Debian Codegraph @

Changelog

raw
1 version 1.7.4 (tag v1.7.4rel)
2 ================================
3 * Make sure automatic conversion of character arrays <--> string arrays works for Unicode strings (issue #1440).
4 (previously only worked correctly for encoding="ascii").
5 * Add netcdf plugins (blosc, zstd, bzip2) in wheels. Blosc plugin doesn't work in Windows wheels.
6 Macos wheels now use conda provided libs. (PR #1450)
7 * Add windows/arm (PR #1453) and free-threaded python wheels (issue #1454). Windows wheels now use netcdf-c 4.9.3.
8 WARNING: netcdf-c is not thread-safe and netcdf4-python does have internal locking so expect segfaults if you
9 use netcdf4-python on multiple threads with free-threaded python. Users must exercise care to only call netcdf from
10 a single thread.
11
12 version 1.7.3 (tag v1.7.3rel)
13 =============================
14 * Python 3.14 wheels (issue #1432)
15 * support os.PathLike arguments for `Dataset.fromcdl` and raise a `FileNotFoundError`
16 if the cdl is missing and a `FileExistsError` if the nc file already exists (PR #1387)
17 * raise more informative error when trying to iterate or
18 perform a membership operation on a Dataset (issue #1383)
19 * fix type hint for createEnumType (issue #1378)
20 * add python 3.13 to windows wheel builds (PR #1377)
21 * allow slicing of vlen and string variables with non-unitary strides (issue #1408).
22
23 version 1.7.2 (tag v1.7.2rel)
24 =============================
25 * add static type hints (PRs #1302, #1349)
26 * Expose nc_rc_set, nc_rc_get (via rc_set, rc_get module functions). (PR #1348)
27 * Add Variable.get_fill_value and allow `fill_value='default'` to
28 set `_FillValue` attribute using default fill values. (issue #1374, PR #1375).
29 * Fix NETCDF3 endian error (issue #1373, PR #1355).
30
31 version 1.7.1 (tag v1.7.1rel)
32 ===============================
33 * include nc_complex source code from v0.2.0 tag (instead of using submodule).
34 * add aarch64 wheels.
35
36 version 1.7.0 (tag v1.7.0rel)
37 ===============================
38 * add support for complex numbers via `auto_complex` keyword to `Dataset` (PR #1295)
39 * fix for deprecated Cython `DEF` and `IF` statements using compatibility header
40 with shims for unavailable functionality (PR #1277)
41 * use `szip` as the library name on Windows (PR #1304)
42 * add support for MS-MPI `MPI_Message` detection (PR #1305)
43 * fix for issue #1306 - surprising result when indexing vlen str with non-contiguous
44 indices.
45 * Fix bug in set_collective introduced in PR #1277 (collective mode was
46 always set).
47
48 version 1.6.5 (tag v1.6.5rel)
49 ===============================
50 * fix for issue #1271 (mask ignored if bool MA assigned to uint8 var)
51 * include information on specific object when reporting errors from netcdf-c
52 * python 3.12 wheels added, support for python 3.7 removed.
53
54 version 1.6.4 (tag v1.6.4rel)
55 ===============================
56 * set path to SSL certificates internally, so https DAP URLs work with wheels
57 (issue #1246, requires nc_rc_set function available starting with netcdf-c
58 4.9.1, plus bugfix in netcdf-c PR #2690).
59 Added certifi as a dependency.
60 * Added `isopen` method to `MFDataset` object to check if underlying files are open.
61
62 version 1.6.3 (tag v1.6.3rel)
63 ==============================
64 * Use ``nc_put_vars`` for strided writes for netcdf-c >= 4.6.2 (issue #1222).
65 * _Unsigned="false" should be same as not having _Unsigned set (issue #1232).
66 _Unsigned now must be set to "true" or "True" for variable to be interpreted
67 as unsigned, instead of just having _Unsigned be set (to anything).
68 * pypi wheels built with netcdf-c 4.9.1.
69
70 version 1.6.2 (tag v1.6.2rel)
71 ==============================
72 * Added ``netCDF4.__has_set_alignment__`` property to help identify if the
73 underlying netcdf4 supports setting the HDF5 alignment.
74 * Slicing multi-dimensional variables with an all False boolean index array
75 now returns an empty numpy array (instead of raising an exception - issue #1197).
76 Behavior now consistent with numpy slicing.
77 * fix problem with compiling using netcdf-c < 4.9.0 (issue #1209)
78 * pypi wheels build with netcdf-c 4.9.0.
79
80 version 1.6.1 (tag v1.6.1rel)
81 ==============================
82 * add Dataset methods has_<name>_filter (where <name>=zstd,blosc,bzip2,szip)
83 to check for availability of extra compression filters.
84 * release GIL for all C-lib calls (issue #1180).
85 * Add support for nc_set_alignment and nc_get_alignment to control alignment
86 of data within HDF5 files.
87
88 version 1.6.0 (tag v1.6.0rel)
89 ==============================
90 * add support for new quantization functionality in netcdf-c 4.9.0 via "signficant_digits"
91 and "quantize_mode" kwargs in Dataset.createVariable. Default quantization_mode is "BitGroom",
92 but alternate methods "BitRound" and GranularBitRound" also supported.
93 * opening a Dataset in append mode (mode = 'a' or 'r+') creates a Dataset
94 if one does not already exist (similar to python open builtin). Issue #1144.
95 Added a mode='x' option (as in python open) which is the same as mode='w' with
96 clobber=False.
97 * allow createVariable to accept either Dimension instances or Dimension
98 names in "dimensions" tuple kwarg (issue #1145).
99 * remove all vestiges of python 2 in _netCDF4.pyx and set cython language_level
100 directive to 3 in setup.py.
101 * add 'compression' kwarg to createVariable to enable new compression
102 functionality in netcdf-c 4.9.0. 'None','zlib','szip','zstd','bzip2'
103 'blosc_lz','blosc_lz4','blosc_lz4hc','blosc_zlib' and 'blosc_zstd'
104 are currently supported. 'blosc_shuffle',
105 'szip_mask' and 'szip_pixels_per_block' kwargs also added.
106 compression='zlib' is equivalent to (the now deprecated) zlib=True.
107 If the environment variable NETCDF_PLUGIN_DIR is set to point to the
108 directory with the compression plugin lib__nc* files, then the compression plugins will
109 be installed within the package and be automatically available (the binary
110 wheels have this). Otherwise, the environment variable HDF5_PLUGIN_PATH
111 needs to be set at runtime to point to plugins in order to use the new compression
112 options.
113 * MFDataset did not aggregate 'name' variable attribute (issue #1153).
114 * issue warning instead of raising an exception if missing_value or
115 _FillValue can't be cast to the variable type when creating a
116 masked array (issue #1152).
117 * Define MPI_Session for compatibility with current mpi4py (PR #1156).
118
119 version 1.5.8 (tag v1.5.8rel)
120 ==============================
121 * Fix Enum bug (issue #1128): the enum_dict member of an EnumType read from a file
122 contains invalid values when the enum is large enough (more than 127 or 255
123 members).
124 * Binary wheels for aarch64 and python 3.10.
125
126 version 1.5.7 (tag v1.5.7rel)
127 ==============================
128 * don't try to mask vlens with default _FillValue, since vlens don't have a default _FillValue.
129 This gets rid of numpy DeprecationWarning (issue #1099).
130 * update docs to reflect the fact that a variable must be in collective mode before writing
131 compressed data to it in parallel. Added a test for this (examples/mpi_example_compressed.py).
132 Issue #1108.
133 * Fix OverflowError when dimension sizes become greater than 2**32-1 elements on Windows (Issue #1112).
134 * Don't return masked arrays for vlens (only for primitive and enum types - issue #1115).
135
136 version 1.5.6 (tag v1.5.6rel)
137 ==============================
138 * move CI/CD tests from travis/appveyor to Github Actions (PR #1061).
139 * move netCDF4 dir under src so module can be imported in source directory (PR #1062).
140 * change numpy.bool to numpy.bool_ and numpy.float to numpy.float_ (float and
141 bool are deprecated in numpy 1.20, issue #1065)
142 * clean up docstrings so that they work with latest pdoc.
143 * update cython numpy API to remove deprecation warnings.
144 * Add "fromcdl" and "tocdl" Dataset methods for import/export of CDL
145 via ncdump/ncgen called externally via the subprocess module (issue #1078).
146 * remove python 2.7 support.
147 * broadcast data (if possible)to conform to variable shape when writing to a slice
148 (issue #1083).
149
150 version 1.5.5.1 (tag v1.5.5.1rel)
151 ==================================
152 * rebuild binary wheels for linux and OSX to link netcdf-c 4.7.4 and hdf5 1.12.0.
153
154 version 1.5.5 (tag v1.5.5rel)
155 ==============================
156 * have setup.py always try use nc-config first to find paths to netcdf and
157 hdf5 libraries and headers. Don't use pkg-config to find HDF5 if HDF5 env
158 vars are set (or read from setup.cfg).
159 * Change MIT license text to standard OSI wording (PR #1046).
160
161 version 1.5.4 (tag v1.5.4rel)
162 ==============================
163 * fix printing of variable objects for variables that end with the letter 'u'
164 (issue #983).
165 * make sure root group has 'name' attribute (issue #988).
166 * add the ability to pack vlen floats to integers using
167 scale_factor/add_offset (issue #1003)
168 * use len instead of deprecated numpy.alen (issue #1008)
169 * check size on valid_range instead of using len (issue #1013).
170 * add `set_chunk_cache/get_chunk_cache` module functions to reset the
171 default chunk cache sizes before opening a Dataset (issue #1018).
172 * replace use of numpy's deprecated tostring() method with tobytes()
173 (issue #1023).
