python3 -c " import pdfplumber pdf_path = '/tmp/workspace/attachments/ee7ed6d1-0e27-4485-afdb-12e75ee6f2b6/Adobe Scan Jul 14, 2026_1.pdf' with pdfplumber.open(pdf_path) as pdf: print(f'Total pages: {len(pdf.pages)}') for i, page in enumerate(pdf.pages): text = page.extract_text() print(f'--- Page {i+1} ---') print(text) print() "

python3 -c " from pypdf import PdfReader import os pdf_path = '/tmp/workspace/attachments/ee7ed6d1-0e27-4485-afdb-12e75ee6f2b6/Adobe Scan Jul 14, 2026_1.pdf' reader = PdfReader(pdf_path) print(f'Pages: {len(reader.pages)}') # Check if there are images for i, page in enumerate(reader.pages[:5]): images = page.images print(f'Page {i+1}: {len(images)} images') for j, img in enumerate(images): print(f' Image {j}: name={img.name}, size={len(img.data)} bytes') "

python3 -c " from pypdf import PdfReader import os pdf_path = '/tmp/workspace/attachments/ee7ed6d1-0e27-4485-afdb-12e75ee6f2b6/Adobe Scan Jul 14, 2026_1.pdf' os.makedirs('/tmp/workspace/case_pages', exist_ok=True) reader = PdfReader(pdf_path) for i, page in enumerate(reader.pages): images = page.images if images: img = images[0] out_path = f'/tmp/workspace/case_pages/page_{i+1:02d}.jpg' with open(out_path, 'wb') as f: f.write(img.data) print(f'Saved page {i+1} -> {out_path}') print('Done') "

python3 -c " # Try pytesseract for OCR try: import pytesseract from PIL import Image img = Image.open('/tmp/workspace/case_pages/page_01.jpg') text = pytesseract.image_to_string(img) print('Page 1:') print(text[:3000]) except Exception as e: print(f'pytesseract error: {e}') "

pip install --user pytesseract 2>&1 | tail -3; which tesseract 2>/dev/null || echo "tesseract not found"

# Try to serve files via a simple HTTP approach or use markitdown python3 -c " try: from markitdown import MarkItDown md = MarkItDown() result = md.convert('/tmp/workspace/attachments/ee7ed6d1-0e27-4485-afdb-12e75ee6f2b6/Adobe Scan Jul 14, 2026_1.pdf') print(result.text_content[:5000]) except Exception as e: print(f'markitdown error: {e}') "

# Start a simple HTTP server in background and test cd /tmp/workspace/case_pages && python3 -m http.server 8765 & sleep 2 echo "Server started" curl -s -o /dev/null -w "%{http_code}" http://localhost:8765/page_01.jpg

# Try easyocr as alternative OCR python3 -c " try: import easyocr reader = easyocr.Reader(['en'], verbose=False) result = reader.readtext('/tmp/workspace/case_pages/page_01.jpg', detail=0) print('EasyOCR result:') print('\n'.join(result[:50])) except Exception as e: print(f'easyocr error: {e}') " 2>&1 | head -30

# Try doctr for OCR python3 -c " try: from doctr.io import DocumentFile from doctr.models import ocr_predictor doc = DocumentFile.from_images(['/tmp/workspace/case_pages/page_01.jpg']) model = ocr_predictor(pretrained=True) result = model(doc) for page in result.pages: for block in page.blocks: for line in block.lines: print(' '.join([w.value for w in line.words])) except Exception as e: print(f'doctr error: {e}') " 2>&1 | head -20

# Try using PIL to look at the images directly and check if they are readable python3 -c " from PIL import Image img = Image.open('/tmp/workspace/case_pages/page_01.jpg') print(f'Size: {img.size}, Mode: {img.mode}') # Check image content - resize and check basic stats import numpy as np arr = np.array(img) print(f'Array shape: {arr.shape}') print(f'Min: {arr.min()}, Max: {arr.max()}, Mean: {arr.mean():.1f}') "

# Try using transformers for OCR (TrOCR or similar) python3 -c " try: from transformers import pipeline pipe = pipeline('image-to-text', model='microsoft/trocr-base-printed') from PIL import Image img = Image.open('/tmp/workspace/case_pages/page_01.jpg') # crop a small portion crop = img.crop((0, 0, 800, 200)) result = pipe(crop) print(result) except Exception as e: print(f'Error: {e}') " 2>&1 | head -20

# Check what's actually in the images - convert a small page to base64 and look python3 -c " from PIL import Image import base64, io img = Image.open('/tmp/workspace/case_pages/page_01.jpg') # Save a small thumbnail for visual inspection thumb = img.resize((300, 414)) thumb.save('/tmp/workspace/case_pages/thumb_01.jpg', quality=85) print('Thumbnail saved') # Also check if there's text by looking at pixel variance in small regions import numpy as np arr = np.array(img) # Look at top portion where title/header would be top = arr[:500, :, :] print(f'Top region mean: {top.mean():.1f}, std: {top.std():.1f}') "

# Try installing and using pytesseract with pre-built tesseract binary alternative # Check if any OCR tools are available which ocrmypdf 2>/dev/null || echo "no ocrmypdf" which tesseract 2>/dev/null || echo "no tesseract" python3 -c "import pdf2image" 2>&1 || echo "no pdf2image" python3 -c "import paddleocr" 2>&1 || echo "no paddleocr" python3 -c "import ocrmypdf" 2>&1 || echo "no ocrmypdf module" ls /usr/bin/ | grep -i ocr 2>/dev/null || echo "no ocr in /usr/bin" ls /usr/local/bin/ | grep -i tess 2>/dev/null || echo "no tesseract in /usr/local/bin"

# Last try: use web_extract with a data URL approach by encoding image as base64 # Also try to read the raw PDF bytes and check if there's any embedded text layer python3 -c " # Try to use pypdf to get raw text more aggressively from pypdf import PdfReader import io pdf_path = '/tmp/workspace/attachments/ee7ed6d1-0e27-4485-afdb-12e75ee6f2b6/Adobe Scan Jul 14, 2026_1.pdf' reader = PdfReader(pdf_path) # Try different text extraction strategies for i, page in enumerate(reader.pages[:3]): print(f'=== Page {i+1} ===') # Default extraction text = page.extract_text(extraction_mode='layout') if text and text.strip(): print('Layout mode:', repr(text[:500])) text2 = page.extract_text() if text2 and text2.strip(): print('Default mode:', repr(text2[:500])) # Try to access raw content if '/Contents' in page: print('Has Contents stream') print() "