#!/usr/bin/env python3

# pdfspotthediff-20251230
# djb
# SPDX-License-Identifier: LicenseRef-PD-hp OR CC0-1.0 OR 0BSD OR MIT-0 OR MIT

# dependencies: apt install python3-fitz (new name: python3-pymupdf)
# usage: pdfspotthediff old.pdf new.pdf
# output: old-spotthediff.pdf and new-spotthediff.pdf
#         (overwriting those if they exist already)

# some caveats:
# * does not mark changes in positioning, color, font, glyphs, images
# * for highlighting, does not avoid colors already used in document

# comparison to (some) alternatives:
# * diff-pdf-wx: visual differences; each page separately; comparison window
# * diffpdf: word differences; each page separately; comparison window
# * https://github.com/malavika-suresh/multiple_pdf_comparison: word differences; each page separately; comparison PDF
# * https://github.com/cascremers/pdfdiff: sentence differences; crosses pages; vimdiff comparison
# * https://github.com/ssibb/PDF-Diff-Viewer: word differences; crosses pages; comparison window
# * this tool: pdfspotthediff: word differences; crosses pages; two comparison PDFs

tag2color = {
  'delete': (1.0,0.8,0.8), # light red in old-spotthediff.pdf for removed text
  'insert': (0.8,1.0,1.0), # light blue in new-spotthediff.pdf for added text
  'replace': (1.0,1.0,0.2), # light yellow in both for modified text
}
notecolor = 0.8,1.0,0.8 # light green for words before+after blue/red in other PDF

import sys
import difflib
try:
  import pymupdf
except:
  import fitz as pymupdf

pdf1 = pymupdf.open(sys.argv[1])
pdf2 = pymupdf.open(sys.argv[2])

words1 = [w[4] for page in pdf1 for w in page.get_text('words',sort=False)]
words2 = [w[4] for page in pdf2 for w in page.get_text('words',sort=False)]

diffs = difflib.SequenceMatcher(None,words1,words2,autojunk=False)

pos2color1 = {}
pos2color2 = {}

for tag,i1,i2,j1,j2 in diffs.get_opcodes():
  if tag == 'equal': continue
  for i in range(i1,i2): pos2color1[i] = tag2color[tag]
  if i1 == i2:
    pos2color1[i1-0.5] = notecolor
    pos2color1[i1+0.5] = notecolor
  for j in range(j1,j2): pos2color2[j] = tag2color[tag]
  if j1 == j2:
    pos2color2[j1-0.5] = notecolor
    pos2color2[j1+0.5] = notecolor

def highlight(pdf,pos2color):
  pos = 0
  for page in pdf:
    for w in page.get_text('words',sort=False):
      for offset in 0,0.5:
        if pos+offset in pos2color:
          annotation = page.add_highlight_annot(pymupdf.Rect(w[:4]))
          annotation.set_colors(stroke=pos2color[pos+offset])
          annotation.update()
      pos += 1

highlight(pdf1,pos2color1)
highlight(pdf2,pos2color2)

for pdf,fn in (pdf1,sys.argv[1]),(pdf2,sys.argv[2]):
  if fn.endswith('.pdf'): fn = fn[:-4]
  pdf.save(fn+'-spotthediff.pdf')