cd /home/daytona/workspace/attachments/668e09e8-00b3-4772-845f-e77794ba2bab && python3 -c " import pdfplumber from pypdf import PdfReader reader = PdfReader('urology.pdf') print(f'Total pages: {len(reader.pages)}') meta = reader.metadata print(f'Title: {meta.title}') print(f'Author: {meta.author}') # Check first page text with pdfplumber.open('urology.pdf') as pdf: print(f'\n--- PAGE 1 TEXT ---') print(pdf.pages[0].extract_text()) print(f'\n--- PAGE 2 TEXT ---') if len(pdf.pages) > 1: print(pdf.pages[1].extract_text()) "

python3 -c " import fitz # PyMuPDF doc = fitz.open('/home/daytona/workspace/attachments/668e09e8-00b3-4772-845f-e77794ba2bab/urology.pdf') print(f'Total pages: {len(doc)}') # Check images on page 1 page = doc[0] print('Page 1 text:', page.get_text()) img_list = page.get_images(full=True) print('Images on page 1:', len(img_list)) # Check page dimensions print('Page 1 size:', page.rect) doc.close() "

# Install pymupdf locally and check what OCR tools are available pip install --user pymupdf --quiet 2>&1 | tail -3 which tesseract 2>/dev/null && tesseract --version 2>&1 | head -2 which pdftoppm 2>/dev/null python3 -c "import pytesseract; print('pytesseract ok')" 2>&1 python3 -c "from PIL import Image; print('PIL ok')" 2>&1

# Convert first few pages to images to inspect content and detect language mkdir -p /home/daytona/workspace/urology-english/pages pdftoppm -r 200 -png /home/daytona/workspace/attachments/668e09e8-00b3-4772-845f-e77794ba2bab/urology.pdf /home/daytona/workspace/urology-english/pages/page ls /home/daytona/workspace/urology-english/pages/ | head -10

# Convert just first 3 pages at lower resolution to check content mkdir -p /home/daytona/workspace/urology-english/pages pdftoppm -r 150 -png -f 1 -l 3 /home/daytona/workspace/attachments/668e09e8-00b3-4772-845f-e77794ba2bab/urology.pdf /home/daytona/workspace/urology-english/pages/page ls /home/daytona/workspace/urology-english/pages/

# Run OCR on first page to detect the language python3 -c " import pytesseract from PIL import Image img = Image.open('/home/daytona/workspace/urology-english/pages/page-01.png') # Try Arabic first text = pytesseract.image_to_string(img, lang='ara') print('=== ARABIC OCR RESULT (page 1) ===') print(text[:1000]) print() # Also try with OSD to detect orientation/script osd = pytesseract.image_to_osd(img) print('=== OSD ===') print(osd) "

# Check what languages are available for tesseract tesseract --list-langs 2>&1

# English OCR on page 1 - check if it's an English PDF or another language python3 -c " import pytesseract from PIL import Image img = Image.open('/home/daytona/workspace/urology-english/pages/page-01.png') text = pytesseract.image_to_string(img, lang='eng') print(repr(text[:2000])) "

# Convert remaining pages (we already have 1-16, need 17-33) pdftoppm -r 150 -png -f 17 -l 33 /home/daytona/workspace/attachments/668e09e8-00b3-4772-845f-e77794ba2bab/urology.pdf /home/daytona/workspace/urology-english/pages/page ls /home/daytona/workspace/urology-english/pages/ | wc -l

