forked from extern/Unexpected-Keyboard
b61e9db21e
Make compose sequences ending in the same character to share the ending state. This reduce the compiled compose key data size from 33kB to 27kB.
278 lines
9.6 KiB
Python
278 lines
9.6 KiB
Python
import textwrap, sys, re, string, json, os
|
|
from array import array
|
|
|
|
# Compile compose sequences from Xorg's format or from JSON files into an
|
|
# efficient state machine.
|
|
# See [ComposeKey.java] for the interpreter.
|
|
#
|
|
# Takes input files as arguments and generate a Java file.
|
|
# The initial state for each input is generated as a constant named after the
|
|
# input file.
|
|
|
|
# Parse symbol names from keysymdef.h. Many compose sequences in
|
|
# en_US_UTF_8_Compose.pre reference theses. For example, all the sequences on
|
|
# the Greek, Cyrillic and Hebrew scripts need these symbols.
|
|
def parse_keysymdef_h(fname):
|
|
with open(fname, "r") as inp:
|
|
keysym_re = re.compile(r'^#define XK_(\S+)\s+\S+\s*/\*.U\+([0-9a-fA-F]+)\s')
|
|
for line in inp:
|
|
m = re.match(keysym_re, line)
|
|
if m != None:
|
|
yield (m.group(1), chr(int(m.group(2), 16)))
|
|
|
|
dropped_sequences = 0
|
|
|
|
# Parse XKB's Compose.pre files
|
|
def parse_sequences_file_xkb(fname, xkb_char_extra_names):
|
|
# 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(sc):
|
|
uchar, named_char = sc
|
|
if uchar != "":
|
|
c = chr(int(uchar, 16))
|
|
elif len(named_char) == 1:
|
|
c = named_char
|
|
else:
|
|
if not named_char in char_names:
|
|
raise Exception("Unknown char: " + named_char)
|
|
c = char_names[named_char]
|
|
# The state machine can't represent sequence characters that do not fit
|
|
# in a 16-bit char.
|
|
if len(c) > 1 or ord(c[0]) > 65535:
|
|
raise Exception("Char out of range: " + r)
|
|
return c
|
|
# 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]
|
|
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
|
|
|
|
# Basic support for comments in json files. Reads a file
|
|
def strip_cstyle_comments(inp):
|
|
def strip_line(line):
|
|
i = line.find("//")
|
|
return line[:i] + "\n" if i >= 0 else line
|
|
return "".join(map(strip_line, inp))
|
|
|
|
# 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.loads(strip_cstyle_comments(inp))
|
|
return list(seqs.items())
|
|
|
|
# Format of the sequences file is determined by its extension
|
|
def parse_sequences_file(fname, xkb_char_extra_names={}):
|
|
if fname.endswith(".pre"):
|
|
return parse_sequences_file_xkb(fname, xkb_char_extra_names)
|
|
if fname.endswith(".json"):
|
|
return parse_sequences_file_json(fname)
|
|
raise Exception(fname + ": Unsupported format")
|
|
|
|
# A sequence directory can contain several sequence files as well as
|
|
# 'keysymdef.h'.
|
|
def parse_sequences_dir(dname):
|
|
compose_files = []
|
|
xkb_char_extra_names = {}
|
|
# Parse keysymdef.h first if present
|
|
for fbasename in os.listdir(dname):
|
|
fname = os.path.join(dname, fbasename)
|
|
if fbasename == "keysymdef.h":
|
|
xkb_char_extra_names = dict(parse_keysymdef_h(fname))
|
|
else:
|
|
compose_files.append(fname)
|
|
sequences = []
|
|
for fname in compose_files:
|
|
sequences.extend(parse_sequences_file(fname, xkb_char_extra_names))
|
|
return sequences
|
|
|
|
# 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(tries):
|
|
previous_leafs = {} # Deduplicate leafs
|
|
states = []
|
|
def add_tree(t):
|
|
this_node_index = len(states)
|
|
# 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()):
|
|
states[i] = (c, add_node(t[c]))
|
|
i += 1
|
|
return this_node_index
|
|
def add_leaf(c):
|
|
if c in previous_leafs:
|
|
return previous_leafs[c]
|
|
this_node_index = len(states)
|
|
previous_leafs[c] = this_node_index
|
|
# There are two encoding for leafs: character final state for 15-bit
|
|
# characters and string final state for the rest.
|
|
if len(c) > 1 or ord(c[0]) > 32767: # String final state
|
|
javachars = array('H', c.encode("UTF-16-LE"))
|
|
states.append((-1, len(javachars) + 1))
|
|
for c in javachars:
|
|
states.append((c, 0))
|
|
else: # Character final state
|
|
states.append((c, 1))
|
|
return this_node_index
|
|
def add_node(n):
|
|
if type(n) == str:
|
|
return add_leaf(n)
|
|
else:
|
|
return add_tree(n)
|
|
states.append((1, 1)) # Add an empty state at the beginning.
|
|
entry_states = { n: add_tree(root) for n, root in tries.items() }
|
|
return entry_states, states
|
|
|
|
# Debug
|
|
def print_automata(automata):
|
|
i = 0
|
|
for (s, e) in automata:
|
|
s = "%#06x" % s if isinstance(s, int) else '"%s"' % str(s)
|
|
print("%3d %8s %d" % (i, s, e), file=sys.stderr)
|
|
i += 1
|
|
|
|
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(entry_states, machine):
|
|
chars_map = {
|
|
# These characters cannot be used in unicode form as Java's parser
|
|
# unescape unicode sequences before parsing.
|
|
-1: "\\uFFFF",
|
|
"\"": "\\\"",
|
|
"\\": "\\\\",
|
|
"\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)))
|
|
def gen_entry_state(s):
|
|
name, state = s
|
|
return " public static final int %s = %d;" % (name, state)
|
|
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();
|
|
|
|
%s
|
|
}""" % (
|
|
# 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)),
|
|
"\n".join(map(gen_entry_state, entry_states.items())),
|
|
))
|
|
|
|
total_sequences = 0
|
|
tries = {} # Orderred dict
|
|
for fname in sorted(sys.argv[1:]):
|
|
tname, _ = os.path.splitext(os.path.basename(fname))
|
|
if os.path.isdir(fname):
|
|
sequences = parse_sequences_dir(fname)
|
|
else:
|
|
sequences = parse_sequences_file(fname)
|
|
add_sequences_to_trie(sequences, tries.setdefault(tname, {}))
|
|
total_sequences += len(sequences)
|
|
entry_states, automata = make_automata(tries)
|
|
gen_java(entry_states, automata)
|
|
print("Compiled %d sequences into %d states. Dropped %d sequences." % (total_sequences, len(automata), dropped_sequences), file=sys.stderr)
|
|
# print_automata(automata)
|