1 | # Copyright (C) 2004 CCLRC & NERC( Natural Environment Research Council ). |
---|
2 | # This software may be distributed under the terms of the |
---|
3 | # Q Public License, version 1.0 or later. http://ndg.nerc.ac.uk/public_docs/QPublic_license.txt |
---|
4 | |
---|
5 | """ |
---|
6 | nc_to_na.py |
---|
7 | ============= |
---|
8 | |
---|
9 | Holds the class NCToNA (sub-classing CDMSToNA) that converts a NetCDF file to |
---|
10 | one or more NASA Ames files. |
---|
11 | |
---|
12 | """ |
---|
13 | |
---|
14 | # Imports from python standard library |
---|
15 | import sys |
---|
16 | |
---|
17 | # Import from nappy package |
---|
18 | import nappy |
---|
19 | from nappy.na_error import na_error |
---|
20 | import nappy.utils |
---|
21 | import nappy.utils.common_utils |
---|
22 | import nappy.nc_interface.cdms_to_na |
---|
23 | import nappy.nc_interface.na_content_collector |
---|
24 | |
---|
25 | # Import external packages (if available) |
---|
26 | if sys.platform.find("win") > -1: |
---|
27 | raise na_error.NAPlatformError("Windows does not support CDMS. CDMS is required to convert to CDMS objects and NetCDF.") |
---|
28 | try: |
---|
29 | import cdms, Numeric |
---|
30 | except: |
---|
31 | raise Exception("Could not import third-party software. Nappy requires the CDMS and Numeric packages to be installed to convert to CDMS and NetCDF.") |
---|
32 | |
---|
33 | cdms.setAutoBounds("off") |
---|
34 | |
---|
35 | # Define global variables |
---|
36 | DEBUG = nappy.utils.getDebug() |
---|
37 | |
---|
38 | class NCToNA(nappy.nc_interface.cdms_to_na.CDMSToNA): |
---|
39 | """ |
---|
40 | Converts a NetCDF file to one or more NASA Ames files. |
---|
41 | """ |
---|
42 | |
---|
43 | def __init__(self, nc_file, var_ids=None, na_items_to_override={}, |
---|
44 | only_return_file_names=False, exclude_vars=[], |
---|
45 | requested_ffi=None): |
---|
46 | """ |
---|
47 | Sets up instance variables. |
---|
48 | Typical usage is: |
---|
49 | >>> import nappy.nc_interface.nc_to_na as nc_to_na |
---|
50 | >>> c = nc_to_na.NCToNA("old_file.nc") |
---|
51 | >>> c.convert() |
---|
52 | >>> c.writeNAFiles("new_file.na", delimiter=",") |
---|
53 | |
---|
54 | OR: |
---|
55 | >>> c = nc_to_na.NCToNA("old_file.nc") |
---|
56 | >>> file_names = c.constructNAFileNames() |
---|
57 | """ |
---|
58 | self.nc_file = nc_file |
---|
59 | |
---|
60 | # Now need to read CDMS file so parent class methods are compatible |
---|
61 | (cdms_variables, global_atts) = self._readCDMSFile(var_ids, exclude_vars) |
---|
62 | nappy.nc_interface.cdms_to_na.CDMSToNA.__init__(self, cdms_variables, global_atts=global_atts, |
---|
63 | na_items_to_override=na_items_to_override, |
---|
64 | only_return_file_names=only_return_file_names, |
---|
65 | requested_ffi=requested_ffi) |
---|
66 | |
---|
67 | |
---|
68 | def _readCDMSFile(self, var_ids=None, exclude_vars=[]): |
---|
69 | """ |
---|
70 | Reads the file and returns all the CDMS variables in a list as well |
---|
71 | as the global attributes: (cdms_variable_list, global_atts_dict) |
---|
72 | If var_ids is defined then only get those. |
---|
73 | """ |
---|
74 | fin = cdms.open(self.nc_file) |
---|
75 | cdms_variables = [] |
---|
76 | |
---|
77 | # Make sure var_ids is a list |
---|
78 | if type(var_ids) == type("string"): |
---|
79 | var_ids = [var_ids] |
---|
80 | |
---|
81 | for var_id in fin.listvariables(): |
---|
82 | if var_ids == None or var_id in var_ids: |
---|
83 | if var_ids not in exclude_vars: |
---|
84 | cdms_variables.append(fin(var_id)) |
---|
85 | |
---|
86 | globals = fin.attributes |
---|
87 | return (cdms_variables, globals) |
---|
88 | |
---|
89 | def constructNAFileNames(self, na_file=None): |
---|
90 | """ |
---|
91 | Works out what the file names of the output NA files will be and |
---|