174 * bump minimal numpy version to 1.9 (first version to have tobytes()).
175
176 version 1.5.3 (tag v1.5.3rel)
177 ==============================
178 * make sure arrays are masked that are not filled when auto_fill is off
179 (issue #972).
180 * python 3.8 binary wheels.
181
182 version 1.5.2 (tag v1.5.2rel)
183 ==============================
184 * fix for scaling bug when _Unsigned attribute is set and byteorder of data
185 does not match native byteorder (issue #930).
186 * revise documentation for Python 3 (issue #946).
187 * establish support for Python 2.7, 3.5, 3.6 and 3.7 (issue #948).
188 * use dict built-in instead of OrderedDict for Python 3.7+
189 (pull request #955).
190 * remove underline ANSI in Dataset string representation (pull request #956).
191 * remove newlines from string representation (pull request #960).
192 * fix for issue #957 (size of scalar var is a float since numpy.prod(())=1.0).
193 * make sure Variable.setncattr fails to set _FillValue (issue #959).
194 * fix detection of parallel HDF5 support with netcdf-c 4.6.1 (issue #964).
195
196 version 1.5.1.2 (tag v1.5.1.2rel)
197 ==================================
198 * fix another slicing bug introduced by the fix to issue #906 (issue #922).
199
200 version 1.5.1.1 (tag v1.5.1.1rel)
201 ==================================
202 * fixed __version__ attribute (was set incorrectly in 1.5.1 release).
203 * fix for issue #919 (assigning 2d array to 3d variable with singleton
204 first dimension with v[:] = a).
205 * minimum numpy changed from 1.9.0 to 1.10.0.
206
207 version 1.5.1 (tag v1.5.1rel)
208 ==============================
209 * fix issue #908 by adding workaround for incorrect value returned
210 by nc_inq_var_fill for netcdf-c < 4.5.1.
211 * fix bug writing slice to unlimited dimension that is not the first
212 (leftmost). Issue #906.
213 * make sure data gets converted to type of scale_factor when add_offset=0
214 and scale_factor=1 (issue #913).
215 * fix for reading empty (NIL) string attributes (issue #915).
216
217 version 1.5.0.1 (tag v1.5.0.1rel)
218 ==================================
219 * binary wheels for linux and macosx rebuilt against netcdf-c 4.6.3 (instead
220 of 4.4.1.1).
221 * add read-shared mode (mode='rs'). Significantly speeds up reads of NETCDF3
222 files (pull request #902).
223
224 version 1.5.0 (tag v1.5.0rel)
225 ===============================
226 * added support for parallel IO in the classic netcdf-3 formats through the
227 pnetcdf library (pull request #897).
228
229 version 1.4.3.2 (tag v1.4.3.2)
230 ===============================
231 * include missing membuf.pyx file in release source tarball.
232
233 version 1.4.3.1 (tag v1.4.3.1)
234 ===============================
235 * fix bug in implementation of NETCDF4_CLASSIC support for parallel IO
236 in v1.4.3 release.
237
238 version 1.4.3 (tag v1.4.3rel)
239 =============================
240 * make set_always_mask work in MFDataset.
241 * fix saving diskless files to disk with netcdf-c >= 4.6.2.
242 * write to an in-memory Dataset, memoryview buffer returned by Dataset.close()
243 (issue #865, requires netcdf-c >= 4.6.2)
244 * fix performance regression when using large sequences of consecutive
245 integers for indexing with netcdf-c >= 4.6.2 (issue #870).
246 * improved error messages for ncinfo and other utilities (issue #873).
247 * fix for int64 attributes not being created for NETCDF3_64BIT_DATA (CDF5)
248 files (issue #878).
249 * fix for MPI parallel error ("NetCDF: Attempt to use feature that was not
250 turned on when netCDF was built") using netcdf-c 4.6.2 (issue #883).
251 * Added methods `set_ncstring_attrs()` to Dataset, Group and Variable that
252 forces all text attributes to be written as variable length strings (netCDF
253 type NC_STRING - issue #882).
254 * Allow parallel mode with NETCDF4_CLASSIC files (issue #890).
255
256 version 1.4.2 (tag v1.4.2rel)
257 =============================
258 * add get_dims Variable method (issue #824)
259 * make sure format keyword not ignored when mode is 'ws' (issue #827)
260 * fix numpy FutureWarning (non-tuple sequence for
261 multidimensional indexing is deprecated), issue #833.
262 * add 'master_file' kwarg to MFDataset.__init__ (issue #835).
263 * always use nc_get_vars for strided access over OpenDAP (issue #838).
264 * raise FutureWarning when trying to set multi-dimensional array attribute
265 while still silently flattening the array (issue #841). Will change
266 to ValueError in next release (1.4.3).
267 * fix parallel writes when both nc4 parallel and pnetcdf parallel options
268 enabled in the netcdf-c library (issue #820).
269 * fix for writing masked scalar character variable (issue #850).
270
271 version 1.4.1 (tag v1.4.1rel)
272 =============================
273 * disable workaround for slow nc_get_vars for __netcdflibversion__ >= 4.6.2,
274 since a fix was added to speed up nc_get_vars in the C library. Issue 680.
275 * new Dataset and Variable methods (set_always_mask) to optionally
276 re-enable old behaviour (return masked arrays
277 only if selected slice contains missing values) (issue #809).
278
279 version 1.4.0 (tag v1.4.0rel)
280 =============================
281 * fixed bug in detection of CDF5 library support in setup.py (pull request
282 #736, issue #713).
283 * fixed reading of variables with zero-length dimensions in NETCDF3_CLASSIC
284 files (issue #743).
285 * allow integer-like objects in VLEN slices (not just python ints, issue
286 #526, pull request #757).
287 * treating _FillValue as a valid_min/valid_max was too surprising, despite
288 the fact the thet netcdf docs 'attribute best practices' suggests that
289 clients should to this. Revert this change from issue #576 (issue #761).
290 * remove netcdftime, since it is now a separate package. date2num, num2date
291 and date2index still importable from netCDF4.
292 * fix 'Unreachable code' cython warning (issue #767).
293 * Change behavior of string attributes so that nc.stringatt = ['foo','bar']
294 produces an vlen string array attribute in NETCDF4, instead of concatenating
295 into a single string ('foobar'). In NETCDF3/NETCDF4_CLASSIC, an IOError
296 is now raised, instead of writing 'foobar'. Issue #770.
297 * fix loading of enum type names (issue #775).
298 * make sure missing_value applies only to scaled short integers if
299 auto-scaling is on (issue #777).
300 * automatically create views of compound types with character arrays as
301 numpy strings (issue #773). Can be disabled using
302 'set_auto_chartostring(False)'. Numpy structured
303 array dtypes with 'SN' string subtypes can now be used to
304 define netcdf compound types (they get converted to ('S1',N)
305 character array types automatically).
306 * always return masked array by default, even if there are no
307 masked values (too surprising to get ndarray or MaskedArray depending
308 on slice, issue #785).
309 * treat valid_min/valid_max/_FillValue/missing_value as unsigned
310 integers if _Unsigned is set (to mimic behaviour of netcdf-java).
311 Conversion to unsigned type now occurs before masking and scale/offset
312 operation. Issue #794.
313
314
315 version 1.3.1 (tag v1.3.1rel)
316 =============================
317 * add parallel IO capabilities. netcdf-c and hdf5 must be compiled with MPI
318 support, and mpi4py must be installed. To open a file for parallel access,
319 use `parallel=True` in `Dataset.__init__` and optionally pass the mpi4py Comm instance
320 using the `comm` kwarg and the mpi4py Info instance using the `info` kwarg.
321 IO can be toggled between collective and independent using `Variable.set_collective`.
322 See `examples/mpi_example.py`. Issue #717, pull request #716.
323 Minimum cython dependency bumped from 0.19 to 0.21.
324 * Add optional `MFTime` calendar overload to use across all files, for example,
325 `'standard'` or `'gregorian'`. If `None` (the default), check that the calendar
326 attribute is present on each variable and values are unique across files raising
327 a `ValueError` otherwise.
328 * Allow _FillValue to be set for vlen string variables (issue #730).
329
330 version 1.3.0 (tag v1.3.0rel)
331 ==============================
332 * always search for HDF5 headers when building, even when nc-config is used
333 (since nc-config does not always include the path to the HDF5 headers).
334 Also use H5get_libversion to obtain HDF5 version info instead of
335 H5public.h. Fixes issue #677.
336 * encoding kwarg added to Dataset.__init__ and Dataset.filepath (default
337 is to use sys.getfilesystemencoding()) so that oddball
338 encodings (such as cp1252 on windows) can be handled in Dataset
339 filepaths (issue #686).
340 * Calls to nc_get_vars are avoided, since nc_get_vars is very slow (issue
341 #680). Strided slices are now converted to multiple calls to
342 nc_get_vara. This speeds up strided slice reads by a factor of 10-100
343 (especially for NETCDF4/HDF5 files) in most cases. In some cases, strided reads
344 using nc_get_vars are faster (e.g. strided reads over many dimensions
345 such as var[:,::2,::2,::2])), so a variable method use_nc_get_vars was added.
346 var.use_nc_get_vars(True) will tell the library to use nc_get_vars instead
347 of multiple calls to nc_get_vara, which was the default behaviour previous
348 to this change.
349 * fix utc offset time zone conversion in netcdftime - it was being done
350 exactly backwards (issue #685 - thanks to @pgamez and @mdecker).
351 * Fix error message for illegal ellipsis slicing, add test (issue #701).
352 * Improve timezone format parsing in netcdftime
353 (https://github.com/Unidata/netcdftime/issues/17).
