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
|
"""NetCDF file reader.
This is adapted from Roberto De Almeida's Pupynere PUre PYthon NEtcdf REader.
classes changed to underscore_separated instead of CamelCase
TODO:
Add write capability.
"""
#__author__ = "Roberto De Almeida <rob@pydap.org>"
__all__ = ['netcdf_file', 'netcdf_variable']
import struct
import itertools
import mmap
from numpy import ndarray, zeros, array
ABSENT = '\x00' * 8
ZERO = '\x00' * 4
NC_BYTE = '\x00\x00\x00\x01'
NC_CHAR = '\x00\x00\x00\x02'
NC_SHORT = '\x00\x00\x00\x03'
NC_INT = '\x00\x00\x00\x04'
NC_FLOAT = '\x00\x00\x00\x05'
NC_DOUBLE = '\x00\x00\x00\x06'
NC_DIMENSION = '\x00\x00\x00\n'
NC_VARIABLE = '\x00\x00\x00\x0b'
NC_ATTRIBUTE = '\x00\x00\x00\x0c'
class netcdf_file(object):
"""A NetCDF file parser."""
def __init__(self, file, mode):
mode += 'b'
self._buffer = open(file, mode)
if mode in ['rb', 'r+b']:
self._parse()
elif mode == 'ab':
raise NotImplementedError
def flush(self):
pass
def sync(self):
pass
def close(self):
pass
def create_dimension(self, name, length):
pass
def create_variable(self, name, type, dimensions):
pass
def read(self, size=-1):
"""Alias for reading the file buffer."""
return self._buffer.read(size)
def _parse(self):
"""Initial parsing of the header."""
# Check magic bytes.
assert self.read(3) == 'CDF'
# Read version byte.
byte = self.read(1)
self.version_byte = struct.unpack('>b', byte)[0]
# Read header info.
self._numrecs()
self._dim_array()
self._gatt_array()
self._var_array()
def _numrecs(self):
"""Read number of records."""
self._nrecs = self._unpack_int()
def _dim_array(self):
"""Read a dict with dimensions names and sizes."""
assert self.read(4) in [ZERO, NC_DIMENSION]
count = self._unpack_int()
self.dimensions = {}
self._dims = []
for dim in range(count):
name = self._read_string()
length = self._unpack_int()
if length == 0: length = None # record dimension
self.dimensions[name] = length
self._dims.append(name) # preserve dim order
def _gatt_array(self):
"""Read global attributes."""
self.attributes = self._att_array()
# Update __dict__ for compatibility with S.IO.N
self.__dict__.update(self.attributes)
def _att_array(self):
"""Read a dict with attributes."""
assert self.read(4) in [ZERO, NC_ATTRIBUTE]
count = self._unpack_int()
# Read attributes.
attributes = {}
for attribute in range(count):
name = self._read_string()
nc_type = self._unpack_int()
n = self._unpack_int()
# Read value for attributes.
attributes[name] = self._read_values(n, nc_type)
return attributes
def _var_array(self):
"""Read all variables."""
assert self.read(4) in [ZERO, NC_VARIABLE]
# Read size of each record, in bytes.
self._read_recsize()
# Read variables.
self.variables = {}
count = self._unpack_int()
for variable in range(count):
name = self._read_string()
self.variables[name] = self._read_var()
def _read_recsize(self):
"""Read all variables and compute record bytes."""
pos = self._buffer.tell()
recsize = 0
count = self._unpack_int()
for variable in range(count):
name = self._read_string()
n = self._unpack_int()
isrec = False
for i in range(n):
dimid = self._unpack_int()
name = self._dims[dimid]
dim = self.dimensions[name]
if dim is None and i == 0:
isrec = True
attributes = self._att_array()
nc_type = self._unpack_int()
vsize = self._unpack_int()
begin = [self._unpack_int, self._unpack_int64][self.version_byte-1]()
if isrec: recsize += vsize
self._recsize = recsize
self._buffer.seek(pos)
def _read_var(self):
dimensions = []
shape = []
n = self._unpack_int()
isrec = False
for i in range(n):
dimid = self._unpack_int()
name = self._dims[dimid]
dimensions.append(name)
dim = self.dimensions[name]
if dim is None and i == 0:
dim = self._nrecs
isrec = True
shape.append(dim)
dimensions = tuple(dimensions)
shape = tuple(shape)
attributes = self._att_array()
nc_type = self._unpack_int()
vsize = self._unpack_int()