python3 << 'EOF' import pytesseract from PIL import Image from reportlab.lib.pagesizes import A4 from reportlab.lib.styles import getSampleStyleSheet, ParagraphStyle from reportlab.lib.units import inch, cm from reportlab.lib import colors from reportlab.platypus import SimpleDocTemplate, Paragraph, Spacer, Image as RLImage, PageBreak from reportlab.lib.enums import TA_LEFT, TA_CENTER import os, io pages_dir = '/home/daytona/workspace/urology-english/pages' output_pdf = '/home/daytona/workspace/urology-english/urology_english.pdf' # OCR all pages page_files = sorted([f for f in os.listdir(pages_dir) if f.endswith('.png')]) print(f"Processing {len(page_files)} pages...") all_texts = [] for pf in page_files: img_path = os.path.join(pages_dir, pf) img = Image.open(img_path) text = pytesseract.image_to_string(img, lang='eng', config='--psm 6') all_texts.append((pf, text, img_path)) print(f" OCR done: {pf} ({len(text)} chars)") print("\nAll OCR complete. Building PDF...") # Build PDF doc = SimpleDocTemplate( output_pdf, pagesize=A4, rightMargin=2*cm, leftMargin=2*cm, topMargin=2*cm, bottomMargin=2*cm ) styles = getSampleStyleSheet() title_style = ParagraphStyle( 'CustomTitle', parent=styles['Title'], fontSize=16, spaceAfter=20, textColor=colors.HexColor('#1a237e'), alignment=TA_CENTER, ) normal_style = ParagraphStyle( 'CustomNormal', parent=styles['Normal'], fontSize=10, leading=14, spaceAfter=6, ) page_header_style = ParagraphStyle( 'PageHeader', parent=styles['Heading2'], fontSize=9, textColor=colors.HexColor('#555555'), spaceAfter=4, ) story = [] # Title page story.append(Spacer(1, 2*inch)) story.append(Paragraph("UROLOGY", title_style)) story.append(Paragraph("Question Bank", title_style)) story.append(Spacer(1, 0.5*inch)) story.append(Paragraph("English Language Edition", ParagraphStyle( 'Sub', parent=styles['Normal'], fontSize=12, alignment=TA_CENTER, textColor=colors.HexColor('#444444') ))) story.append(PageBreak()) # Add each page as image + OCR text W, H = A4 available_width = W - 4*cm available_height = H * 0.45 # max half page for image for i, (pf, text, img_path) in enumerate(all_texts): page_num = i + 1 # Add page header story.append(Paragraph(f"Page {page_num}", page_header_style)) # Add the page image (original scan) img = Image.open(img_path) iw, ih = img.size aspect = ih / iw display_w = available_width display_h = display_w * aspect if display_h > available_height: display_h = available_height display_w = display_h / aspect story.append(RLImage(img_path, width=display_w, height=display_h)) story.append(Spacer(1, 0.15*inch)) # Add OCR text story.append(Paragraph("<b>Extracted Text:</b>", ParagraphStyle( 'Bold', parent=styles['Normal'], fontSize=9, textColor=colors.HexColor('#333333') ))) # Clean and add text lines lines = text.strip().split('\n') text_block = [] for line in lines: line = line.strip() if line: # Escape XML special chars for reportlab line = line.replace('&', '&amp;').replace('<', '&lt;').replace('>', '&gt;') text_block.append(line) if text_block: combined = '<br/>'.join(text_block) story.append(Paragraph(combined, normal_style)) story.append(PageBreak()) doc.build(story) print(f"\nPDF created: {output_pdf}") import os size = os.path.getsize(output_pdf) print(f"File size: {size/1024/1024:.1f} MB") EOF

# Check which file is truncated and re-render that page ls -la /home/daytona/workspace/urology-english/pages/ | sort -k5 -n | head -5 # Re-render pages 16-33 which may have been cut off pdftoppm -r 150 -png -f 16 -l 33 /home/daytona/workspace/attachments/668e09e8-00b3-4772-845f-e77794ba2bab/urology.pdf /home/daytona/workspace/urology-english/pages/page ls -la /home/daytona/workspace/urology-english/pages/ | sort -k5 -n | head -5