354 * make sure numpy datatypes used to define CompoundTypes have
355 isalignedstruct flag set to True (issue #705), otherwise.
356 segfaults can occur. Fix required raising them minimum numpy requirement
357 from 1.7.0 to 1.9.0.
358 * ignore missing_value, _FillValue, valid_range, valid_min and valid_max
359 when creating masked arrays if attribute cannot be safely
360 cast to variable data type (and issue a warning). When setting
361 these attributes don't cast to variable dtype unless it can
362 be done safely and issue a warning. Issue #707.
363
364 version 1.2.9 (tag v1.2.9rel)
365 ==============================
366 * Fix for auto scaling and masking when _Unsigned attribute set (create
367 view as unsigned type after scaling and masking). Issue #671.
368 * Always mask values outside valid_min, valid_max (not just when
369 missing_value attribute present). Issue #672.
370 * Fix setup.py so pip install doesn't fail if cython not installed.
371 setuptools >= 18.0 now required for installation (Issue #666).
372
373 version 1.2.8 (tag v1.2.8rel)
374 ==============================
375 * recognize _Unsigned attribute used by netcdf-java to designate unsigned
376 integer data stored with a signed integer type in netcdf-3 (issue #656).
377 * add Dataset init memory parameter to allow loading a file from memory
378 (pull request #652, issues #406 and #295).
379 * fix for negative times in num2date (issue #659).
380 * fix for failing tests in numpy 1.13 due to changes in numpy.ma (issue #662).
381 * Checking for _Encoding attribute for NC_STRING variables, otherwise use
382 'utf-8'. 'utf-8' is used everywhere else, 'default_encoding' global module
383 variable is no longer used. getncattr method now takes optional kwarg
384 'encoding' (default 'utf-8') so encoding of attributes can be specified
385 if desired. If _Encoding is specified for an NC_CHAR ('S1') variable,
386 the chartostring utility function is used to convert the array of
387 characters to an array of strings with one less dimension (the last
388 dimension is interpreted as the length of each string) when reading the
389 data. When writing the data, stringtochar is used to convert a numpy
390 array of fixed length strings to an array of characters with one more
391 dimension. chartostring and stringtochar now also have an 'encoding' kwarg.
392 Automatic conversion to/from character to string arrays can be turned off
393 via a new set_auto_chartostring Dataset and Variable method (default
394 is True). Addresses issue #654.
395 * Cython >= 0.19 now required, _netCDF4.c and _netcdftime.c removed from
396 repository.
397
398 version 1.2.7 (tag v1.2.7rel)
399 ==============================
400 * fix for issue #624 (error in conversion to masked array when variable slice
401 returns a scalar). This is a regression introduced in 1.2.5 associated
402 with support for vector missing_values. Test (tst_masked5.py) added for
403 vector missing_values.
404 * fix for python 3.6 compatibility (error retrieving character _FillValue attribute,
405 issue #626). Test with python 3.6 using travis CI.
406
407 version 1.2.6 (tag v1.2.6rel)
408 ==============================
409 * fix some test failures on big endian PPC64 that were due to
410 errors in byte-swapping logic. Also fixed bug in enum
411 code exposed on PPC64 (issue #608).
412 * remove support for python 2.6 (it probably still will work for a while
413 though).
414 * Sometimes checking that data being assigned to a variable has
415 an 'ndim' attribute is not sufficient, instead check to see that
416 the object supports the buffer interface (issue #613).
417 * make get_variables_by_attributes work in MFDataset (issue #610)
418 The hack is also applied for set_auto_maskandscale, set_auto_scale,
419 set_automask, so these don't have to be duplicated in MFDataset (pull
420 request #571).
421
422 version 1.2.5 (tag v1.2.5rel)
423 ==============================
424 * Add MFDataset.set_auto_maskandscale (plus set_auto_scale, set_auto_mask).
425 Fixes issue #570.
426 * Use valid_min/valid_max/valid_range attributes when defining mask
427 (issue #576). Values outside the valid range are considered to
428 be missing when defining the mask.
429 * Fix for issue #584 (add support for dates before -4712-1-1 in 360_day
430 and 365_day calendars to netcdftime.utime).
431 * Fix for issue #593: add support for datetime.timedelta operations
432 (adding and subtracting timedelta, subtracting two datetime
433 instances to compute time duration between them), implement
434 datetime.replace() and datetime.__str__(). datetime.__repr__()
435 includes the full state of an instance. Add datetime.calendar.
436 datetime comparison operators have full accuracy now.
437 * Fix for issue #585 by increasing the size of the buffer used to store the
438 filepath.
439 * Fix for issue #592: Add support for string array attributes. (When
440 reading, a vlen string array attribute is returned as a list of
441 strings. To write, use var.setncattr_string("name", ["two", "strings"]).)
442 * Fix for issue #596 - julian day calculations wrong for negative years,
443 caused incorrect roundtrip num2date(date2num(date)) roundtrip for dates with year
444 < 0.
445 * Make sure negative years work in utime.num2date (issue #596).
446 * raise NotImplementedError when trying to pickle Dataset, Variable,
447 CompoundType, VLType, EnumType and MFDataset (issue #602).
448 * Fix for issue #527: initialize vldata[i].p in Variable._get(...).
449
450 version 1.2.4 (tag v1.2.4rel)
451 ==============================
452 * Fix for issue #554. It is now ensured that data is in native endian
453 byte order before passing to netcdf-c library. Data read from variable
454 with non-native byte order is also byte-swapped, so that dtype remains
455 consistent with netcdf variable. Behavior now consistent with h5py.
456 * raise warning for HDF5 1.10.x (issue #549), since backwards
457 incompatible files may be created.
458 * raise AttributeError instead of RuntimeError when attribute operation
459 fails. raise IOError instead of RuntimeError when nc_create or
460 nc_open fails (issue #546).
461 * Use NamedTemporaryFile instead of deprecated mktemp in tests
462 (pull request #543).
463 * add AppVeyor automated windows tests (pull request #540).
464
465 version 1.2.3.1 (tag v1.2.3.1rel)
466 ==================================
467 * fix bug in setup.py (pull request #539, introduced in issue #518).
468
469 version 1.2.3 (tag v1.2.3rel)
470 ==============================
471 * try to avoid writing NC_STRING attributes if possible, by
472 trying to convert unicode strings to ascii and write as NC_CHAR (issue
473 #529). This preserves compatibility with clients (like Matlab) that
474 can't deal with NC_STRING attributes. A 'setncattr_string' method
475 was added for Dataset and Variable to that users can force attributes
476 to be written as NC_STRING if necessary.
477 * fix failing tests with numpy 1.11 (issues #521 and #522).
478 * fix indentation bug in nc4tonc3 utility (issue #519).
479 * add the capability in setup.py to use pkg-config instead of
480 nc-config (pull request #518).
481 * make sure slices which return scalar masked arrays
482 are consistent with numpy.ma (issue #515).
483 * add test/tst_cdf5.py and test/tst_filepath.py (to test new
484 NETCDF3_64BIT_DATA format and filepath Dataset method).
485 * expose netcdftime.__version__ (issue #504).
486 * fix potential memory leak in Dataset.filepath in attempt to fix
487 mysterious segfaults on CentOS6 (issue #506). Segfaults
488 can apparently still occur on systems like CentOS6 with old versions of glibc.
489
490 version 1.2.2 (tag v1.2.2rel)
491 =============================
492 * fix failing tests on python 2.6 (issue #497). Change minimum required
493 python from 2.5 to 2.6.
494 * Potential memory leaks fixed by freeing string pointers internally allocated
495 in netcdf-c using nc_free_string. Also use nc_free_vlens to free space allocated
496 for vlens inside netcdf-c (issue #495).
497 * invoke str on filename argument to Dataset constructor, so pathlib
498 instances can be used (issue #489).
499 * don't use hardwired NC_MAX_DIMS or NC_MAX_VARS internally to allocate space
500 for dimension or variable ids. Instead, find out the number of dims
501 and vars and use malloc. NC_MAX_NAME is still used to allocate space
502 for attribute and variable names, since there is no obvious way to
503 determine the length of these names.
504 * if trying to write a unicode attribute, check to see if it exists
505 first and is NC_CHAR, and if so, delete it and recreate it. Workaround for C
506 lib bug discovered in issue #485.
507 * support for NETCDF3_64BIT_DATA format supported in netcdf-c 4.4.0.
508 Similar to NETCDF3_64BIT (now NETCDF3_64BIT_OFFSET), but includes
509 64 bit dimensions and sizes, plus unsigned and 64 bit integer
510 data types.
511 * make sure chunksize does not exceed dimension size
512 (for non-unlimited dimensions) on variable creation (issue #480).
513 * add 'size' attribute to Dimension (same as len(d), where
514 d is a Dimension instance, issue #477).
515 * fix bug in nc3tonc4 with --unpackshort=1 (issue #474).
516 * dates do not have to be contiguous, i.e. can be before and after the
517 missing dates in Gregorian calendar (pull request #476).
518
519 version 1.2.1 (tag v1.2.1rel)
520 =============================
521 * add the capability to slice variables with unsorted integer sequences,
522 or integer sequences with duplicates (issue #467). This was done
523 by converting boolean array slices to integer array slices internally,
524 instead of the other way around.
525 * raise TypeError if masked array assigned to a VLEN str variable slice
526 (issue #464).
527 * Ellipsis now can be used with scalar VLEN str variables (issue #458).
528 Slicing of scalar VLEN (non-str) variables now works.
529 * Allow non-positive reference years in non-real-world calendars
530 (issue #442).
531
532 version 1.2.0 (tag v1.2.0rel)
533 =============================
534 * Fixes to setup.py for building on windows (issue #460).
