cd "/home/daytona/workspace/attachments/8c81eed9-ce97-4963-9229-861dfaf2c8ad/" && python3 -c " import pdfplumber with pdfplumber.open('DOC-20260710-WA0006. (1).pdf') as pdf: print(f'Total pages: {len(pdf.pages)}') # Inspect first 3 pages to understand header/footer regions for i in range(min(3, len(pdf.pages))): page = pdf.pages[i] print(f'\n--- Page {i+1} ---') print(f'Page size: width={page.width}, height={page.height}') # Get all text with bounding boxes words = page.extract_words() if words: # Show top items (potential header) print('TOP 5 words (potential header):') sorted_top = sorted(words, key=lambda w: float(w[\"top\"])) for w in sorted_top[:5]: print(f' text={w[\"text\"]!r}, top={w[\"top\"]}, bottom={w[\"bottom\"]}, x0={w[\"x0\"]}') # Show bottom items (potential footer) print('BOTTOM 5 words (potential footer):') sorted_bot = sorted(words, key=lambda w: float(w[\"top\"]), reverse=True) for w in sorted_bot[:5]: print(f' text={w[\"text\"]!r}, top={w[\"top\"]}, bottom={w[\"bottom\"]}, x0={w[\"x0\"]}') "

python3 -c " import pypdf from pypdf import PdfReader, PdfWriter from pypdf.generic import ArrayObject, FloatObject, NameObject, RectangleObject import io input_path = '/home/daytona/workspace/attachments/8c81eed9-ce97-4963-9229-861dfaf2c8ad/DOC-20260710-WA0006. (1).pdf' output_path = '/home/daytona/workspace/attachments/8c81eed9-ce97-4963-9229-861dfaf2c8ad/DOC-20260710-no_header_footer.pdf' reader = PdfReader(input_path) writer = PdfWriter() # Page dimensions: 595.25 x 842 # Header region: top=29.75, bottom=37.75 (pdfplumber coords, origin top-left) # Footer region: top=805.475, bottom=813.475 # In PDF coords (origin bottom-left): # header_pdf_bottom = 842 - 37.75 = 804.25, header_pdf_top = 842 - 29.75 = 812.25 # footer_pdf_bottom = 842 - 813.475 = 28.525, footer_pdf_top = 842 - 805.475 = 36.525 # We will add white rectangle annotations to cover header and footer on each page from pypdf.generic import ( DictionaryObject, ArrayObject, FloatObject, NameObject, NumberObject, ByteStringObject, StreamObject, DecodedStreamObject ) # Use page cropping: keep only the body area # Body starts after header ends: pdfplumber y ~45 (adding a small margin) # Body ends before footer starts: pdfplumber y ~800 # In PDF coordinates (origin bottom-left, y increases upward): # pdfplumber top y -> pdf y = page_height - pdfplumber_top # Header bottom in pdfplumber = 37.75 -> PDF y = 842 - 37.75 = 804.25 # Footer top in pdfplumber = 805.475 -> PDF y = 842 - 805.475 = 36.525 # We'll crop each page to exclude header and footer with a small margin # Keep area from PDF y=45 (above footer) to PDF y=800 (below header) # That means: mediabox bottom=45, top=800 (in original PDF coords) # But we want to KEEP content and just HIDE header/footer, not change content area # Better approach: use white annotation rectangles to cover header/footer import pikepdf with pikepdf.open(input_path) as pdf: for page in pdf.pages: page_height = float(page.mediabox[3]) page_width = float(page.mediabox[2]) # Header: pdfplumber y 29.75 to 37.75 -> PDF bottom-left coords # pdfplumber uses top-left origin, so PDF y = page_height - pdfplumber_y header_pdf_y_bottom = page_height - 45 # a bit of margin below header header_pdf_y_top = page_height # top of page # Footer: pdfplumber y 805 to 814 -> PDF y footer_pdf_y_bottom = 0 # bottom of page footer_pdf_y_top = page_height - 800 # a bit above footer top # Add white rectangle annotations annots = [] for (x0, y0, x1, y1) in [ (0, header_pdf_y_bottom, page_width, header_pdf_y_top), # header (0, footer_pdf_y_bottom, page_width, footer_pdf_y_top), # footer ]: annot = pikepdf.Dictionary( Type=pikepdf.Name('/Annot'), Subtype=pikepdf.Name('/Square'), Rect=pikepdf.Array([x0, y0, x1, y1]), C=pikepdf.Array([1, 1, 1]), # white border IC=pikepdf.Array([1, 1, 1]), # white fill BS=pikepdf.Dictionary(W=0), # no border width F=pikepdf.Integer(4), # print flag ) annots.append(pdf.make_indirect(annot)) if '/Annots' in page: page['/Annots'].extend(annots) else: page['/Annots'] = pikepdf.Array(annots) pdf.save(output_path) print('Done! Saved to:', output_path) import os print('File size:', os.path.getsize(output_path), 'bytes') "

