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
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
|
#cython: language_level=3
import scipy.sparse as sp
from libc.stdio cimport fopen, fclose, fgetc, ungetc, EOF, FILE
from libc.string cimport strchr
cdef enum State:
BEGIN_LINE, READ_START_ATOM, READ, QUOTED, END_QUOTED, END_ATOM,
COMMENT, CARRIAGE_RETURNED, ESCAPE, EQUALS,
SET_VALUE, WAIT_VALUE, READ_VALUE, READ_DECS, TO_NEXT
cdef enum ColKinds:
ATTRIBUTE, CLASS, META
def sparse_read_float(fname):
cdef:
State state
char c
int ci
char *not_in_atom = "#,|;\n\r\x00"
int col, line, cur_line
int in_line
char atom[10240]
char *atomp
char *atome = atom + 10240
char *endc
char f_eof = 0
int attr_index
float value, decs
char col_kind
list X_data = []
list X_rows = []
list X_cols = []
list Y_data = []
list Y_rows = []
list Y_cols = []
list M_data = []
list M_rows = []
list M_cols = []
dict attr_indices = {}
dict class_indices = {}
dict meta_indices = {}
cdef FILE *f = fopen(fname, "rb")
if f == NULL:
raise IOError("File '{}' cannot be opened".format(fname))
line = 0
cur_line = 0
in_line = 0
try:
state = BEGIN_LINE
while not (f_eof and state == BEGIN_LINE):
if state == BEGIN_LINE:
col_kind = ATTRIBUTE
if in_line or line == 0:
line += 1
t_names = attr_indices
cur_line += 1
col = in_line = 0
state = READ_START_ATOM
if state != SET_VALUE:
ci = fgetc(f)
if ci == EOF:
f_eof = 1
c = b"\x00"
else:
c = <char>ci
col += 1
if state == READ_START_ATOM:
atomp = atom
value = 1
if c == b"," or c == b" " or c == b"\t":
continue
elif c == b'"':
state = QUOTED
continue
elif c == b"=":
raise ValueError("{}:{}:{}: missing value name"
.format(fname, cur_line, col))
# fall through
state = READ
if state == ESCAPE:
if c == b"t": c = b"\t"
elif c == b"n": c = b"\n"
elif c == b"r": c = b"\r"
elif c == b'"' or c == b"'" or c == b"\\" or c == b" ":
pass
elif c == b"\r" or c == b"\n":
raise ValueError("{}:{}:{}: end of line in escape sequence"
.format(fname, cur_line, col))
elif f_eof:
raise ValueError("{}:{}:{}: end of file in escape sequence"
.format(fname, cur_line, col))
else:
raise ValueError("{}:{}:{}: unrecognized escape sequence"
.format(fname, cur_line, col))
# fall through
state = READ
if state == READ:
if c == b"\\":
state = ESCAPE
continue
endc = strchr(not_in_atom, c)
if endc == NULL and c != b"=":
atomp[0] = c
atomp += 1
if atomp == atome:
raise ValueError("{}:{}:{}: value name too long"
.format(fname, cur_line, col))
continue
else:
# fall through to END_ATOM
state = END_ATOM
if state == QUOTED:
if c == b"\r" or c == b"\n":
raise ValueError(
"{}:{}:{}: end of line within a quoted value"
.format(fname, cur_line, col))
elif f_eof:
raise ValueError(
"{}:{}:{}: end of file within a quoted value"
.format(fname, cur_line, col))
elif c != b'"':
atomp[0] = c
atomp += 1
if atomp == atome:
raise ValueError("{}:{}:{}: value name too long"
.format(fname, cur_line, col))
else:
state = END_QUOTED
continue
if state == END_QUOTED:
if c == b" " or c == b"\t":
continue
endc = strchr(not_in_atom, c)
if endc == NULL and c != b"=":
raise ValueError("{}:{}:{}: quoted value should be "
"followed by value separator or end of line"
.format(fname, cur_line, col))
# fall through
state = END_ATOM
if state == END_ATOM:
while atomp != atom and (
atomp[-1] == b" " or atomp[-1] == b"\t"):
atomp -= 1
if atomp == atom:
if c == b"=":
raise ValueError("{}:{}:{}: empty value name"
.format(fname, cur_line, col))
else:
state = TO_NEXT
else:
atomp[0] = 0
b_atom = atom
"""
attr_index = t_names.get(b_atom, -1)
if attr_index < 0:
attr_index = t_names[b_atom] = len(t_names)
"""
attr_index = t_names.setdefault(b_atom,len(t_names))
atomp = atom
if c == b"=":
value = 0
state = WAIT_VALUE
continue
# fall through
state = SET_VALUE
if state == WAIT_VALUE:
if c == b" " or c == b"\t":
continue
else:
# fall through
state = READ_VALUE
if state == READ_VALUE:
if b"0" <= c <= b"9":
value = value * 10 + (c & 0xf)
elif c == b".":
decs = 0.1
state = READ_DECS
else:
endc = strchr(not_in_atom, c)
if endc != NULL:
state = SET_VALUE
else:
raise ValueError("{}:{}:{}: invalid value"
.format(fname, cur_line, col))
continue
if state == READ_DECS:
if b"0" <= c <= b"9":
value = value * decs + (c & 0xf)
decs /= 10
else:
endc = strchr(not_in_atom, c)
if endc != NULL:
state = SET_VALUE
else:
raise ValueError("{}:{}:{}: invalid value"
.format(fname, cur_line, col))
continue
if state == SET_VALUE:
if col_kind == ATTRIBUTE:
X_data.append(value)
X_rows.append(line - 1)
X_cols.append(attr_index)
elif col_kind == CLASS:
Y_data.append(value)
Y_rows.append(line - 1)
Y_cols.append(attr_index)
else:
M_data.append(value)
M_rows.append(line - 1)
M_cols.append(attr_index)
in_line += 1
state = TO_NEXT
if state == TO_NEXT:
if c == b"|":
if col_kind != ATTRIBUTE:
raise ValueError(
"{}:{}:{}: classes should follow attributes"
.format(fname, cur_line, col))
col_kind = CLASS
t_names = class_indices
state = READ_START_ATOM
elif c == b";":
if col_kind == META:
raise ValueError("{}:{}:{} duplicated semi-colons"
.format(fname, cur_line, col))
col_kind = META
t_names = meta_indices
state = READ_START_ATOM
elif c == b",":
state = READ_START_ATOM
elif c == b"#":
state = COMMENT
elif c == b"\\":
state = ESCAPE
elif c == b"\n" or f_eof:
state = BEGIN_LINE
elif c == b"\r":
state = CARRIAGE_RETURNED
else:
state = READ_START_ATOM
continue
elif state == COMMENT:
if c == b"\n" or f_eof:
state = BEGIN_LINE
elif c == b"\r":
state = CARRIAGE_RETURNED
elif state == CARRIAGE_RETURNED:
if c != b"\n" and not f_eof:
ungetc(c, f)
state = BEGIN_LINE
finally:
fclose(f)
if in_line == 0:
line -= 1
res = []
for t_data, t_rows, t_cols, ll in (
(X_data, X_rows, X_cols, attr_indices),
(Y_data, Y_rows, Y_cols, class_indices),
(M_data, M_rows, M_cols, meta_indices)):
if len(t_data):
mat = sp.coo_matrix((t_data, (t_rows, t_cols)), (line, len(ll))).tocsr()
mat.sort_indices()
else:
mat = None
res.append(mat)
return tuple(res) + (attr_indices, class_indices, meta_indices)
# TODO: this function needs to be bottlenecked (not literally;
# adding 'char' and 'float32' should suffice)
|