535 * warnings now issued if file being read contains unsupported
536 variables or data types (they were previously being silently
537 skipped).
538 * added 'get_variables_by_attributes' method (issue #454).
539 * check for 'units' attribute in date2index (issue #453).
540 * added support for enum types (issue #452).
541 * added 'isopen' Dataset method (issue #450).
542 * raise ValueError if year 0 or negative year used in time units string.
543 The year 0 does not exist in the Julian and Gregorian
544 calendars (issue #442).
545
546 version 1.1.9 (tag v1.1.9rel)
547 =============================
548 * fix for issue #391 (data is already byte-swapped to native
549 endian format by the HDF4 library).
550 * fix for issue #415 (copy.deepcopy does not work on
551 netcdftime datetime object).
552 * fix for issue #420 - len(v) where v is a scalar variable returned
553 unexpected IndexError, now returns "TypeError: len() on unsized object"
554 (same as numpy does for len() on a scalar array).
555 * translate docstrings from epydoc markup to markdown, so
556 pdoc can be used (epydoc is dead).
557 * add small offset in conversion to Julian date for numerical
558 stability (more accurate round trip calculations). This offset is removed
559 in back conversion only from microseconds. Pull request #433.
560 * add detection of unsigned integers to handling of automatic
561 packing (set_auto_scale and set_auto_maskandscale) when writing.
562 Pull request #435.
563 * use USE_SETUPCFG env var to over-ride use of setup.cfg. If USE_SETUPCFG
564 evaluates to false, setup.cfg will not be used and all configuration
565 variables can be set from environment variables. Useful when using 'pip
566 install' and nc-config is broken (issue #438).
567 * fix for integer overflow in date2index (issue #444).
568
569 version 1.1.8 (tag v1.1.8rel)
570 =============================
571 * v[...] now returns a numpy scalar array (not just a scalar) when
572 v is a scalar netcdf variable (issue #413).
573 * unix-like paths can now be used in createVariable and createGroup.
574 v = nc.createVariable('/path/to/var1',('xdim','ydim'),float)
575 will create a Variable named 'var1', while also creating the Groups
576 'path' and 'path/to' if they do not already exist.
577 Similarly, g = nc.createGroup('/path/to') acts like 'mkdir -p' in unix, creating
578 the Groups 'path' and '/path/to', if they don't already exist.
579 Users who relied on nc.createGroup(groupname) failing when the
580 group already exists will have to modify their code, since nc.createGroup
581 will now return the existing group instance.
582 Dataset.__getitem__ also added. nc['/path/to'] returns
583 a Group instance, and nc['/path/to/var1'] returns a Variable
584 instance.
585 * change minimum required numpy to 1.7.0, fix so all tests pass with 1.7.0.
586 Added travis tests for minimum required cython, numpy (issue #404).
587 * enable abbreviations to time units specification, as allowed in CF (issue
588 #402). Now, instead of just 'seconds' and 'seconds', 'secs', 'sec' and 's'
589 are also allowed (similar to minutes, days and hours).
590 * install utility scripts in utils directory with setuptools entry points
591 (pull request #392 from @mindw). Code for utilities moved
592 to netCDF4_utils.py - makes utilities more windows-friendly.
593 * make sure booleans are treated correctly in setup.cfg. Add
594 use_cython (default True) to setup.cfg. If set to False, then
595 cython will not be used to compile netCDF4.pyx (existing netCDF4.c
596 will be used instead).
597 * use "from Cython.Build import cythonize" instead of
598 "from Cython.Distutils import build_ext" in setup.py (issue #393)
599 to conform to new cython build mechanism (CEP 201, described at
600 https://github.com/cython/cython/wiki/enhancements-distutils_preprocessing).
601 * unicode attributes now written as strings, not bytes (using
602 nc_put_att_string instead of nc_put_att_text, issue #388).
603 * add __orthogonal_indexing__ attribute to Variable, Dataset and Group (issue #385) to
604 denote that Variable objects do not follow numpy indexing semantics for integer and
605 boolean array indices.
606 * make sure application of scale_factor and add_offset works correctly when
607 scale_factor not given (issue #381).
608 * add man pages for nc3tonc4, nc4tonc3, ncinfo in man directory.
609 Not installed by setup.py (contributed by Ross Gammon, issue #383).
610 * replace tabs with spaces by running reindent.py on all *.py and *.pyx files
611 (issue #378).
612 * refactor netCDF4_utils and netCDF4 module into netCDF4 package.
613 Refactoring effectively removes netCDF4 utils private attributes from
614 netCDF4 namespace, so has the potential to break code using private
615 attributes (issue #409).
616
617 version 1.1.7 (tag v1.1.7rel)
618 =============================
619 * check to make sure cython >= 0.19 is available before trying
620 to use it (otherwise compilation with fail). Issue 367.
621 * add ipython notebooks from Unidata workshop in examples directory.
622 * fix ellipsis variable slicing regression (issue 371).
623 * release the Global Interpreter Lock (GIL) when calling the C
624 library for read operations. Speeds up multi-threaded reads
625 (issue 369). Caution - the HDF5 library may need to be compiled
626 with the threadsafe option to ensure that global data structures
627 are not corrupted by simultaneous manipulation by different threads.
628 * Make sure USE_NCCONFIG environment variable takes precedence over value
629 of use_ncconfig in setup.cfg. With this change, 'pip install netCDF4'
630 with USE_NCCONFIG=1 will use environment variables to find paths to
631 libraries and include files, instead of relying on nc-config (issue #341).
632
633 version 1.1.6 (tag v1.1.6rel)
634 =============================
635
636 * fix for issue 353 (num2date can no longer handle units like 'hours since
637 2000-01-01 0').
638 * fix for issue 354 (num2date no longer supports multi-dimensional arrays).
639 * fix for spurious UserWarning about endian-ness mismatch (issue 364).
640 * make calendar name keyword for num2date/date2num case insensitive
641 (issue 362).
642 * make sure units parser returns time-zone naive datetime instance that
643 includes UTC offset (issue 357). UTC offset was applied incorrectly in
644 netcdftime.date2num and num2date. No longer need to depend on
645 python-dateutil.
646
647 version 1.1.5 (tag v1.1.5rel)
648 =============================
649
650 * add dependency on python-dateutil in setup.py and install docs.
651 * use python datetime in num2date and date2num whenever possible. Remove
652 duplicate num2date and date2num functions from netcdftime. Addresses issue
653 #344. Add microsecond capability to netcdftime.datetime. Roundtrip
654 accuracy of num2date/date2num now down to less than a millisecond.
655 * use nc-config by default to find dependencies. setup.py modified to handle failure
656 to find nc-config more gracefully (issue #340). If you wish to use env vars to point
657 to the libs, you must first move the setup.cfg file out of the way
658 (rename it to setup.cfg.save), or set USE_NCCONFIG to 0.
659 * if endian-ness of variable is specified, adjust datatype to reflect this when opening
660 a file (issue 346).
661 * fix for issue #349 (seconds outside the range 0-59 in netcdftime.num2date).
662
663 version 1.1.4 (tag v1.1.4rel)
664 =============================
665
666 * speedup conversion of array indices to slices (issue #325).
667 * fix for issue #330 (incorrect values for seconds returned by netcdftime).
668 * fix reading of scalar vlen variables (issue #333).
669 * setting fill_value=False in createVariable for vlen and compound variables
670 now does nothing, instead of causing an error when the Dataset is closed
671 (issue #331).
672 * cython will regenerate netCDF4.c when install is run, not just build.
673 Makes 'pip install' do the right thing when cython is installed (issue #263).
674
675 version 1.1.3 (tag v1.1.3rel)
676 =============================
677
678 * checked in _datetime.c to git (resolves issue #315). Note - _datetime.c
679 was *not* included in the 1.1.2 release.
680 * Changed __str__ to __repr__ in MFDataset, to be consistent with Dataset
681 (issue #317). IPython uses __repr__ to make use-friendly human-readable summaries
682 of objects in the terminal.
683
684 version 1.1.2 (tag v1.1.2rel)
685 =============================
686
687 * fix for issue 312 (allow slicing with objects that can be cast to ints).
688 * indexing netCDF variables with integer sequences and boolean arrays now
689 behave the same way (integer sequences are converted to boolean arrays
690 internally). Addresses issue #300. Since indexing using integer sequences
691 does not behave exactly as before, some client code may break. For example,
692 previously when integer index arrays had the same length, and that length
693 was equal to the number of dimensions of the array being indexed,
694 netcdf4-python mirrored the numpy indexing behavior and treated the elements
695 of the index arrays as individual sets of integer indices. This special
696 case has been removed. An IndexError is now raised when the new behavior
697 would produce a different result than the old, i.e. when the
698 indices in an integer sequence are not sorted, or there are duplicate
699 indices in the sequence.
700 * fix for issue #310 (masked arrays not returned correctly when variable
701 has non native endian-ness).
702 * fix for issue #306 (slicing variable with "-1" when there is only
703 one element along that dimension).
704 * Improved speed of num2date and date2num for standard, julian, gregorian
705 and proleptic gregorian calendars by vectorizing the functions. See Issue #296
706 * Fix for issue #301 ("Datestring parser chokes on years with extra space").
707 * Add name property for Dimension, Variable and Group instances (to access string
708 name associated with instance).
709 * Allow for null byte attributes (so _FillValue='\x00' can be set manually).
710 Issue 273.
711 * Added __repr__ (matching __str__) for all types (pull request #291).
712 IPython uses __repr__ to make use-friendly human-readable summaries
713 of objects in the terminal.
714
715
716 version 1.1.1 (tag v1.1.1rel)
717 ==============================
718
719 * make sure _FillValue is a byte for character arrays in Python 3 (issue
720 271).