python3 << 'EOF' import pytesseract from PIL import Image, ImageFile ImageFile.LOAD_TRUNCATED_IMAGES = True from reportlab.lib.pagesizes import A4 from reportlab.lib.styles import getSampleStyleSheet, ParagraphStyle from reportlab.lib.units import inch, cm from reportlab.lib import colors from reportlab.platypus import SimpleDocTemplate, Paragraph, Spacer, Image as RLImage, PageBreak from reportlab.lib.enums import TA_LEFT, TA_CENTER import os pages_dir = '/home/daytona/workspace/urology-english/pages' output_pdf = '/home/daytona/workspace/urology-english/urology_english.pdf' page_files = sorted([f for f in os.listdir(pages_dir) if f.endswith('.png')]) print(f"Processing {len(page_files)} pages...") all_texts = [] for pf in page_files: img_path = os.path.join(pages_dir, pf) try: img = Image.open(img_path) img.load() text = pytesseract.image_to_string(img, lang='eng', config='--psm 6') except Exception as e: print(f" WARNING: {pf} error: {e}, using blank") text = "" all_texts.append((pf, text, img_path)) print(f" OCR: {pf} ({len(text)} chars)") print("\nBuilding PDF...") doc = SimpleDocTemplate( output_pdf, pagesize=A4, rightMargin=2*cm, leftMargin=2*cm, topMargin=2*cm, bottomMargin=2*cm ) styles = getSampleStyleSheet() title_style = ParagraphStyle('T', parent=styles['Title'], fontSize=18, spaceAfter=12, textColor=colors.HexColor('#1a237e'), alignment=TA_CENTER) subtitle_style = ParagraphStyle('S', parent=styles['Normal'], fontSize=13, spaceAfter=8, alignment=TA_CENTER, textColor=colors.HexColor('#333333')) hdr_style = ParagraphStyle('H', parent=styles['Normal'], fontSize=8, textColor=colors.HexColor('#777777'), spaceAfter=3) text_style = ParagraphStyle('N', parent=styles['Normal'], fontSize=10.5, leading=15, spaceAfter=4) label_style = ParagraphStyle('L', parent=styles['Normal'], fontSize=9, textColor=colors.HexColor('#555555'), spaceBefore=6, spaceAfter=2, fontName='Helvetica-Bold') story = [] # Title page story.append(Spacer(1, 1.8*inch)) story.append(Paragraph("UROLOGY", title_style)) story.append(Paragraph("Question Bank — English Edition", subtitle_style)) story.append(Spacer(1, 0.3*inch)) story.append(Paragraph(f"Total Pages: {len(page_files)}", ParagraphStyle( 'Info', parent=styles['Normal'], fontSize=11, alignment=TA_CENTER, textColor=colors.HexColor('#666666')))) story.append(PageBreak()) W, H = A4 avail_w = W - 4*cm max_img_h = H * 0.48 for i, (pf, text, img_path) in enumerate(all_texts): page_num = i + 1 story.append(Paragraph(f"— Page {page_num} of {len(all_texts)} —", hdr_style)) # Image try: img = Image.open(img_path) iw, ih = img.size aspect = ih / iw dw = avail_w dh = dw * aspect if dh > max_img_h: dh = max_img_h dw = dh / aspect story.append(RLImage(img_path, width=dw, height=dh)) except Exception as e: story.append(Paragraph(f"[Image error: {e}]", text_style)) story.append(Spacer(1, 0.12*inch)) # OCR text lines = [l.strip() for l in text.strip().split('\n') if l.strip()] if lines: story.append(Paragraph("Extracted Text:", label_style)) combined = '<br/>'.join( l.replace('&','&amp;').replace('<','&lt;').replace('>','&gt;') for l in lines ) story.append(Paragraph(combined, text_style)) story.append(PageBreak()) doc.build(story) sz = os.path.getsize(output_pdf) print(f"\nDone! PDF: {output_pdf}") print(f"Size: {sz/1024/1024:.1f} MB") EOF