python3 -c " import reportlab from reportlab.lib.pagesizes import A4 from pypdf import PdfReader, PdfWriter from reportlab.pdfgen import canvas import io input_path = '/home/daytona/workspace/attachments/8c81eed9-ce97-4963-9229-861dfaf2c8ad/DOC-20260710-WA0006. (1).pdf' output_path = '/home/daytona/workspace/attachments/8c81eed9-ce97-4963-9229-861dfaf2c8ad/DOC-20260710-no_header_footer.pdf' reader = PdfReader(input_path) writer = PdfWriter() num_pages = len(reader.pages) print(f'Processing {num_pages} pages...') for i, page in enumerate(reader.pages): page_width = float(page.mediabox.width) page_height = float(page.mediabox.height) # Create a white rectangle overlay page using reportlab # Header region: pdfplumber top=29.75 bottom=37.75 # In PDF coords (y from bottom): # header top = page_height - 29.75, header bottom = page_height - 37.75 # Add generous margin: cover from page_height-45 to page_height (top) # Footer: cover from 0 to page_height - 800 packet = io.BytesIO() c = canvas.Canvas(packet, pagesize=(page_width, page_height)) c.setFillColorRGB(1, 1, 1) # white c.setStrokeColorRGB(1, 1, 1) # Cover header (top strip) header_cover_bottom = page_height - 45 c.rect(0, header_cover_bottom, page_width, 45, fill=1, stroke=0) # Cover footer (bottom strip) footer_cover_top = page_height - 800 # ~42 pts from bottom c.rect(0, 0, page_width, footer_cover_top, fill=1, stroke=0) c.save() packet.seek(0) # Merge overlay onto original page from pypdf import PdfReader as PR overlay_reader = PR(packet) overlay_page = overlay_reader.pages[0] # Merge: original page first, then white overlay on top page.merge_page(overlay_page) writer.add_page(page) with open(output_path, 'wb') as f: writer.write(f) import os print('Done! Output:', output_path) print('File size:', os.path.getsize(output_path), 'bytes') "

python3 -c " import pdfplumber output_path = '/home/daytona/workspace/attachments/8c81eed9-ce97-4963-9229-861dfaf2c8ad/DOC-20260710-no_header_footer.pdf' with pdfplumber.open(output_path) as pdf: print(f'Total pages: {len(pdf.pages)}') for i in range(min(3, len(pdf.pages))): page = pdf.pages[i] words = page.extract_words() if words: print(f'\n--- Page {i+1} ---') sorted_top = sorted(words, key=lambda w: float(w['top'])) print('Top 5 words:') for w in sorted_top[:5]: print(f' text={w[\"text\"]!r}, top={w[\"top\"]:.2f}') sorted_bot = sorted(words, key=lambda w: float(w['top']), reverse=True) print('Bottom 5 words:') for w in sorted_bot[:5]: print(f' text={w[\"text\"]!r}, top={w[\"top\"]:.2f}') "