721 * add numpy to install_requires in setup.py (issue #282, fixes issue #211).
722 'pip install netcdf4-python' will no longer fail if numpy not installed.
723 * Fix for issue 278 (UnicodeDecodeError reading netcdf.h from setup.py with
724 Python 3.4).
725 * Make netcdftime.datetime immutable and hashable (issue 255).
726 * Fix issue with slicing of scalar VLEN arrays (issue 270).
727 * Add set_auto_mask and set_auto_scale methods to control auto scaling and
728 auto masking separately. (issue 269). Also added set_auto_maskandscale,
729 set_auto_scale, set_auto_mask Dataset/Group methods that recursively walk
730 through all variables in the Dataset/Group.
731 * Make sure file_format attribute propagated to Group instances (issue 265).
732 * Fix for issue #259 ("Cannot use broadcasting to set all elements of a
733 Variable to a given value").
734
735 version 1.1.0 (tag v1.1.0rel)
736 =============================
737
738 * revert weakref change, so that previous behaviour (Dimensions and Variables
739 keep strong references to parent Dataset) is the default. New keyword
740 argument 'keepweakref' for Dataset.__init__ can be set to true to get
741 weak references.
742
743 version 1.0.9 (tag v1.0.9rel)
744 =============================
745
746 * speed up the creation of new Group instances (issue 239).
747
748 * fix logic errors in setup.py (issue 236).
749
750 * it is now possible to create and set variable length string variables with
751 numpy string datatypes (pull request 224).
752
753 * add .travis.yml (for travis-ci testing on github), silence warnings from
754 test output (issue 225).
755
756 * modify __unicode__ for Variable and Dimension to return more useful error
757 message when Dataset object has been garbage collected.
758
759 * use weak references to group instances when creating Dimension and Variable
760 objects. This prevents cyclic references messing up garbage collection (issue
761 218, pull request 219).
762
763 * accessing values from a 0-dimensional Variable now returns a 0-dimensional
764 numpy array, not a 1-dimensional array (issue 220). To write code
765 compatible with both the old and new (fixed) behavior, wrap values accessed
766 from a 0-dimensional Variable with numpy.asscalar.
767
768 * add an __array__ method to Variable to make numpy ufuncs faster (issue 216).
769
770 * change download_url in setup.py to point to pypi instead of googlecode.
771
772 * fix for date2index error when time variable has only one entry (issue 215).
773
774 * silence warnings ("Non-trivial type declarators in shared declaration (e.g.
775 mix of pointers and values). Each pointer declaration should be on its own
776 line") with Cython 0.2.
777
778 * reduced memory usage for Variable.__getitem__ under Python 2.
779
780 version 1.0.8 (tag v1.0.8rel)
781 =============================
782
783 * change file_format Dataset attribute to data_model (keeping file_format
784 for backward compatibility). Add disk_format attribute (underlying
785 disk format, one of NETCDF3, HDF4, HDF5, DAP2, DAP4, PNETCDF or UNDEFINED).
786 Uses nc_inq_format_extended, added in version 4.3.1 of library. If using
787 earlier version of lib, disk_format will be set to UNDEFINED.
788
789 * default _FillValue now ignored for byte data types (int8 and uint8) as per
790 http://www.unidata.ucar.edu/software/netcdf/docs/netcdf-c/Fill-Values.html#Fill-Values
791 "If you need a fill value for a byte variable, it is recommended that you
792 explicitly define an appropriate _FillValue attribute, as generic utilities
793 such as ncdump will not assume a default fill value for byte variables".
794 ncinfo now returns fill mode information (issue 209).
795
796 * check to see if filling was disabled before masking data equal to
797 default fill value (issue 209).
798
799 * add variable type information to Dataset.__repr__ (output of ncinfo).
800
801 version 1.0.7 (tag v1.0.7rel)
802 =============================
803
804 * add the ability to specify the locations of hdf4,jpeg and curl libs, in
805 case netCDF4 was built statically with HDF4 and/or OpenDAP support (issue
806 207).
807
808 * add 'ncinfo' utility (like 'ncdump -h' but less verbose).
809
810 * more information displayed when Dataset or Group instance is printed.
811
812 * fix for issue 194 (versions after 1.0.5 fail for netcdf 4.1.1, due to
813 call to nc_inq_path, which was added in netcdf 4.1.2). Fixed by adding
814 compile time API check similar to what was done for nc_rename_grp. If
815 filepath Dataset method is called an exception will be raised at runtime
816 if the module was built with netcdf < 4.1.2, or cython was not installed
817 at build time.
818
819 * fix for issues 202 and 206 (exception raised by numpy.isnan for character data types).
820
821 * if dateutils not installed and time unit accuracy < 1 second requested, have
822 netcdftime raise an ImportError.
823
824 version 1.0.6 (svn revision 1312)
825 ================================
826
827 * issue warning of endian-ness of dtype argument does not match endian kwarg in
828 createVariable.
829
830 * make sure netcdf type NC_CHAR always returned in numpy array dtype 'S1'
831 (sometimes arrays of type 'U1' were being returned). Fixes intermittently failing
832 test tst_compoundatt.py on python 3.3.
833
834 * fix for issue 201 (if data associated with numpy array not the
835 same endian-ness as dtype, data was written incorrectly). Now
836 bytes are swapped if necessary. Variable.endian() now returns
837 'native' instead of None for NETCDF3 formatted files. createVariable
838 now enforces endian='native' for NETCDF3 files. Added tst_endian.py
839 test case.
840
841 * fix for issue 200 (library version detection failed on cygwin).
842
843 * fix for issue 199 (nc4tonc3 utility not copying global attributes).
844
845 * fix for issue 198 (setup.py chokes when no arguments given).
846
847 * fix for issue 197 (slicing of netCDF variables using lists of integers).
848
849 * create 'path' attribute for group instance using posixpath, instead
850 of os.path (to ensure the unix path is used on all platforms). Issue 196.
851
852 * fix for issue 196 (test failures on win32 due to files being deleted
853 before they are closed).
854
855 version 1.0.5 (svn revision 1278)
856 ================================
857
858 * change setup.py to compile the Cython sources directly, if
859 cython is available.
860 This allows for "ifdef" like capability to modify source at compile time to
861 account for changes in netcdf API (e.g. the forthcoming addition of the
862 nc_rename_grp in version 4.3.1).
863
864 * added a "renameGroup" method, which raises an exception if
865 the netcdf lib version linked does not support it. Requires
866 netcdf >= 4.3.1.
867
868 * support for more than one missing value (missing_value attribute
869 is a vector) when converting to masked array.
870
871 * add 'renameAttribute' method to Dataset, Group and Variable.
872
873 * fix so that var[:] = x works if x is a scalar, and var is a netcdf variable
874 with an unlimited dimension that has shape () - i.e. no data has been
875 written to it yet. Before this change, var[:] = x did not write any data.
876 Now the scalar x will be written as the first entry in var along the
877 unlimited dimension.
878
879 * remove dos line feeds from nc3tonc4 (issue 181).
880
881 * add datatype property for Variable that returns numpy dtype for primitive
882 datatypes (same as dtype attribute) but returns CompoundType or VLType
883 instance for compound or vlen variables (issue 178).
884
885 * fix logic for deciding where to look for nc-config in setup.py (issue 177).
886
887 * issue a warning and don't try to apply scale_factor or add_offset if these
888 attributes are not convertible to floats (issue 176).
889
890 * add filepath method to Dataset instance to return file path (or opendap
891 URL) used to create Dataset (issue 172).
892
893 * fix for issue 170 (opening a remote DAP dataset fails after creating
894 a NETCDF4 formatted file).
895
896 * fix for issue 169 (error in chartostring function on 64-bit windows).
897
898 * add support for missing_value or _FillValue == NaN (issue 168).
899
900 * added a Dimension.group() method (issue 165).
901
902 version 1.0.4 (svn revision 1229)
903 =================================
904
905 * fixed alignment bug that could cause memory corruption
906 when reading compound type variables. All users of compound
907 types should upgrade.
908
909 version 1.0.3 (svn revision 1219)
910 =================================
911
912 * don't try to write empty data array to netcdf file (fixed failing
913 test with netcdf 4.3.0rc2).
914
915 * date2num, num2date and date2index now can handle units of microseconds and milliseconds
916 (for proleptic_gregorian calendar, or gregorian and standard calendars
917 as long as the time origin is after 1582-10-15). Issue 159.
918
919 * Added a _grp attribute to Dimension (issue 165).
920
921 * don't bundle ordereddict (issue 164).
922
923 * support reading of vlen string attributes (issue 156).
924
925 * add --vars option to nc3tonc4 (issue 154).
926
927 * Don't try to set fletcher32 checksum on scalar variables (it causes HDF5
928 to crash). Fixes issue 150.
929
930 * Add --istart/--istop options to nc3tonc4 (issue 148, courtesy of
931 Rich Signell).
932
933 * fix for proleptic_gregorian in netcdftime.py (courtesy of Matthias Cuntz).
934
935 version 1.0.2 (svn revision 1196)
936 =================================
937
938 * disable version check for HDF5, which is broken by hdf5 1.8.10.
939
940 * make sure all files have a calendar attribute in MFTime (issue 144).
941
942 * more robust fix to issue 90 (array shape modified by assignment to
943 a netCDF variable with one more dimension), including test case.
944
945 version 1.0.1 (svn revision 1190)
946 =================================
947
948 * fix error that occurred when retrieving data from a variable that has a
949 missing_value attribute specified as a string (issue 142).
950
951 * automatically close netcdf files when there are no references
952 left to Dataset object (using __dealloc__ method). Fixes issue 137.
953
954 * fix for slicing of scalar vlen string variables (issue 140).