# Check if PDF was partially created ls -la /home/daytona/workspace/urology-english/urology_english.pdf 2>/dev/null && echo "exists" # Also try a faster approach: just embed original pages as images directly without OCR text python3 << 'EOF' from PIL import Image, ImageFile ImageFile.LOAD_TRUNCATED_IMAGES = True from reportlab.lib.pagesizes import A4 from reportlab.lib.styles import getSampleStyleSheet, ParagraphStyle from reportlab.lib.units import cm from reportlab.lib import colors from reportlab.platypus import SimpleDocTemplate, Paragraph, Spacer, Image as RLImage, PageBreak from reportlab.lib.enums import TA_CENTER import os, sys pages_dir = '/home/daytona/workspace/urology-english/pages' output_pdf = '/home/daytona/workspace/urology-english/urology_english.pdf' page_files = sorted([f for f in os.listdir(pages_dir) if f.endswith('.png')]) print(f"Total pages: {len(page_files)}", flush=True) # Pre-load all OCR text files (if they exist) or use cached ocr_cache = '/home/daytona/workspace/urology-english/ocr_cache.txt' all_texts = {} if os.path.exists(ocr_cache): with open(ocr_cache) as f: content = f.read() # Parse cached for block in content.split('===PAGE==='): parts = block.strip().split('\n', 1) if len(parts) == 2: all_texts[parts[0].strip()] = parts[1].strip() print(f"Loaded {len(all_texts)} cached OCR results") else: print("No cache found, running OCR now...", flush=True) import pytesseract cache_lines = [] for pf in page_files: img_path = os.path.join(pages_dir, pf) try: img = Image.open(img_path) img.load() text = pytesseract.image_to_string(img, lang='eng', config='--psm 6') except Exception as e: text = f"[OCR error: {e}]" all_texts[pf] = text cache_lines.append(f"===PAGE===\n{pf}\n{text}") print(f" {pf}: {len(text)} chars", flush=True) with open(ocr_cache, 'w') as f: f.write('\n'.join(cache_lines)) print("OCR cache saved.", flush=True) print("Building PDF...", flush=True) styles = getSampleStyleSheet() W, H = A4 doc = SimpleDocTemplate(output_pdf, pagesize=A4, rightMargin=1.5*cm, leftMargin=1.5*cm, topMargin=1.5*cm, bottomMargin=1.5*cm) story = [] avail_w = W - 3*cm max_h = H - 3*cm for i, pf in enumerate(page_files): img_path = os.path.join(pages_dir, pf) try: img = Image.open(img_path) iw, ih = img.size dw = avail_w dh = dw * (ih/iw) if dh > max_h: dh = max_h dw = dh * (iw/ih) story.append(RLImage(img_path, width=dw, height=dh)) except Exception as e: story.append(Paragraph(f"[Image error p{i+1}: {e}]", styles['Normal'])) story.append(PageBreak()) if i % 5 == 0: print(f" added page {i+1}", flush=True) print("doc.build()...", flush=True) doc.build(story) sz = os.path.getsize(output_pdf) print(f"Done: {output_pdf} ({sz/1024/1024:.1f} MB)") EOF

PDF="/home/daytona/workspace/attachments/668e09e8-00b3-4772-845f-e77794ba2bab/urology (1).pdf" mkdir -p /home/daytona/workspace/urology-out/pages # Check page count first python3 -c " from pypdf import PdfReader r = PdfReader('/home/daytona/workspace/attachments/668e09e8-00b3-4772-845f-e77794ba2bab/urology (1).pdf') print('Pages:', len(r.pages)) # Try text extraction for i,p in enumerate(r.pages[:3]): t = p.extract_text() print(f'Page {i+1} text len: {len(t)}') if t: print(t[:300]) "

# Convert all 33 pages to images at 180 DPI pdftoppm -r 180 -png \ "/home/daytona/workspace/attachments/668e09e8-00b3-4772-845f-e77794ba2bab/urology (1).pdf" \ /home/daytona/workspace/urology-out/pages/p ls /home/daytona/workspace/urology-out/pages/ | wc -l

ls /home/daytona/workspace/urology-out/pages/ | wc -l ls /home/daytona/workspace/urology-out/pages/ | tail -5

