        print(f"⚠️ ERROR: The folder '{IMAGES_DIR}' was not found.")
-        print("Please create it in: /Users/guillemhernandezsola/code/scripts/file-management/")
-        print("And make sure you exported your slide images from OnlyOffice into it!")
-    else:
-        create_handout_pdf(PPTX_FILE, IMAGES_DIR, OUTPUT_FILE)