955
956 * fix to allow writing of unicode data to a NC_CHAR variable.
957
958 * allow for writing of large variables (> 2**32 elements). Fixes issue 130.
959
960 version 1.0fix1
961 ===============
962
963 * fix python 3 incompatibility in setup.py (issue 125).
964
965 version 1.0 (svn revision 1164)
966 ===============================
967 * add 'aggdim' keyword to MFDataset, so the name of the dimension
968 to aggregate over can be specified (instead of using the unlimited
969 dimension). aggdim=None by default, which results in the previous
970 behavior. aggdim must be the leftmost dimension of all the variables
971 to be aggregated.
972
973 * raise IndexError when indexing a netcdf variable out of range
974 so iterating over a variable in a for loop behaves as expected
975 (as described in http://effbot.org/zone/python-for-statement.htm).
976 Fixes issue 121.
977
978 * added MacPorts portfile (so it can be installed via MacPorts
979 on macosx using a "local Portfile repository"). Installs
980 from svn HEAD using 'port install netcdf4-python'.
981
982 * added experimental 'diskless' file capability (only added to the C
983 lib after the 4.2 release). Controlled by kwarg 'diskless' to
984 netCDF4.Dataset (default False). diskless=True when creating a file
985 results in a file that exists only in memory, closing the file
986 makes the data disappear, except if persist=True keyword given in
987 which case it is persisted to a disk file on close. diskless=True
988 when opening a file creates an in-memory copy of the file for faster access.
989
990 * add the ability to specify the location of the required libs
991 (and whether to use nc-config) with setup.cfg,
992 instead of using environment variables.
993
994 * fix ISO9601 date parser so it recognizes time zone offsets in time unit
995 strings (contributed by David Hassel, issue 114, r1117).
996
997 * add setncatts Dataset,Group and Variable method to add a bunch of
998 attributes (given in a python dictionary) at once. Speeds things
999 up for NETCDF3 and NETCDF4_CLASSIC files a lot, since nc_redef/nc_enddef
1000 not need to be called for each attribute (issue 85, r1113). Adding 1000
1001 attributes is about 35 times faster using setncatts to add them all at once.
1002 Makes no difference for NETCDF4 formatted files, since nc_redef/nc_enddef
1003 is not called.
1004
1005 * only round after apply scale_factor and add_offset if variable type is
1006 integer (issue 111, r1109).
1007
1008 * Fixed bug with all False Boolean index (r1107).
1009
1010 * added support for after, before and nearest selection method to date2index
1011 fast "first guess" indexing (r1106).
1012
1013 * Remove white space in time units string (netcdftime._parse_date).
1014 An extra space in the time units of one CMIP3 model caused an error
1015 (r1105).
1016
1017 * based on results with examples/bench_compress2.py, change
1018 default complevel for zlib compression from 6 to 4. If complevel=0,
1019 turn compression off entirely (set zlib=False) (r1102).
1020
1021 version 0.9.9 (svn revision 1099)
1022 ================================
1023 * changed default unicode encoding from "latin-1" to "utf-8", since
1024 this is the python 3 default, and the only encoding that appears to
1025 work for dimension and variable names.
1026
1027 * added test case for unicode attributes, variable and dimension names.
1028
1029 * fixes for unicode variable, dimension and group names.
1030
1031 * fix for unicode attributes in python3 (ncdump did not intrepret them as
1032 text strings). Issue 107.
1033
1034 * add --format option to nc4tonc3 utility (can be either NETCDF3_CLASSIC or
1035 NETCDF3_64BIT). Fixes issue 104.
1036
1037 version 0.9.8 (svn revision 1080)
1038 ================================
1039 * use numpy.ma.isMA to check for masked array (instead of checking
1040 for presence of 'mask' attribute).
1041
1042 * fixes for AIX with ibm xlc compiler.
1043
1044 * make sure unicode attributes don't get converted to ascii strings (issue
1045 98).
1046
1047 version 0.9.7 (svn revision 1073)
1048 ================================
1049 * Added __str__ methods to Dataset, Variable, Dimension, CompoundType,
1050 VLType and MFDataset, so useful human-readable information is provided when these
1051 objects are printed in an interactive session.
1052
1053 * don't try to apply scale_factor and offset if scale_factor=1 and
1054 add_offset=0 (to avoid making copies of large arrays).
1055
1056 * changed netCDF4._default_fillvals to netCDF4.default_fillvals (to make part
1057 of public API). Added to docs (issue 94).
1058
1059 version 0.9.6 (svn revision 1043)
1060 =================================
1061 * changed default unicode encoding from "ascii" to "latin-1" (iso-8859-1).
1062
1063 * add "unicode_error" module variable to control what happens when characters
1064 cannot be decoded by the encoding specified by the "default_encoding" module
1065 variable (which is "ascii" by default). unicode_error = "replace" by default
1066 which means bad characters are replace by "?". Previously an error was
1067 raised, the old behavior can be obtained by setting unicode_error = 'strict'.
1068 Fixes issue 92.
1069
1070 * add __enter__ and __exit__ methods so you can do "with Dataset(url) as f:"
1071 (issue 89).
1072
1073 * don't add extra singleton dimensions to rhs numpy arrays when assigning to
1074 a netcdf variable. Fixes issue 90.
1075
1076 * coerce missing_value attribute to same type as variable (for primitive
1077 types). Fixes issue 91.
1078
1079 version 0.9.5 (svn revision 1031)
1080 ================================
1081 * fix for compound variables on python 3.2.
1082
1083 * fix slicing of masked MFDataset variables (issue 83).
1084
1085 * round to nearest integer after packing with scale_factor and add_offset
1086 (instead of truncation) (issue 84).
1087
1088 * if add_offset missing, but scale_factor present, assume add_offset zero.
1089 if scale_factor missing, but add_offset present, assume scale_factor one.
1090 (this is consistent with unidata recommendations - issue 86).
1091
1092 * only try to convert strings to bytes for python 3 so Dataset can
1093 be subclassed (issue 87).
1094
1095 version 0.9.4 (svn revision 1018)
1096 ================================
1097 * tested with python 2.7.1/3.1.3 using netcdf 4.1.2 and hdf5 1.8.6.
1098
1099 * Added a 'default_encoding' module variable that controls how unicode
1100 strings are encoded into bytes. Default is 'ascii'.
1101
1102 * now works on Python 3.
1103
1104 * netCDF3 module removed. If you still need it, get it from netCDF4 0.9.3.
1105
1106 * regenerated C source with Cython 0.14.1.
1107
1108 * Added a MFTime class. Provide a unified interface to MFDataset time
1109 variable using different time units.
1110
1111 * Fixed bug in netcdftime (issue 75) that occurs when time specified is within
1112 one second of the end of the month.
1113
1114 * on unix-like systems, the environment variable USE_NCCONFIG can be set to
1115 tell setup.py to use the nc-config script installed by netcdf to figure out
1116 where all the libs and headers are (without having to specify NETCDF_DIR,
1117 HDF5_DIR, etc). Only works with netcdf 4.1.2.
1118
1119 version 0.9.3 (svn revision 930)
1120 ================================
1121 * fix chunk sizes bug (chunk sizes pointer should be size_t, not int). Fixes
1122 issue 66. Added test in tst_compression.py
1123
1124 * fixed writing of data with missing values with scale/offset packing. Added
1125 test (tst_masked2.py).
1126
1127 * fix iso8601 regex in netcdftime date parser so it can parse 'hours since 1-1-1 ...'
1128 (year had to be 4 digits previously)
1129
1130 version 0.9.2 (svn revision 907)
1131 ================================
1132 * fix netcdftime bug with '360_day' calendar. Fixes issue 59.
1133
1134 * make sure scalar slice of 1d variable returns array scalar (not array of
1135 shape (1,)). Fixes issue 57.
1136
1137 * updated date parser in netcdftime. Can now handle units like
1138 "seconds since 1970-01-01T00:00:00Z".
1139
1140 * added support in setup.py for specifying the locations of the HDF5/netcdf-4 headers
1141 and libs separately with environment variables (HDF5_INCDIR, HDF5_LIBDIR).i
1142 Patch contributed by Patrice Dumas.
1143
1144 * add masked array support to num2date (dates for missing times set to None).
1145
1146 * add chunk_cache keyword to createVariable. HDF5 default is 1mb, which can
1147 cause problems when creating 1000's of variables. In such cases,
1148 chunk_cache can be reduced, or set to zero.
1149
1150 * add set_var_chunk_cache and get_var_chunk_cache Variable methods.
1151
1152 * raise AttributeError when trying to set _FillValue attribute (it can only
1153 be reliably set on variable creation, using the fill_value keyword to
1154 createVariable).
1155
1156 version 0.9.1 (svn revision 879)
1157 ================================
1158 * raise ImportError if netcdf-4 < 4.1.1 or hdf5 <= 1.8.4.
1159
1160 * add __netcdf4libversion__ and __hdf5libversion__ module variables.
1161
1162 * make sure data is not truncated to integers before scale_factor and
1163 add_offset is applied (issue 46).
1164
1165 * fix bug in date2num with noleap calendar in netcdftime (issue 45).
1166
1167 * fix bug in 360day calendar in netcdftime (issue 44).
1168
1169 * python 2.4 compatibility restored (by modifying OrderedDict). Fixes issue
1170 37.
1171
1172 * make sure int64 attributes cast to int32 when format=NETCDF4_CLASSIC. This
1173 was causing tst_multifile.py to fail on 64-bit platforms.
1174
1175 * fix tutorial.py to cast 64 bit integers to 32 bit when writing to 32-bit
1176 integer vlen (was causing tutorial.py to fail on 64-bit platforms).
1177
1178 * remove nose dependency from tst_netcdftime.py.