python3 -c " from pypdf import PdfReader, PdfWriter from reportlab.pdfgen import canvas from reportlab.lib.colors import white import io input_path = '/home/daytona/workspace/attachments/8c81eed9-ce97-4963-9229-861dfaf2c8ad/DOC-20260710-WA0006. (1).pdf' output_path = '/home/daytona/workspace/attachments/8c81eed9-ce97-4963-9229-861dfaf2c8ad/DOC-20260710-no_header_footer.pdf' reader = PdfReader(input_path) writer = PdfWriter() for i, page in enumerate(reader.pages): page_width = float(page.mediabox.width) page_height = float(page.mediabox.height) # pdfplumber found: # Header text at top=29.75, bottom=37.75 (from top of page) # Footer text at top=805.475, bottom=813.475 (from top of page) # Page height = 842 # In PDF coordinate system (origin bottom-left, y increases upward): # Header occupies: y from (842-37.75)=804.25 to (842-29.75)=812.25 # We'll whiteout: y from 795 to 842 (top 47 pts, generous margin) # Footer occupies: y from (842-813.475)=28.525 to (842-805.475)=36.525 # We'll whiteout: y from 0 to 47 (bottom 47 pts, generous margin) # Strategy: use cropbox to show only the body content area, # which effectively hides header and footer from view AND from extraction # CropBox in PDF coords: left, bottom, right, top body_bottom = 47 # hide footer region body_top = page_height - 47 # hide header region # Set cropbox to clip header and footer page.cropbox.lower_left = (0, body_bottom) page.cropbox.upper_right = (page_width, body_top) writer.add_page(page) with open(output_path, 'wb') as f: writer.write(f) import os print('Done! Output:', output_path) print('File size:', os.path.getsize(output_path), 'bytes') "

python3 -c " import pdfplumber output_path = '/home/daytona/workspace/attachments/8c81eed9-ce97-4963-9229-861dfaf2c8ad/DOC-20260710-no_header_footer.pdf' with pdfplumber.open(output_path) as pdf: print(f'Total pages: {len(pdf.pages)}') for i in range(min(3, len(pdf.pages))): page = pdf.pages[i] words = page.extract_words() if words: print(f'\n--- Page {i+1} | Page dims: {page.width:.1f} x {page.height:.1f} ---') sorted_top = sorted(words, key=lambda w: float(w['top'])) print('Top 5 words:') for w in sorted_top[:5]: print(f' text={w[\"text\"]!r}, top={w[\"top\"]:.2f}') sorted_bot = sorted(words, key=lambda w: float(w['top']), reverse=True) print('Bottom 5 words:') for w in sorted_bot[:5]: print(f' text={w[\"text\"]!r}, top={w[\"top\"]:.2f}') else: print(f'\n--- Page {i+1}: no words extracted ---') "

python3 -c "import fitz; print(fitz.__version__)"

pip install --user pikepdf 2>&1 | tail -5

python3 -c " import pikepdf import io input_path = '/home/daytona/workspace/attachments/8c81eed9-ce97-4963-9229-861dfaf2c8ad/DOC-20260710-WA0006. (1).pdf' output_path = '/home/daytona/workspace/attachments/8c81eed9-ce97-4963-9229-861dfaf2c8ad/DOC-20260710-no_header_footer.pdf' with pikepdf.open(input_path) as pdf: for page in pdf.pages: mediabox = page.mediabox page_width = float(mediabox[2]) page_height = float(mediabox[3]) # Header: pdfplumber top=29.75 to 37.75 (from top) # PDF y (from bottom): header top = page_height - 29.75 = 812.25 # header bottom = page_height - 37.75 = 804.25 # Footer: pdfplumber top=805.475 to 813.475 (from top) # PDF y: footer top = page_height - 805.475 = 36.525 # footer bottom = page_height - 813.475 = 28.525 # Set CropBox to hide header and footer # Keep body: from y=44 (above footer) to y=800 (below header) crop_bottom = 44.0 crop_top = page_height - 44.0 page.cropbox = pikepdf.Array([ pikepdf.Real(0), pikepdf.Real(crop_bottom), pikepdf.Real(page_width), pikepdf.Real(crop_top) ]) # Also set MediaBox equal to CropBox so content is actually clipped # and redraw with a clipping path in the content stream pdf.save(output_path) print('Saved with CropBox.') # Now verify visual removal using pdfplumber on cropped region import pdfplumber with pdfplumber.open(output_path) as pdf2: page = pdf2.pages[0] # Restrict to cropbox area print(f'Page bbox: {page.bbox}') words = page.extract_words() sorted_top = sorted(words, key=lambda w: float(w['top'])) print('Top 5 words after crop:') for w in sorted_top[:5]: print(f' text={w[\"text\"]!r}, top={w[\"top\"]:.2f}') sorted_bot = sorted(words, key=lambda w: float(w['top']), reverse=True) print('Bottom 5 words after crop:') for w in sorted_bot[:5]: print(f' text={w[\"text\"]!r}, top={w[\"top\"]:.2f}') "