# Convert remaining pages 24-33 pdftoppm -r 180 -png -f 24 -l 33 \ "/home/daytona/workspace/attachments/668e09e8-00b3-4772-845f-e77794ba2bab/urology (1).pdf" \ /home/daytona/workspace/urology-out/pages/p ls /home/daytona/workspace/urology-out/pages/ | wc -l

python3 << 'PYEOF' import pytesseract, os from PIL import Image, ImageFile ImageFile.LOAD_TRUNCATED_IMAGES = True pages_dir = '/home/daytona/workspace/urology-out/pages' cache_file = '/home/daytona/workspace/urology-out/ocr_all.txt' page_files = sorted([f for f in os.listdir(pages_dir) if f.endswith('.png')]) print(f"OCR-ing {len(page_files)} pages...", flush=True) results = [] for pf in page_files: path = os.path.join(pages_dir, pf) try: img = Image.open(path) img.load() text = pytesseract.image_to_string(img, lang='eng', config='--psm 6') text = text.strip() except Exception as e: text = f"[OCR error: {e}]" results.append((pf, text)) print(f" {pf}: {len(text)} chars", flush=True) # Save cache with open(cache_file, 'w', encoding='utf-8') as f: for pf, text in results: f.write(f"===PAGE:{pf}===\n{text}\n\n") print(f"\nSaved to {cache_file}", flush=True) PYEOF

python3 << 'PYEOF' import os from PIL import Image, ImageFile ImageFile.LOAD_TRUNCATED_IMAGES = True from reportlab.lib.pagesizes import A4 from reportlab.lib.styles import getSampleStyleSheet, ParagraphStyle from reportlab.lib.units import cm from reportlab.lib import colors from reportlab.platypus import (SimpleDocTemplate, Paragraph, Spacer, Image as RLImage, PageBreak, HRFlowable) from reportlab.lib.enums import TA_CENTER, TA_LEFT pages_dir = '/home/daytona/workspace/urology-out/pages' cache_file = '/home/daytona/workspace/urology-out/ocr_all.txt' output_pdf = '/home/daytona/workspace/urology-out/Urology_Questions_English.pdf' # Parse OCR cache ocr = {} current_page = None buf = [] with open(cache_file, encoding='utf-8') as f: for line in f: line = line.rstrip('\n') if line.startswith('===PAGE:') and line.endswith('==='): if current_page is not None: ocr[current_page] = '\n'.join(buf).strip() current_page = line[8:-3] buf = [] else: buf.append(line) if current_page is not None: ocr[current_page] = '\n'.join(buf).strip() page_files = sorted([f for f in os.listdir(pages_dir) if f.endswith('.png')]) print(f"Pages: {len(page_files)}, OCR entries: {len(ocr)}", flush=True) # Styles styles = getSampleStyleSheet() W, H = A4 avail_w = W - 4*cm title_sty = ParagraphStyle('TI', parent=styles['Title'], fontSize=20, textColor=colors.HexColor('#1a237e'), alignment=TA_CENTER, spaceAfter=10) sub_sty = ParagraphStyle('SU', parent=styles['Normal'], fontSize=12, alignment=TA_CENTER, textColor=colors.HexColor('#444444'), spaceAfter=6) pg_hdr_sty = ParagraphStyle('PH', parent=styles['Normal'], fontSize=8, textColor=colors.HexColor('#888888'), alignment=TA_CENTER, spaceAfter=4) label_sty = ParagraphStyle('LB', parent=styles['Normal'], fontSize=9, fontName='Helvetica-Bold', textColor=colors.HexColor('#1a237e'), spaceBefore=8, spaceAfter=3) text_sty = ParagraphStyle('TX', parent=styles['Normal'], fontSize=10.5, leading=16, spaceAfter=4, fontName='Helvetica') doc = SimpleDocTemplate(output_pdf, pagesize=A4, rightMargin=2*cm, leftMargin=2*cm, topMargin=2*cm, bottomMargin=2*cm) story = [] # ── Title page ── story.append(Spacer(1, 2*cm)) story.append(Paragraph("UROLOGY", title_sty)) story.append(Paragraph("Question Bank – English Edition", sub_sty)) story.append(Spacer(1, 0.4*cm)) story.append(HRFlowable(width='80%', thickness=1, color=colors.HexColor('#1a237e'), spaceAfter=12)) story.append(Paragraph(f"Total exam cards: {len(page_files)}", sub_sty)) story.append(PageBreak()) # ── Content pages ── for i, pf in enumerate(page_files): img_path = os.path.join(pages_dir, pf) page_num = i + 1 story.append(Paragraph(f"Card {page_num} / {len(page_files)}", pg_hdr_sty)) story.append(HRFlowable(width='100%', thickness=0.5, color=colors.HexColor('#cccccc'), spaceAfter=6)) # ── original scan image ── try: img = Image.open(img_path) iw, ih = img.size dw = avail_w dh = dw * (ih / iw) max_h = H * 0.52 if dh > max_h: dh = max_h dw = dh * (iw / ih) story.append(RLImage(img_path, width=dw, height=dh)) except Exception as e: story.append(Paragraph(f"[Image unavailable: {e}]", text_sty)) story.append(Spacer(1, 0.2*cm)) # ── OCR text ── raw = ocr.get(pf, '').strip() if raw: story.append(Paragraph("Extracted Text (English):", label_sty)) lines = [l.strip() for l in raw.split('\n') if l.strip()] # Filter out obvious garbage lines (pure symbols / very short junk) clean = [] for ln in lines: alnum = sum(c.isalnum() for c in ln) if alnum >= 3 or (len(ln) <= 6 and alnum >= 1): clean.append(ln) if clean: combined = '<br/>'.join( l.replace('&','&amp;').replace('<','&lt;').replace('>','&gt;') for l in clean) story.append(Paragraph(combined, text_sty)) else: story.append(Paragraph("[Text not clearly legible – see image above]", text_sty)) else: story.append(Paragraph("[No text detected – see image above]", text_sty)) story.append(PageBreak()) print("Building PDF...", flush=True) doc.build(story) sz = os.path.getsize(output_pdf) print(f"Done: {output_pdf} ({sz/1024/1024:.1f} MB)", flush=True) PYEOF