1179
1180 version 0.9 (svn revision 846)
1181 ==============================
1182 * fixed bug (issue 30) with date2index occurring with dates outside the support.
1183
1184 * make sure that auto masking works with MFDataset.
1185
1186 * fix bug (issue 34) when slicing MFDataset variables with dimensions of
1187 length 1.
1188
1189 * used ordered dictionaries for variables, dimensions, groups etc to
1190 preserve creation order (makes it easier to copy files, fixes issue 28).
1191
1192 * change auto_maskandscale default to True. This means data will
1193 automatically be converted to and from masked arrays. Data scaled as short
1194 integers using the scale_factor and add_offset attributes will also be
1195 automatically converted to/from float arrays.
1196
1197 * add setncattr, getncattr, delncattr methods (for setting/getting/deleting netcdf
1198 attributes with names that clash with the reserved python attributes).
1199
1200 version 0.8.2 (svn revision 769)
1201 ================================
1202 * compound type tests re-enabled. Compound and vlen types now fully
1203 supported in netcdf-4.1-beta2.
1204
1205 * make sure data retrieved from a netCDF variable is not coerced to a python
1206 scalar (it should remain a numpy scalar array).
1207
1208 * fix docs to point out that an unlimited dimension can be created by setting
1209 size to *either* None or 0 in createDimension.
1210
1211 * fix another slicing corner case.
1212
1213 * remove auto pickling/unpickling into vlen strings (too cute, sometimes
1214 produced surprising results).
1215
1216 version 0.8.1 (svn revision 744)
1217 ================================
1218 * added 'cmptypes' and 'vltypes' Group/Dataset attributes, which contain
1219 dictionaries that map the names of compound and vlen types to
1220 CompoundType and VLType instances.
1221
1222 * Experimental variable-length (vlen) data type support added.
1223
1224 * changes to accommodate compound types in netcdf-4.1-beta snapshots.
1225 Compound types now work correctly for snapshots >= 20090603.
1226
1227 * Added __len__ method and 'size' property to Variable class.
1228
1229 * In date2index, replaced the brute force method by the bisection method and
1230 added a 'select' keyword to find the index of the date before, after or
1231 nearest the given date.
1232
1233 * Fixed bug occurring when indexing with a numpy array of length 1.
1234
1235 * Fixed bug that occurred when -1 was used as a variable index.
1236
1237 * enabled 'shared access' mode for NETCDF3 formatted files (mode='ws',
1238 'r+s' or 'as'). Writes in shared mode are unbuffered, which can
1239 improve performance for non-sequential access.
1240
1241 * fixed bug in renameVariable that caused failure when new name is longer
1242 than old name, and file format is NETCDF3_64BIT or NETCDF3_CLASSIC.
1243
1244 version 0.8 (svn revision 685)
1245 ==============================
1246 * added 'stringtoarr' utility function for converting python strings
1247 to numpy character arrays of a specified size.
1248
1249 * initial support for compound data types (which are mapped to structured
1250 numpy arrays). Compound data types are created with the createCompoundTYpe
1251 Dataset or Group method. Both attributes and variables can be compound
1252 types.
1253
1254 * make sure 64-bit integer attributes converted to 32 bits when writing to
1255 a NETCDF3 formatted file.
1256
1257 * added nc4tonc3 utility for converted NETCDF4_CLASSIC files to NETCDF3_64BIT
1258 files (useful for sharing data with colleagues that don't have netcdf-4
1259 capable clients).
1260
1261 version 0.7.7 (svn revision 626)
1262 ================================
1263
1264 * David Huard reworked fancy indexing - it is now much more efficient and
1265 less of a memory hog. Now works differently than numpy fancy indexing - 1d
1266 arrays of boolean or integer indices work independently on each dimension.
1267 This enables things like:
1268 >>> tempdat = temp[[0,1,3],lats>0,lons>0]
1269 (retrieves 1st, 2nd and 4th levels, all Northern Hem. and Eastern
1270 Hem. grid points - note that this would raise an IndexError in numpy)
1271
1272 * added opendap test (tst_dap.py).
1273
1274 * bugfix for nc3tonc4 utility.
1275
1276 * fix MFDataset.Variable. __getattr__ to raise AttributeError instead of
1277 KeyError when attribute not found.
1278
1279 * netcdftime version number upped to 0.7.
1280
1281 version 0.7.6 (svn revision 574)
1282 ================================
1283
1284 * added date2index function, courtesy of David Huard, which finds the indices
1285 in a netCDF time variable corresponding to a sequence of datetime instances.
1286
1287 * make _get_att/_set_att raise AttributeError instead of RuntimeError, so that
1288 getattr(object, 'nonexistantattribute', None) works. (thanks David Huard)
1289
1290 * v[:] = data now works along unlim dim, i.e. you can do this:
1291
1292 file = Dataset('test.nc', "w")
1293 file.createDimension("time", None) # unlimited dimension
1294 var = file.createVariable("var", 'd', ("time",))
1295 # you used to have to do this
1296 #var[0:10] = numpy.arange(10)
1297 # but now you can simply do this
1298 var[:] = numpy.arange(10)
1299
1300 version 0.7.5 (svn revision 549)
1301 ================================
1302 * return a scalar array, not a python scalar, when a slice returns a single
1303 number. This is more consistent with numpy behavior, and fixes a bug
1304 in MFDataset slicing.
1305 * added 'exclude' parameter to MFDataset.__init__
1306 * added set_auto_maskandscale method to MFDataset variables.
1307
1308 version 0.7.4 (svn revision 540)
1309 ================================
1310 * ensure all arithmetic is done with float64 in netcdftime (Rob Hetland).
1311 * fixes for netcdf-4.0-beta2 ('chunking' keyword to createVariable
1312 replaced by 'contiguous'). Now works with netcdf-4.0-beta2 and hdf5-1.8.0
1313 final, but is incompatible with netcdf-4.0-beta1.
1314
1315 version 0.7.3.1 (svn revision 507)
1316 ==================================
1317 * netCDF3 docs were missing from 0.7.3.
1318 * make sure quantization function preserves fill_value of masked arrays.
1319
1320 version 0.7.3 (svn revision 501)
1321 ================================
1322 * MFnetCDF4 module merged into netCDF4 and netCDF3 (now
1323 called MFDataset).
1324 * added netCDF3 module for those who can't install the netCDF 4 lib.
1325 * added set_auto_maskandscale Variable method to enable automatic
1326 packing and unpacking of short integers (using scale_factor
1327 and add_offset attributes) and automatic conversion to/from
1328 masked arrays (using missing_value or _FillValue attribute)
1329 on a per-variable basis.
1330 var.set_auto_maskandscale(True) turns automatic
1331 conversion on (it is off by default).
1332 * automatically pack/unpack short integer variables
1333 if scale_factor and add_offset variable attributes are set.
1334 * added support for masked arrays. If you try to write a masked
1335 array to a variable with the missing_value or _FillValue attributes
1336 set, the masked array is filled with that value before being written
1337 to the file. If you read data from a variable with the missing_value
1338 or _FillValue attribute set, a masked array is returned with the
1339 appropriate values masked.
1340 * added date2num and num2date functions.
1341 * added capability to use 'fancy indexing' with variable objects
1342 (i.e. using sequences of integers or booleans in slices). WARNING:
1343 if a sequence of integers or booleans is used to slice a netCDF4
1344 variable, all of the data in that dimension is read into a numpy
1345 array, and then the sequence is used to slice the numpy array,
1346 returning just the requested elements to the user. This can
1347 potentially gobble a lot of memory and degrade performance
1348 (especially if 'fancy indexing' is done on the left-most dimension).
1349 * added convenience functions stringtochar and chartostring for
1350 converting character arrays to arrays of fixed-length strings and
1351 vice-versa. Example usage in examples/test_stringarr.py.
1352
1353 20070826 - version 0.7.1 (svn revision 400)
1354 ===========================================
1355 * added 'endian()' and 'chunking()' Variable methods (to inquire about
1356 endian and chunking variable settings).
1357
1358 * 'ndim' attribute was not public (so it couldn't be accessed from python).
1359 Fixed.
1360
1361 * added 'endian' kwarg to createVariable (to set the endian-ness
1362 used in the HDF5 file).
1363
1364 * can now manually set HDF5 chunksizes for each dimension at
1365 variable creation, using 'chunksizes' kwarg to createVariable.
1366
1367 * added "getlibversion()" function to get info about version
1368 of netcdf-4 library used to build module.
1369
1370 * if a variable has an unsupported datatype (such as 'compound', or
1371 'vlen'), then instead of raising an exception, just skip it.
1372 Print a useful error message when an attribute with an unsupported
1373 datatype is accessed.
1374
1375 * if variable dimension is specified as 'dimname' or ('dimname')
1376 in createVariable, it is automatically converted to a tuple ('dimname',).
1377 Better error messages when specified dimension can't be found.
1378
1379 * createVariable accepts numpy dtype object as datatype. dtype variable
1380 attribute is now a numpy dtype object.
1381
1382 20070723 - version 0.7 (svn revision 361)
1383 =========================================
1384 * renamed MFnetCDF4_classic --> MFnetCDF4.
1385
1386 * eliminated netCDF4_classic module (all file formats handled by
1387 netCDF4 module now).
1388
1389 * removed all user-defined data type stuff (it was hacky and made
1390 the code too complex - wait till there is a real use case to
1391 refactor and put back in).
1392
1393 * added 'ndim' variable attribute (number of variable dimensions).
1394
1395 20070424 - version 0.6.3 (svn revision 302)
1396 ===========================================
1397 * passes all tests with netcdf-4.0-beta1/hdf5-1.8.0-beta1.