92 | returns a list of them. |
---|
93 | """ |
---|
94 | self.convert() |
---|
95 | |
---|
96 | file_names = [] |
---|
97 | # create file name if not given |
---|
98 | if na_file == None: |
---|
99 | base_name = self.nc_file |
---|
100 | if base_name[-3:] == ".nc": |
---|
101 | base_name = base_name[:-3] |
---|
102 | na_file = base_name + ".na" |
---|
103 | |
---|
104 | file_counter = 1 |
---|
105 | # Now, create some valid file names |
---|
106 | for this_na_dict in self.na_dict_list: |
---|
107 | if len(self.na_dict_list) == 1: |
---|
108 | suffix = "" |
---|
109 | else: |
---|
110 | suffix = "_%s" % file_counter |
---|
111 | |
---|
112 | # Create file name |
---|
113 | name_parts = na_file.split(".") |
---|
114 | new_name = (".".join(name_parts[:-1])) + suffix + "." + name_parts[-1] |
---|
115 | file_names.append(new_name) |
---|
116 | file_counter += 1 |
---|
117 | |
---|
118 | return file_names |
---|
119 | |
---|
120 | def writeNAFiles(self, na_file=None, delimiter=" ", |
---|
121 | float_format="%g", size_limit=None): |
---|
122 | """ |
---|
123 | Writes the self.na_dict_list content to one or more NASA Ames files. |
---|
124 | Output file names are based on the self.nc_file name unless specified |
---|
125 | in the na_file_name argument in which case that provides the main name |
---|
126 | that is appended to if multiple output file names are required. |
---|
127 | """ |
---|
128 | self.convert() # just in case not already called |
---|
129 | |
---|
130 | # Gets a list of NA file_names that will be produced. |
---|
131 | file_names = self.constructNAFileNames(na_file) |
---|
132 | |
---|
133 | # Set up some counters: file_counter is the expected number of files. |
---|
134 | # full_file_counter includes any files that have been split across multiple output NA files |
---|
135 | # because size_limit exceeded. |
---|
136 | file_counter = 1 |
---|
137 | full_file_counter = 1 |
---|
138 | |
---|
139 | # Now loop through writing the outputs |
---|
140 | for na_dict_and_var_ids in self.na_dict_list: |
---|
141 | file_name = file_names[file_counter - 1] |
---|
142 | msg = "\nWriting output NASA Ames file: %s" % file_name |
---|
143 | if DEBUG: print msg |
---|
144 | self.output_message.append(msg) |
---|
145 | |
---|
146 | # Set up current na dict |
---|
147 | (this_na_dict, vars_to_write) = na_dict_and_var_ids |
---|
148 | |
---|
149 | # Override content of NASA Ames if they are permitted |
---|
150 | for key in self.na_items_to_override.keys(): |
---|
151 | if key in permitted_overwrite_metadata: |
---|
152 | if key in items_as_lists: |
---|
153 | new_item = self.na_items_to_override[key].split() |
---|
154 | else: |
---|
155 | new_item = self.na_items_to_override[key] |
---|
156 | |
---|
157 | if new_item != this_na_dict[key]: |
---|
158 | this_na_dict[key] = new_item |
---|
159 | msg = "Metadata overwritten in output file: '%s' is now '%s'" % (key, this_na_dict[key]) |
---|
160 | if DEBUG: print msg |
---|
161 | self.output_message.append(msg) |
---|
162 | |
---|
163 | file_list = [] |
---|
164 | # Cope with size limits if specified and FFI is 1001 |
---|
165 | # Seems to be writing different chunks of a too long array to different na_dicts to then write to separate files. |
---|
166 | if size_limit is not None and (this_na_dict["FFI"] == 1001 and len(this_na_dict["V"][0]) > size_limit): |
---|
167 | files_written = self._writeNAFileSubsetsWithinSizeLimit(this_na_dict, file_name, delimiter=delimiter, |
---|
168 | float_format=float_format, size_limit=size_limit) |
---|
169 | file_list.extend(files_written) |
---|
170 | |
---|
171 | # If not having to split file into multiple outputs (normal condition) |
---|
172 | else: |
---|
173 | x = nappy.openNAFile(file_name, 'w', this_na_dict) |
---|
174 | x.write(delimiter=delimiter, float_format=float_format) |