ls -la "/home/daytona/workspace/attachments/8c81eed9-ce97-4963-9229-861dfaf2c8ad/" && mkdir -p /home/daytona/workspace/output/

python3 -c " import pikepdf input_path = '/home/daytona/workspace/attachments/8c81eed9-ce97-4963-9229-861dfaf2c8ad/DOC-20260710-WA0006. (1).pdf' output_path = '/home/daytona/workspace/output/DOC-20260710-no_header_footer.pdf' with pikepdf.open(input_path) as pdf: for page in pdf.pages: mediabox = page.mediabox page_width = float(mediabox[2]) page_height = float(mediabox[3]) # Crop away header (top 44pt) and footer (bottom 44pt) crop_bottom = 44.0 crop_top = page_height - 44.0 page.cropbox = pikepdf.Array([ pikepdf.Real(0), pikepdf.Real(crop_bottom), pikepdf.Real(page_width), pikepdf.Real(crop_top) ]) pdf.save(output_path) print('Saved.') import os print('File size:', os.path.getsize(output_path), 'bytes') # Verify import pdfplumber with pdfplumber.open(output_path) as pdf2: page = pdf2.pages[0] print(f'Page bbox: {page.bbox}') words = page.extract_words() if words: sorted_top = sorted(words, key=lambda w: float(w['top'])) print('Top 5 words:') for w in sorted_top[:5]: print(f' {w[\"text\"]!r} top={w[\"top\"]:.2f}') sorted_bot = sorted(words, key=lambda w: float(w['top']), reverse=True) print('Bottom 5 words:') for w in sorted_bot[:5]: print(f' {w[\"text\"]!r} top={w[\"top\"]:.2f}') "

python3 << 'EOF' import pikepdf input_path = '/home/daytona/workspace/attachments/8c81eed9-ce97-4963-9229-861dfaf2c8ad/DOC-20260710-WA0006. (1).pdf' output_path = '/home/daytona/workspace/output/DOC-20260710-no_header_footer.pdf' # Margins to remove: header top 44pt, footer bottom 44pt HEADER_CUT = 44.0 FOOTER_CUT = 44.0 with pikepdf.open(input_path) as pdf: for page in pdf.pages: mediabox = page.mediabox page_height = float(mediabox[3]) page_width = float(mediabox[2]) clip_bottom = FOOTER_CUT clip_top = page_height - HEADER_CUT # Build a prefix that: # q - save graphics state # W n - set clipping path (the rect we draw) # and a suffix that restores state # We wrap existing content stream with: # q <clip rect> W n ... original content ... Q clip_prefix = ( f"q " f"{0} {clip_bottom} {page_width} {clip_top - clip_bottom} re W n " ).encode() clip_suffix = b" Q" # Collect existing content streams into a single bytes blob if "/Contents" not in page: continue contents_obj = page["/Contents"] existing_data = b"" if isinstance(contents_obj, pikepdf.Array): for ref in contents_obj: stream_obj = ref existing_data += stream_obj.read_bytes() else: existing_data = contents_obj.read_bytes() # Build new combined stream new_data = clip_prefix + existing_data + clip_suffix # Replace with a single new stream new_stream = pikepdf.Stream(pdf, new_data) page["/Contents"] = new_stream # Also set CropBox = clipped region so page size reflects body only page.cropbox = pikepdf.Array([ pikepdf.Real(0), pikepdf.Real(clip_bottom), pikepdf.Real(page_width), pikepdf.Real(clip_top) ]) pdf.save(output_path) print("Done. Saved to:", output_path) import os print("File size:", os.path.getsize(output_path)) # Quick verification with pdfplumber using cropbox import pdfplumber with pdfplumber.open(output_path) as pdf2: page = pdf2.pages[0] # Manually crop to the visible region cropped = page.within_bbox((0, 44, page.width, page.height - 44)) words = cropped.extract_words() if words: sorted_top = sorted(words, key=lambda w: float(w['top'])) print("Top 5 words (body only):") for w in sorted_top[:5]: print(f" {w['text']!r} top={w['top']:.2f}") sorted_bot = sorted(words, key=lambda w: float(w['top']), reverse=True) print("Bottom 5 words (body only):") for w in sorted_bot[:5]: print(f" {w['text']!r} top={w['top']:.2f}") EOF