1398
1399 * if slice index is not a slice object, assume it's an integer (and
1400 try to convert to one if it is not). This allows numpy scalar arrays
1401 to work as slice indices.
1402
1403 * (netCDF4_classic only) try to make sure file is not left in 'define mode'
1404 when exception is raised.
1405
1406 * if slicing a variable results in a array with shape (1,), just return
1407 a scalar (except for compound types).
1408
1409 * added instructions for using the netCDF4_classic module to serve
1410 data over http with the DAP using pydap (http://pydap.org).
1411
1412 * added --quiet and --chunk options to nc3tonc4.
1413
1414 * Turned off zlib compression by default so as not to violate the
1415 'principle of least surprise'. Shuffle filter still activated
1416 by default when zlib compression turned on.
1417
1418 * Fixed bug in fletcher32 checksum activation call. Renamed compression()
1419 variable method to filters(), include fletcher32 checksum flag in output.
1420
1421 * added utility for converting GRIB1 files to compressed
1422 NETCDF4_CLASSIC files (requires PyNIO).
1423
1424 * added 'compression()' variable method that returns a dict with
1425 compression filter parameter settings for that variable. (rev 237)
1426
1427 * reimplemented 'shape' and 'dimensions' variable attributes as
1428 properties.
1429
1430 * fixed bug when 'chunking' keyword in createVariable was set to 'sub'
1431 (caused Bus Error on MacOS X).
1432
1433 * Setting 'shuffle=0' keyword in createVariable was turning off
1434 zlib compression filter instead of shuffle filter. Fixed.
1435
1436 20070213 - version 0.6.2
1437 ========================
1438 * updated for compatibility with netcdf-4.0-alpha18 and hdf5 1.8.0alpha5
1439 (shared dimensions actually work now).
1440
1441 * netCDF4.createVariable can now use old single character Numeric typecodes
1442 for datatype specification.
1443
1444 * Improvements to MFDataset (now called MFnetCDF4_classic) by Rob Hetland.
1445
1446 20061121 - version 0.6.1
1447 ========================
1448 * bugfixes for negative strides.
1449
1450 * bugfix for empty string attributes.
1451
1452 * support for shared dimensions (variables can use dimensions defined
1453 only in a parent group). This doesn't actually work yet, because of
1454 a bug in netcdf-4.0-alpha17.
1455
1456 * now requires Pyrex (C source files generated on the fly when setup.py
1457 is run).
1458
1459 20061003 - version 0.6
1460 ======================
1461 * if fill_value keyword to createVariable is set to the Boolean
1462 False (not an integer that evaluates to False), no pre-filling
1463 is done for that variable.
1464
1465 * updated to be compatible with netcdf-4.0-alpha17.
1466 Can now install pure-python netcdftime separately with setup-netcdftime.py.
1467 netcdftime will try to use numpy, but fall back to Numeric if numpy
1468 not installed.
1469
1470 * generated source files with a version of pyrex
1471 (from http://codespeak.net/svn/lxml/pyrex/) that produces
1472 extensions compatible with python 2.5.
1473
1474 * added new module for multi-file access of NETCDF3 and NETCDF4_CLASSIC
1475 files (MFDataset). Based on CDFMF from pycdf.
1476
1477 * implement negative strides in variable slicing
1478 (feature missing from Scientific.IO.NetCDF). Now variables support
1479 full python extended slicing syntax.
1480
1481 20060925 - version 0.5.1
1482 ========================
1483 * on 64-bit systems integer attributes in netCDF4_classic failed, since there
1484 is no 64-bit integer data type. Fixed by downcasting to 32-bit integer.
1485
1486 20060920 - version 0.5
1487 ======================
1488 * Compound type support! (members must be fixed data primitive types -
1489 no user-defined types or NC_STRING variables allowed). Attributes
1490 are still restricted to primitive data types (no vlen or compound
1491 type attributes).
1492
1493 * Assigning single values to a slice now does the Right Thing, i.e.
1494 >>> data[:] = 1
1495 fills all the elements with 1 (instead of raising an IndexError).
1496
1497 * Tested with numpy 1.0b5, netcdf-4.0-alpha16, HDF5 1.7.52 alpha.
1498
1499 * Added renameDimension and renameVariable methods to Dataset and Group classes.
1500
1501 * netCDF attributes can be deleted using python del (i.e. 'del dset.foo').
1502
1503 * Moved examples from test and test_classic to examples and
1504 examples_classic directories.
1505
1506 * Added proper unit tests (in test and test_classic directories).
1507
1508 * NULL characters are removed from text attributes.
1509
1510 * Variable _FillValue can be set using new keyword argument 'fill_value'
1511 to createVariable Dataset and Group method.
1512
1513 * docstrings now formatted with epydoc (http://epydoc.sf.net).
1514
1515 * improved Scientific.IO.NetCDF compatibility for netCDF4_classic
1516 (typecode method, ability to use old Numeric typecodes).
1517
1518 * zlib=False or complevel=0 disables shuffle filter in createVariable.
1519
1520 * subversion repository hosted on Google projects
1521 (http://code.google.com/p/netcdf4-python/).
1522
1523 * examples_classic/bench2.py is a performance comparison with
1524 Scientific.IO.NetCDF (the numpy version provided by pynetcdf).
1525
1526 * __dict__ attribute of Dataset, Group or Variable provides a python
1527 dictionary with all netCDF attribute name/value pairs (just like
1528 Scientific.IO.NetCDF).
1529
1530 20060710 - version 0.4.5
1531 ========================
1532 * fixed to work with recent svn versions of numpy
1533
1534 * Now requires at least numpy 0.9.8.
1535
1536 * Raise a AttributeError if user tries to rebind a private attribute
1537 (like 'variables', 'dimensions' or 'dtype').
1538
1539 20060629 - version 0.4.4
1540 ========================
1541 * fixed to work with netcdf-4.0-alpha14.
1542
1543 * automatically cast _FillValue attribute to variable type, to
1544 avoid surprising error message.
1545
1546 20060320 - version 0.4.3
1547 ========================
1548 updated netcdftime module yet again
1549 added 'all_leap'/'366_day' and '360_day' calendars.
1550 netCDFTime class renamed utime, fwd and inv methods
1551 renamed date2num and num2date. These methods can now handle
1552 numpy arrays as well as scalars.
1553 a 'real' python datetime instance is returned if calendar
1554 is gregorian, otherwise a 'datetime-like' instance is returned
1555 (python datetime can't handle funky dates in 'all_leap' and '360_day'
1556 calendars).
1557
1558 20060316 - version 0.4.2
1559 ========================
1560 udunits module replaced by pure python version, renamed 'netcdftime'
1561 No longer requires udunits library. Includes 4 calendars
1562 ('julian','standard'/'gregorian','proleptic_gregorian','noleap'/'365_day').
1563 Calendar names and their interpretations follow the CF metadata convention.
1564
1565 20060310 - version 0.4.1
1566 ========================
1567 udunits module included for doing time conversions.
1568
1569 20060306 - version 0.4
1570 ======================
1571
1572 netCDF4_classic module can now write NETCDF3_CLASSIC, NETCDF4_64BIT
1573 as well as NETCDF4_CLASSIC files. The file format is given as
1574 an optional keyword to the Dataset constructor ('NETCDF4_CLASSIC'
1575 is the default). Preliminary work on compound types done - but
1576 awaiting the next alpha of the netCDF 4 library to complete (bugs
1577 in alpha12 prevent it from working properly if the compound type
1578 has fields which are arrays).
1579
1580 20060217 - version 0.3.1
1581 ========================
1582 refactored user-defined data type support - user-defined
1583 data types are now described by an instance of the class
1584 UserType. usertype and usertype_name keyword args
1585 eliminated from createVariable.
1586
1587 20060214 - version 0.3
1588 ======================
1589 support for variable length strengths (typecode = 'S') and
1590 variable-length, or 'ragged' arrays (vlen user-defined datatype).
1591 Arrays of python objects can be saved as pickled strings with
1592 datatype = 'S'.
1593
1594 20050128 - version 0.2.5
1595 ========================
1596 added support for scalar variables (and assignValue, getValue
1597 Variable methods for Scientific.IO.NetCDF compatibility).
1598
1599 20051123 - version 0.2.4
1600 ========================
1601 numpy 0.9.4 compatibility
1602 Changed data type codes from ('d', 'f', 'i', 'h', ...) to
1603 ('f8', 'f4', 'i4', 'i2', ...).
1604
1605 20050110 - version 0.2.3
1606 ========================
1607 added ellipsis slicing capability
1608
1609 20050106 - version 0.2.2
1610 ========================
1611 changed scipy_core to numpy.
1612
1613 20051228 - version 0.2.1
1614 ========================
1615 bugfixes, added 'nc3tonc4' utility to convert netCDF version 3 files
1616 to NETCDF4_CLASSIC files (with compression). The converted files
1617 can be read from netCDF 3 clients that have been re-linked to the netCDF 4
1618 library. 'chunking' keyword added to createVariable in netCDF4 module.
1619
1620 20051224 - version 0.2
1621 ======================
1622 Added netCDF4_classic module - which creates files in NETCDF4_CLASSIC
1623 format. These files are compatible with netCDF 3 clients which have
1624 been linked against the netCDF 4 lib. This module does not use any
1625 new features of the netCDF 4 API except zlib compression. Unlike
1626 any other netCDF 3 python client, it can transparently compress data
1627 with zlib compression and the HDF5 shuffle filter.
1628
1629 20051222 - version 0.1
1630 ======================
1631 First release. Supports groups, multiple unlimited dimensions, zlib
1632 compression (plus shuffle filter and fletcher32 checksum) and all new
1633 primitive data types. No support for user-defined data types yet.