---|
175 | x.close() |
---|
176 | file_list.append(file_name) |
---|
177 | |
---|
178 | # Report on what has been written |
---|
179 | msg = "\nWrote the following variables:" + "\n\t" + ("\n\t".join(vars_to_write[0])) |
---|
180 | if DEBUG: print msg |
---|
181 | self.output_message.append(msg) |
---|
182 | |
---|
183 | msg = "" |
---|
184 | aux_var_count = vars_to_write[1] |
---|
185 | if len(aux_var_count) > 0: |
---|
186 | msg = "\nWrote the following auxiliary variables:" + "\n\t" + ("\n\t".join(aux_var_count)) |
---|
187 | |
---|
188 | singleton_var_count = vars_to_write[2] |
---|
189 | if len(singleton_var_count) > 0: |
---|
190 | msg = "\nWrote the following Singleton variables:" + "\n\t" + ("\n\t".join(singleton_var_count)) |
---|
191 | |
---|
192 | if len(file_list) > 0: |
---|
193 | msg = msg + ("\n\nNASA Ames file(s) written successfully: \n%s" % "\n".join(file_list)) |
---|
194 | |
---|
195 | full_file_counter += len(file_list) |
---|
196 | file_counter += 1 |
---|
197 | |
---|
198 | if DEBUG: print msg |
---|
199 | self.output_message.append(msg) |
---|
200 | |
---|
201 | full_file_count = full_file_counter - 1 |
---|
202 | if full_file_count == 1: |
---|
203 | plural = "" |
---|
204 | else: |
---|
205 | plural = "s" |
---|
206 | msg = "\n%s file%s written." % (full_file_count, plural) |
---|
207 | |
---|
208 | if DEBUG: print msg |
---|
209 | self.output_message.append(msg) |
---|
210 | return self.output_message |
---|
211 | |
---|
212 | def _writeNAFileSubsetsWithinSizeLimit(self, this_na_dict, file_name, delimiter, float_format, size_limit): |
---|
213 | """ |
---|
214 | If self.size_limit is specified and FFI is 1001 we can chunk the output into |
---|
215 | different files in a NASA Ames compliant way. |
---|
216 | Returns list of file names of outputs written. |
---|
217 | """ |
---|
218 | file_names = [] |
---|
219 | var_list = this_na_dict["V"] |
---|
220 | array_length = len(var_list[0]) |
---|
221 | nvol_info = divmod(array_length, size_limit) |
---|
222 | nvol = nvol_info[0] |
---|
223 | |
---|
224 | # create the number of volumes (files) that need to be written. |
---|
225 | if nvol_info[1] > 0: nvol = nvol + 1 |
---|
226 | |
---|
227 | start = 0 |
---|
228 | letter_count = 0 |
---|
229 | ivol = 0 |
---|
230 | |
---|
231 | # Loop through until full array length has been written to a set of files. |
---|
232 | while start < array_length: |
---|
233 | ivol = ivol + 1 |
---|
234 | end = start + size_limit |
---|
235 | |
---|
236 | if end > array_length: |
---|
237 | end = array_length |
---|
238 | |
---|
239 | current_block = [] |
---|
240 | # Write new V array |
---|
241 | for v in var_list: |
---|
242 | current_block.append(v[start:end]) |
---|
243 | |
---|
244 | # Adjust X accordingly in the na dictionary, because independent variable has been reduced in size |
---|
245 | na_dict_copy = nappy.utils.common_utils.modifyNADictCopy(this_na_dict, current_block, |
---|
246 | start, end, ivol, nvol) |
---|
247 | # Append a letter to the file name for writing this block to |
---|
248 | file_name_plus_letter = "%s-%.3d.na" % (file_name[:-3], ivol) |
---|
249 | file_list.append(file_name_plus_letter) |
---|
250 | |
---|
251 | # Write data to output file |
---|
252 | x = nappy.openNAFile(file_name_plus_letter, 'w', na_dict_copy) |
---|
253 | x.write(delimiter=delimiter, float_format=float_format) |
---|
254 | x.close() |
---|
255 | |
---|
256 | msg = "\nOutput files split on size limit: %s\nFilename used: %s" % (size_limit, file_name_plus_letter) |
---|
257 | if DEBUG: print msg |
---|
258 | self.output_message.append(msg) |
---|
259 | letter_count = letter_count + 1 |
---|
260 | start = end |
---|
261 | |
---|
262 | file_names.append(file_name_plus_letter) |
---|
263 | |
---|
264 | return file_names |
---|