1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
|
#!/usr/bin/python
# -*- coding:utf8 -*-
# Copyright 2007 Behdad Esfahbod <besfahbo@redhat.com>
# A slides file should populate the variable slides with
# a list of tuples. Each tuple should have:
#
# - Slide content
# - User data
# - Canvas width
# - Canvas height
#
# Slide content can be a string, a list of strings,
# a function returning one of those, or a generator
# yielding strings. The user data should be a dictionary or
# None, and is both used to communicate options to the
# renderer and to pass extra options to the theme functions.
#
# A function-based slide content will be passed a renderer object.
# Renderer is an object similar to a cairo.Context and
# pangocairo.CairoContext but has its own methods too.
# The more useful of them here are put_text, put_image, and
# set_allocation. See their pydocs.
slides = []
def slide_add(f, data=None, width=800, height=600):
#slides[:0] = [(f, data, width, height)]
slides.append ((f, data, width, height))
return f
import pango, pangocairo, cairo, os, signal
# We use slide data to tell the theme who's speaking.
# That is, which side the bubble should point to.
behdad = -1
whois = None
def who(name):
global whois
whois = name
# And convenience functions to add a slide. Can be
# used as a function decorator, or called directly.
def slide_who(f, who, data=None):
if data:
data = dict (data)
else:
data = {}
data['who'] = who
return slide_add (f, data)
def slide(f, data=None):
return slide_who (f, whois, data=data)
def slide_noone(f, data=None):
return slide_who (f, None, data=data)
def slide_behdad(f, data=None):
return slide_who (f, behdad, data=data)
#
# Slides start here
#
@slide_noone
def title_slide (r):
r.move_to (400, 100)
r.put_text (
"""Generating perfectly\ntext-extractable\nPDFs""", width=800, valign=1)
r.move_to (0, 450)
r.put_text ("""Behdad Esfahbod\n<span font_desc="16">besfahbo@<span foreground="#c00">redhat</span>.com</span>""",
desc="20", halign=1, valign=-1)
who (behdad)
def list_slide (l, data=None):
def s (r):
return '\n'.join (l)
#yield l[0]
#for i in l[1:]:
# yield '\n'+i
s.__name__ = l[0]
slide (s, data)
slide_noone("Terminology 101")
list_slide ([
"• Character",
"• Glyph",
"• <i>Codepoint</i>",
], data={'align': pango.ALIGN_LEFT})
list_slide ([
"• Input text",
"• Output glyph",
"• <i>Encoded codepoints</i>",
], data={'align': pango.ALIGN_LEFT})
list_slide ([
'• "Hello"',
"• 〈1,2,3,3,4〉",
"• 〈0102030304〉",
], data={'align': pango.ALIGN_LEFT})
slide_noone('<span size="larger">Overview of\nPDF Fonts</span>\n<span size="smaller">(subsetted,\ncustom encoding)</span>')
list_slide ([
'• Simple',
'• Composite',
], data={'align': pango.ALIGN_LEFT})
slide ('<b>Simple Fonts</b>\neach byte of\nencoded codepoints\nrepresents one codepoint')
list_slide ([
'<b>Simple Fonts</b>',
'• Type1',
'• TrueType',
'• Type3',
], data={'align': pango.ALIGN_LEFT})
slide ('<b>Composite Fonts</b>\ncomplex mapping of\nencoded codepoints\nto codepoints')
list_slide ([
'<b>Composite Fonts</b>',
'<b>(CID-keyed Fonts)</b>',
'• Type0 (CFF)',
'• Type2 (TrueType)',
], data={'align': pango.ALIGN_LEFT})
slide ('A CID is exactly\nwhat we called\na codepoint')
slide ('<b>Observation</b>\nAll PDF fonts can\nhave arbitrary\ncodepoint-to-glyph\nmapping')
slide ('<b>ToUnicode</b>\nAll PDF fonts can\nhave a codepoint-to-\nsequence-of-characters\nmapping')
slide ('ToUnicode mapping\nis the only\nstandard way for\ntext extraction')
slide ("<b>Broken Practice</b>\nUse original font's\ncharacter-to-glyph\nmapping to generate\nToUnicode mapping")
slide ('Assumes\nreversible\nchar-to-glyph\nmapping')
slide_noone('What\nTo Do?')
slide('Admit that\nchar-to-glyph\nmapping is\n not 1-to-1')
slide('<b>Corollary</b>\noutput glyphs\nare not enough\nto produce\ngood PDF text')
slide_noone("Terminology 201")
list_slide ([
"• Cluster",
], data={'align': pango.ALIGN_LEFT})
@slide_noone
def cluster_image (r):
r.move_to (400, 300)
r.put_image ("Fi_garamond_sort_001.png", height=650)
r.set_allocation (200, 0, 400, 600)
yield ""
list_slide ([
"<b>Cluster</b>",
"• Grapheme cluster: ö",
"• Ligature cluster:",
' - Typographical: 〈<span font_desc="Doulos SIL">f</span>,<span font_desc="Doulos SIL">i</span>〉 → 〈<span font_desc="Doulos SIL">fi</span>〉',
' - Orthographical: 〈<span font_desc="Nazli">ﻟ</span>,<span font_desc="Nazli">ﺎ</span>〉 → 〈<span font_desc="Nazli">ﻻ</span>〉',
], data={'align': pango.ALIGN_LEFT})
list_slide ([
"<b>Cluster</b>",
"• M input character",
"• N output glyphs",
"• M→N cluster"
], data={'align': pango.ALIGN_LEFT})
list_slide ([
"• Cluster mapping",
], data={'align': pango.ALIGN_LEFT})
slide ('Any good\nPDF generator\nshould take\ninput text,\noutput glyphs,\n<i>and</i>\ncluster mapping')
slide_noone("What To Do\nWith Them?")
slide ("M→1")
slide ("M→0")
slide ("M→N")
slide_noone("Issues")
slide ("0→1")
slide ("How to\nbreak?")
slide ("<i>Cursor</i>\npositions")
slide ("Run\norder")
slide ("Bidi")
slide_noone ('<span font_desc="Doulos SIL">~fin~</span>')
if __name__ == "__main__":
import slippy
import textextraction_theme
slippy.main (slides, textextraction_theme)
|