# Read offset.
begin = [self._unpack_int, self._unpack_int64][self.version_byte-1]()
return netcdf_variable(self._buffer.fileno(), nc_type, vsize, begin, shape, dimensions, attributes, isrec, self._recsize)
def _read_values(self, n, nc_type):
bytes = [1, 1, 2, 4, 4, 8]
typecodes = ['b', 'c', 'h', 'i', 'f', 'd']
count = n * bytes[nc_type-1]
values = self.read(count)
padding = self.read((4 - (count % 4)) % 4)
typecode = typecodes[nc_type-1]
if nc_type != 2: # not char
values = struct.unpack('>%s' % (typecode * n), values)
values = array(values, dtype=typecode)
else:
# Remove EOL terminator.
if values.endswith('\x00'): values = values[:-1]
return values
def _unpack_int(self):
return struct.unpack('>i', self.read(4))[0]
_unpack_int32 = _unpack_int
def _unpack_int64(self):
return struct.unpack('>q', self.read(8))[0]
def _read_string(self):
count = struct.unpack('>i', self.read(4))[0]
s = self.read(count)
# Remove EOL terminator.
if s.endswith('\x00'): s = s[:-1]
padding = self.read((4 - (count % 4)) % 4)
return s
def close(self):
self._buffer.close()
class netcdf_variable(object):
def __init__(self, fileno, nc_type, vsize, begin, shape, dimensions, attributes, isrec=False, recsize=0):
self._nc_type = nc_type
self._vsize = vsize
self._begin = begin
self.shape = shape
self.dimensions = dimensions
self.attributes = attributes # for ``dap.plugins.netcdf``
self.__dict__.update(attributes)
self._is_record = isrec
# Number of bytes and type.
self._bytes = [1, 1, 2, 4, 4, 8][self._nc_type-1]
type_ = ['i', 'S', 'i', 'i', 'f', 'f'][self._nc_type-1]
dtype = '>%s%d' % (type_, self._bytes)
bytes = self._begin + self._vsize
if isrec:
# Record variables are not stored contiguosly on disk, so we
# need to create a separate array for each record.
#
# TEO: This will copy data from the newly-created array
# into the __array_data__ region, thus removing any benefit of using
# a memory-mapped file. You might as well just read the data
# in directly.
self.__array_data__ = zeros(shape, dtype)
bytes += (shape[0] - 1) * recsize
for n in range(shape[0]):
offset = self._begin + (n * recsize)
mm = mmap.mmap(fileno, bytes, access=mmap.ACCESS_READ)
self.__array_data__[n] = ndarray.__new__(ndarray, shape[1:], dtype=dtype, buffer=mm, offset=offset, order=0)
else:
# Create buffer and data.
mm = mmap.mmap(fileno, bytes, access=mmap.ACCESS_READ)
self.__array_data__ = ndarray.__new__(ndarray, shape, dtype=dtype, buffer=mm, offset=self._begin, order=0)
# N-D array interface
self.__array_interface__ = {'shape' : shape,
'typestr': dtype,
'data' : self.__array_data__,
'version': 3,
}
def __getitem__(self, index):
return self.__array_data__.__getitem__(index)
def getValue(self):
"""For scalars."""
return self.__array_data__.item()
def assignValue(self, value):
"""For scalars."""
self.__array_data__.itemset(value)
def typecode(self):
return ['b', 'c', 'h', 'i', 'f', 'd'][self._nc_type-1]
def _test():
import doctest
doctest.testmod()
|
