Add scripts/fill_flat_pdf.py (v1.0.0)

2026-07-20 15:00:27 -05:00
parent 6e665f40d5
commit 24c32614b2
1 changed files with 100 additions and 0 deletions
@@ -0,0 +1,100 @@
+#!/usr/bin/env python3
+"""
+fill_flat_pdf.py — overlay text onto a flat (non-fillable) PDF.
+
+For forms that look like forms but have no AcroForm fields. You supply the
+placements (page index, x, y in PDF points from the bottom-left, and the text);
+this stamps them onto a copy of the PDF.
+
+Usage:
+    python fill_flat_pdf.py INPUT.pdf OUTPUT.pdf PLACEMENTS.json
+    # or pass placements inline:
+    python fill_flat_pdf.py INPUT.pdf OUTPUT.pdf '[{"page":0,"x":150,"y":690,"text":"Message Point Media of Alabama Inc."}]'
+
+PLACEMENTS.json format:
+    [
+      {"page": 0, "x": 150, "y": 690, "text": "Message Point Media of Alabama Inc.", "size": 10},
+      {"page": 0, "x": 150, "y": 668, "text": "4628 Montevallo Rd, Suite 204"}
+    ]
+
+Coordinates: PDF origin is bottom-left; y increases upward; 72 points = 1 inch.
+If your source coordinates are top-left, convert with y_pdf = page_height - y_top.
+Use --list-sizes to print each page's width/height in points to help you place text.
+
+Requires: pypdf and reportlab  (pip install pypdf reportlab --break-system-packages)
+"""
+import sys
+import json
+import io
+
+
+def page_sizes(input_pdf):
+    from pypdf import PdfReader
+    reader = PdfReader(input_pdf)
+    for i, page in enumerate(reader.pages):
+        box = page.mediabox
+        print(f"page {i}: width={float(box.width):.1f} height={float(box.height):.1f} (points)")
+
+
+def fill(input_pdf, output_pdf, placements):
+    from pypdf import PdfReader, PdfWriter
+    from reportlab.pdfgen import canvas
+
+    reader = PdfReader(input_pdf)
+    writer = PdfWriter()
+
+    # group placements by page
+    by_page = {}
+    for p in placements:
+        by_page.setdefault(int(p.get("page", 0)), []).append(p)
+
+    for i, page in enumerate(reader.pages):
+        if i in by_page:
+            box = page.mediabox
+            w, h = float(box.width), float(box.height)
+            buf = io.BytesIO()
+            c = canvas.Canvas(buf, pagesize=(w, h))
+            for p in by_page[i]:
+                size = float(p.get("size", 10))
+                c.setFont(p.get("font", "Helvetica"), size)
+                c.drawString(float(p["x"]), float(p["y"]), str(p["text"]))
+            c.save()
+            buf.seek(0)
+            overlay = PdfReader(buf).pages[0]
+            page.merge_page(overlay)
+        writer.add_page(page)
+
+    with open(output_pdf, "wb") as f:
+        writer.write(f)
+    print(f"Wrote {output_pdf} with {sum(len(v) for v in by_page.values())} text placements.")
+
+
+def main():
+    args = sys.argv[1:]
+    if args and args[0] == "--list-sizes":
+        if len(args) != 2:
+            print("Usage: python fill_flat_pdf.py --list-sizes INPUT.pdf")
+            sys.exit(1)
+        page_sizes(args[1])
+        return
+
+    if len(args) != 3:
+        print(__doc__)
+        sys.exit(1)
+
+    input_pdf, output_pdf, placements_arg = args
+    try:
+        placements = json.loads(placements_arg)
+    except json.JSONDecodeError:
+        with open(placements_arg) as f:
+            placements = json.load(f)
+
+    if not isinstance(placements, list):
+        print("Placements must be a JSON array of {page,x,y,text} objects.")
+        sys.exit(1)
+
+    fill(input_pdf, output_pdf, placements)
+
+
+if __name__ == "__main__":
+    main()