rm /home/daytona/workspace/output/DOC-20260710-no_header_footer.pdf 2>/dev/null; ls /home/daytona/workspace/output/

python3 << 'EOF' import pikepdf input_path = '/home/daytona/workspace/attachments/8c81eed9-ce97-4963-9229-861dfaf2c8ad/DOC-20260710-WA0006. (1).pdf' output_path = '/home/daytona/workspace/output/DOC-20260710-no_header_footer.pdf' HEADER_CUT = 44.0 FOOTER_CUT = 44.0 with pikepdf.open(input_path) as pdf: for page in pdf.pages: mediabox = page.mediabox page_height = float(mediabox[3]) page_width = float(mediabox[2]) clip_bottom = FOOTER_CUT clip_top = page_height - HEADER_CUT clip_prefix = ( f"q {0} {clip_bottom} {page_width} {clip_top - clip_bottom} re W n " ).encode() clip_suffix = b" Q" if "/Contents" not in page: continue contents_obj = page["/Contents"] existing_data = b"" if isinstance(contents_obj, pikepdf.Array): for ref in contents_obj: existing_data += ref.read_bytes() else: existing_data = contents_obj.read_bytes() new_data = clip_prefix + existing_data + clip_suffix new_stream = pikepdf.Stream(pdf, new_data) page["/Contents"] = new_stream page.cropbox = pikepdf.Array([ pikepdf.Real(0), pikepdf.Real(clip_bottom), pikepdf.Real(page_width), pikepdf.Real(clip_top) ]) # Save to a temp path first, then copy import tempfile, shutil, os with tempfile.NamedTemporaryFile(suffix='.pdf', delete=False) as tmp: tmp_path = tmp.name pdf.save(tmp_path) shutil.copy2(tmp_path, output_path) os.unlink(tmp_path) print("Done. Saved to:", output_path) import os print("File size:", os.path.getsize(output_path)) import pdfplumber with pdfplumber.open(output_path) as pdf2: page = pdf2.pages[0] cropped = page.within_bbox((0, 44, page.width, page.height - 44)) words = cropped.extract_words() if words: sorted_top = sorted(words, key=lambda w: float(w['top'])) print("Top 5 words in body:") for w in sorted_top[:5]: print(f" {w['text']!r} top={w['top']:.2f}") sorted_bot = sorted(words, key=lambda w: float(w['top']), reverse=True) print("Bottom 5 words in body:") for w in sorted_bot[:5]: print(f" {w['text']!r} top={w['top']:.2f}") # Verify header/footer area is empty header_region = page.within_bbox((0, 0, page.width, 44)) header_words = header_region.extract_words() print(f"\nHeader region words (should be 0): {len(header_words)}") footer_region = page.within_bbox((0, page.height - 44, page.width, page.height)) footer_words = footer_region.extract_words() print(f"Footer region words (should be 0): {len(footer_words)}") EOF

