annotate tests/gentle_compare.py @ 2:b50eed0cc0ef upstream

ADD: MuPDF v1.26.7: the MuPDF source as downloaded by a default build of PyMuPDF 1.26.4. The directory name has changed: no version number in the expanded directory now.
author Franz Glasner <fzglas.hg@dom66.de>
date Mon, 15 Sep 2025 11:43:07 +0200
parents 1d09e1dec1d9
children a6bc019ac0b2
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
1
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
1 import math
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
2
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
3 import pymupdf
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
4
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
5
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
6 def gentle_compare(w0, w1):
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
7 """Check lists of "words" extractions for approximate equality.
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
8
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
9 * both lists must have same length
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
10 * word items must contain same word strings
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
11 * word rectangles must be approximately equal
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
12 """
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
13 tolerance = 1e-3 # maximum (Euclidean) norm of difference rectangle
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
14 word_count = len(w0) # number of words
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
15 if word_count != len(w1):
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
16 print(f"different number of words: {word_count}/{len(w1)}")
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
17 return False
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
18 for i in range(word_count):
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
19 if w0[i][4] != w1[i][4]: # word strings must be the same
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
20 print(f"word {i} mismatch")
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
21 return False
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
22 r0 = pymupdf.Rect(w0[i][:4]) # rect of first word
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
23 r1 = pymupdf.Rect(w1[i][:4]) # rect of second word
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
24 delta = (r1 - r0).norm() # norm of difference rectangle
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
25 if delta > tolerance:
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
26 print(f"word {i}: rectangle mismatch {delta}")
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
27 return False
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
28 return True
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
29
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
30
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
31 def rms(a, b, verbose=None, out_prefix=''):
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
32 '''
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
33 Returns RMS diff of raw bytes of two sequences.
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
34 '''
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
35 if verbose is True:
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
36 verbose = 100000
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
37 assert len(a) == len(b)
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
38 e = 0
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
39 for i, (aa, bb) in enumerate(zip(a, b)):
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
40 if verbose and (i % verbose == 0):
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
41 print(f'{out_prefix}rms(): {i=} {e=} {aa=} {aa=}.')
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
42 e += (aa - bb) ** 2
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
43 rms = math.sqrt(e / len(a))
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
44 return rms
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
45
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
46
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
47 def pixmaps_rms(a, b, out_prefix=''):
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
48 '''
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
49 Returns RMS diff of raw bytes of two pixmaps.
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
50
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
51 We assert that the pixmaps/sequences are the same size.
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
52
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
53 <a> and <b> can each be a pymupdf.Pixmap or path of a bitmap file.
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
54 '''
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
55 if isinstance(a, str):
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
56 print(f'{out_prefix}pixmaps_rms(): reading pixmap from {a=}.')
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
57 a = pymupdf.Pixmap(a)
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
58 if isinstance(b, str):
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
59 print(f'{out_prefix}pixmaps_rms(): reading pixmap from {b=}.')
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
60 b = pymupdf.Pixmap(b)
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
61 assert a.irect == b.irect, f'Differing rects: {a.irect=} {b.irect=}.'
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
62 a_mv = a.samples_mv
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
63 b_mv = b.samples_mv
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
64 assert len(a_mv) == len(b_mv)
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
65 ret = rms(a_mv, b_mv, verbose=True, out_prefix=out_prefix)
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
66 print(f'{out_prefix}pixmaps_rms(): {ret=}.')
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
67 return ret
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
68
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
69
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
70 def pixmaps_diff(a, b, out_prefix=''):
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
71 '''
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
72 Returns a pymupdf.Pixmap that represents the difference between pixmaps <a>
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
73 and <b>.
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
74
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
75 Each byte in the returned pixmap is `128 + (b_byte - a_byte) // 2`.
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
76 '''
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
77 if isinstance(a, str):
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
78 print(f'{out_prefix}pixmaps_rms(): reading pixmap from {a=}.')
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
79 a = pymupdf.Pixmap(a)
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
80 if isinstance(b, str):
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
81 print(f'{out_prefix}pixmaps_rms(): reading pixmap from {b=}.')
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
82 b = pymupdf.Pixmap(b)
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
83 assert a.irect == b.irect, f'Differing rects: {a.irect=} {b.irect=}.'
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
84 a_mv = a.samples_mv
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
85 b_mv = b.samples_mv
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
86 c = pymupdf.Pixmap(a.tobytes())
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
87 c_mv = c.samples_mv
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
88 assert len(a_mv) == len(b_mv) == len(c_mv)
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
89 if 1:
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
90 print(f'{len(a_mv)=}')
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
91 for i, (a_byte, b_byte, c_byte) in enumerate(zip(a_mv, b_mv, c_mv)):
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
92 assert 0 <= a_byte < 256
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
93 assert 0 <= b_byte < 256
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
94 assert 0 <= c_byte < 256
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
95 # Set byte to 128 plus half the diff so we represent the full
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
96 # -255..+255 range.
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
97 c_mv[i] = 128 + (b_byte - a_byte) // 2
1d09e1dec1d9 ADD: PyMuPDF v1.26.4: the original sdist.
Franz Glasner <fzglas.hg@dom66.de>
parents:
diff changeset
98 return c