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
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
|
import textwrap, sys, re, string, json
# Names not defined in Compose.pre
xkb_char_extra_names = {
"space": " ",
"minus": "-",
"asterisk": "*",
"colon": ":",
"equal": "=",
"exclam": "!",
"grave": "`",
"parenleft": "(",
"parenright": ")",
"percent": "%",
"period": ".",
"plus": "+",
"question": "?",
"semicolon": ";",
"underscore": "_",
}
dropped_sequences = 0
# Parse XKB's Compose.pre files
def parse_sequences_file_xkb(fname):
# Parse a line of the form:
# <Multi_key> <minus> <space> : "~" asciitilde # TILDE
# Sequences not starting with <Multi_key> are ignored.
line_re = re.compile(r'^((?:\s*<[^>]+>)+)\s*:\s*"((?:[^"\\]+|\\.)+)"\s*(\S+)?\s*(?:#.+)?$')
char_re = re.compile(r'\s*<(?:U([a-fA-F0-9]{4,6})|([^>]+))>')
def parse_seq_line(line):
global dropped_sequences
prefix = "<Multi_key>"
if not line.startswith(prefix):
return None
m = re.match(line_re, line[len(prefix):])
if m == None:
return None
def_ = m.group(1)
try:
def_ = parse_seq_chars(def_)
result = parse_seq_result(m.group(2))
except Exception as e:
# print(str(e) + ". Sequence dropped: " + line.strip(), file=sys.stderr)
dropped_sequences += 1
return None
return def_, result
char_names = { **xkb_char_extra_names }
# Interpret character names of the form "U0000" or using [char_names].
def parse_seq_char(c):
uchar, named_char = c
if uchar != "":
return chr(int(uchar, 16))
# else is a named char
if len(named_char) == 1:
return named_char
if not named_char in char_names:
raise Exception("Unknown char: " + named_char)
return char_names[named_char]
# Interpret the left hand side of a sequence.
def parse_seq_chars(def_):
return list(map(parse_seq_char, re.findall(char_re, def_)))
# Interpret the result of a sequence, as outputed by [line_re].
def parse_seq_result(r):
if len(r) == 2 and r[0] == '\\':
return r[1]
# The state machine can't represent characters that do not fit in a
# 16-bit char. This breaks some sequences that output letters with
# combined diacritics or emojis.
if len(r) > 1 or ord(r[0]) > 65535:
raise Exception("Char out of range: " + r)
return r
# Populate [char_names] with the information present in the file.
with open(fname, "r") as inp:
for line in inp:
m = re.match(line_re, line)
if m == None or m.group(3) == None:
continue
try:
char_names[m.group(3)] = parse_seq_result(m.group(2))
except Exception:
pass
# Parse the sequences
with open(fname, "r") as inp:
seqs = []
for line in inp:
s = parse_seq_line(line)
if s != None:
seqs.append(s)
return seqs
# Parse from a json file containing a dictionary sequence → result string.
def parse_sequences_file_json(fname):
with open(fname, "r") as inp:
seqs = json.load(inp)
return list(seqs.items())
# Format of the sequences file is determined by its extension
def parse_sequences_file(fname):
if fname.endswith(".pre"):
return parse_sequences_file_xkb(fname)
if fname.endswith(".json"):
return parse_sequences_file_json(fname)
raise Exception(fname + ": Unsupported format")
# Turn a list of sequences into a trie.
def add_sequences_to_trie(seqs, trie):
def add_seq_to_trie(t_, seq, result):
t_ = trie
i = 0
while i < len(seq) - 1:
c = seq[i]
if c not in t_:
t_[c] = {}
if isinstance(t_[c], str):
global dropped_sequences
dropped_sequences += 1
print("Sequence collide: '%s = %s' '%s = %s'" % (
seq[:i+1], t_[c], seq, result),
file=sys.stderr)
return
t_ = t_[c]
i += 1
c = seq[i]
t_[c] = result
for seq, result in seqs:
add_seq_to_trie(trie, seq, result)
# Compile the trie into a state machine.
def make_automata(tree_root):
states = []
def add_tree(t):
# Index and size of the new node
i = len(states)
s = len(t.keys())
# Add node header
states.append(("\0", s + 1))
i += 1
# Reserve space for the current node in both arrays
for c in range(s):
states.append((None, None))
# Add nested nodes and fill the current node
for c in sorted(t.keys()):
node_i = len(states)
add_node(t[c])
states[i] = (c, node_i)
i += 1
def add_leaf(c):
states.append((c, 1))
def add_node(n):
if type(n) == str:
add_leaf(n)
else:
add_tree(n)
add_tree(tree_root)
return states
def batched(ar, n):
i = 0
while i + n < len(ar):
yield ar[i:i+n]
i += n
if i < len(ar):
yield ar[i:]
# Print the state machine compiled by make_automata into java code that can be
# used by [ComposeKeyData.java].
def gen_java(machine):
chars_map = {
# These characters cannot be used in unicode form as Java's parser
# unescape unicode sequences before parsing.
"\"": "\\\"",
"\\": "\\\\",
"\n": "\\n",
"\r": "\\r",
ord("\""): "\\\"",
ord("\\"): "\\\\",
ord("\n"): "\\n",
ord("\r"): "\\r",
}
def char_repr(c):
if c in chars_map:
return chars_map[c]
if type(c) == int: # The edges array contains ints
return "\\u%04x" % c
if c in string.printable:
return c
return "\\u%04x" % ord(c)
def gen_array(array):
chars = list(map(char_repr, array))
return "\" +\n \"".join(map(lambda b: "".join(b), batched(chars, 72)))
print("""package juloo.keyboard2;
/** This file is generated, see [srcs/compose/compile.py]. */
public final class ComposeKeyData
{
public static final char[] states =
("%s").toCharArray();
public static final char[] edges =
("%s").toCharArray();
}""" % (
# Break the edges array every few characters using string concatenation.
gen_array(map(lambda s: s[0], machine)),
gen_array(map(lambda s: s[1], machine)),
))
total_sequences = 0
trie = {}
for fname in sys.argv[1:]:
sequences = parse_sequences_file(fname)
add_sequences_to_trie(sequences, trie)
total_sequences += len(sequences)
automata = make_automata(trie)
gen_java(automata)
print("Compiled %d sequences into %d states. Dropped %d sequences." % (total_sequences, len(automata), dropped_sequences), file=sys.stderr)
|