mkdir -p /home/daytona/workspace/output && ls /home/daytona/workspace/output/

python3 << 'EOF' import pikepdf import tempfile, shutil, os input_path = '/home/daytona/workspace/attachments/8c81eed9-ce97-4963-9229-861dfaf2c8ad/DOC-20260710-WA0006. (1).pdf' output_path = '/home/daytona/workspace/output/DOC-20260710-no_header_footer.pdf' HEADER_CUT = 44.0 FOOTER_CUT = 44.0 with pikepdf.open(input_path) as pdf: for page in pdf.pages: mediabox = page.mediabox page_height = float(mediabox[3]) page_width = float(mediabox[2]) clip_bottom = FOOTER_CUT clip_top = page_height - HEADER_CUT clip_prefix = ( f"q {0} {clip_bottom} {page_width} {clip_top - clip_bottom} re W n " ).encode() clip_suffix = b" Q" if "/Contents" not in page: continue contents_obj = page["/Contents"] existing_data = b"" if isinstance(contents_obj, pikepdf.Array): for ref in contents_obj: existing_data += ref.read_bytes() else: existing_data = contents_obj.read_bytes() new_data = clip_prefix + existing_data + clip_suffix new_stream = pikepdf.Stream(pdf, new_data) page["/Contents"] = new_stream page.cropbox = pikepdf.Array([ pikepdf.Real(0), pikepdf.Real(clip_bottom), pikepdf.Real(page_width), pikepdf.Real(clip_top) ]) # Save to temp file, then move to output with tempfile.NamedTemporaryFile(suffix='.pdf', delete=False, dir='/home/daytona/workspace/output/') as tmp: tmp_path = tmp.name pdf.save(tmp_path) os.rename(tmp_path, output_path) print("Done. Saved to:", output_path) print("File size:", os.path.getsize(output_path)) import pdfplumber with pdfplumber.open(output_path) as pdf2: page = pdf2.pages[0] # Check body words body = page.within_bbox((0, 44, page.width, page.height - 44)) words = body.extract_words() if words: sorted_top = sorted(words, key=lambda w: float(w['top'])) print("\nTop 5 body words:") for w in sorted_top[:5]: print(f" {w['text']!r} top={w['top']:.2f}") sorted_bot = sorted(words, key=lambda w: float(w['top']), reverse=True) print("Bottom 5 body words:") for w in sorted_bot[:5]: print(f" {w['text']!r} top={w['top']:.2f}") # Check header/footer removed hdr = page.within_bbox((0, 0, page.width, 44)) ftr = page.within_bbox((0, page.height - 44, page.width, page.height)) print(f"\nHeader words remaining: {len(hdr.extract_words())}") print(f"Footer words remaining: {len(ftr.extract_words())}") print(f"\nTotal pages: {len(pdf2.pages)}") EOF

python3 << 'EOF' import pikepdf import io, os input_path = '/home/daytona/workspace/attachments/8c81eed9-ce97-4963-9229-861dfaf2c8ad/DOC-20260710-WA0006. (1).pdf' output_path = '/home/daytona/workspace/output/DOC-20260710-no_header_footer.pdf' HEADER_CUT = 44.0 FOOTER_CUT = 44.0 buf = io.BytesIO() with pikepdf.open(input_path) as pdf: for page in pdf.pages: mediabox = page.mediabox page_height = float(mediabox[3]) page_width = float(mediabox[2]) clip_bottom = FOOTER_CUT clip_top = page_height - HEADER_CUT clip_prefix = f"q {0} {clip_bottom} {page_width} {clip_top - clip_bottom} re W n ".encode() clip_suffix = b" Q" if "/Contents" not in page: continue contents_obj = page["/Contents"] existing_data = b"" if isinstance(contents_obj, pikepdf.Array): for ref in contents_obj: existing_data += ref.read_bytes() else: existing_data = contents_obj.read_bytes() new_data = clip_prefix + existing_data + clip_suffix page["/Contents"] = pikepdf.Stream(pdf, new_data) page.cropbox = pikepdf.Array([ pikepdf.Real(0), pikepdf.Real(clip_bottom), pikepdf.Real(page_width), pikepdf.Real(clip_top) ]) pdf.save(buf) buf.seek(0) with open(output_path, 'wb') as f: f.write(buf.read()) print("Done. Saved to:", output_path) print("File size:", os.path.getsize(output_path)) # Verify import pdfplumber with pdfplumber.open(output_path) as pdf2: page = pdf2.pages[0] hdr = page.within_bbox((0, 0, page.width, 44)) ftr = page.within_bbox((0, page.height - 44, page.width, page.height)) body = page.within_bbox((0, 44, page.width, page.height - 44)) body_words = body.extract_words() sorted_top = sorted(body_words, key=lambda w: float(w['top'])) print("\nTop 5 body words:") for w in sorted_top[:5]: print(f" {w['text']!r} top={w['top']:.2f}") sorted_bot = sorted(body_words, key=lambda w: float(w['top']), reverse=True) print("Bottom 5 body words:") for w in sorted_bot[:5]: print(f" {w['text']!r} top={w['top']:.2f}") print(f"\nHeader region words: {len(hdr.extract_words())} (should be 0)") print(f"Footer region words: {len(ftr.extract_words())} (should be 0)") print(f"Total pages: {len(pdf2.pages)}") EOF