ls /home/daytona/workspace/urology-out/pages/ | sort

# Convert pages to web-accessible JPEGs for viewing (smaller size) mkdir -p /home/daytona/workspace/urology-out/thumbs for i in $(seq -w 1 33); do python3 -c " from PIL import Image img = Image.open('/home/daytona/workspace/urology-out/pages/p-$i.png') img.save('/home/daytona/workspace/urology-out/thumbs/p-$i.jpg', 'JPEG', quality=85) print('saved $i') " 2>/dev/null || echo "skip $i" done ls /home/daytona/workspace/urology-out/thumbs/ | wc -l

python3 -c " from PIL import Image for i in range(1, 34): n = str(i).zfill(2) try: img = Image.open(f'/home/daytona/workspace/urology-out/pages/p-{n}.png') img.save(f'/home/daytona/workspace/urology-out/thumbs/p-{n}.jpg', 'JPEG', quality=80) except: pass import os; print(len(os.listdir('/home/daytona/workspace/urology-out/thumbs')), 'done') "

# Images are large PNGs - use mogrify to batch convert quickly mkdir -p /home/daytona/workspace/urology-out/thumbs which mogrify && mogrify -format jpg -quality 75 -path /home/daytona/workspace/urology-out/thumbs /home/daytona/workspace/urology-out/pages/*.png & echo "started" sleep 50 ls /home/daytona/workspace/urology-out/thumbs/ | wc -l

sleep 20 && ls /home/daytona/workspace/urology-out/thumbs/ | wc -l