1 | #!/usr/bin/env python |
---|
2 | # -*- coding: utf-8 -*- |
---|
3 | #******************************************************************************* |
---|
4 | # @Author: Anne Fouilloux (University of Oslo) |
---|
5 | # |
---|
6 | # @Date: October 2014 |
---|
7 | # |
---|
8 | # @Change History: |
---|
9 | # |
---|
10 | # November 2015 - Leopold Haimberger (University of Vienna): |
---|
11 | # - optimized display_info |
---|
12 | # - optimized data_retrieve and seperate between python and shell |
---|
13 | # script call |
---|
14 | # |
---|
15 | # February 2018 - Anne Philipp (University of Vienna): |
---|
16 | # - applied PEP8 style guide |
---|
17 | # - added documentation |
---|
18 | # - applied some minor modifications in programming style/structure |
---|
19 | # - added writing of mars request attributes to a csv file |
---|
20 | # |
---|
21 | # @License: |
---|
22 | # (C) Copyright 2014-2019. |
---|
23 | # Anne Philipp, Leopold Haimberger |
---|
24 | # |
---|
25 | # This work is licensed under the Creative Commons Attribution 4.0 |
---|
26 | # International License. To view a copy of this license, visit |
---|
27 | # http://creativecommons.org/licenses/by/4.0/ or send a letter to |
---|
28 | # Creative Commons, PO Box 1866, Mountain View, CA 94042, USA. |
---|
29 | #******************************************************************************* |
---|
30 | |
---|
31 | # ------------------------------------------------------------------------------ |
---|
32 | # MODULES |
---|
33 | # ------------------------------------------------------------------------------ |
---|
34 | from __future__ import print_function |
---|
35 | |
---|
36 | import os |
---|
37 | import sys |
---|
38 | import subprocess |
---|
39 | import traceback |
---|
40 | |
---|
41 | # software specific classes and modules from flex_extract |
---|
42 | sys.path.append('../') |
---|
43 | import _config |
---|
44 | try: |
---|
45 | ec_api = True |
---|
46 | import ecmwfapi |
---|
47 | except ImportError: |
---|
48 | ec_api = False |
---|
49 | |
---|
50 | try: |
---|
51 | cds_api = True |
---|
52 | import cdsapi |
---|
53 | except ImportError: |
---|
54 | cds_api = False |
---|
55 | # ------------------------------------------------------------------------------ |
---|
56 | # CLASS |
---|
57 | # ------------------------------------------------------------------------------ |
---|
58 | class MarsRetrieval(object): |
---|
59 | '''Specific syntax and content for submission of MARS retrievals. |
---|
60 | |
---|
61 | A MARS revtrieval has a specific syntax with a selection of keywords and |
---|
62 | their corresponding values. This class provides the necessary functions |
---|
63 | by displaying the selected parameters and their values and the actual |
---|
64 | retrievement of the data through a mars request or a Python web api |
---|
65 | interface. The initialization already expects all the keyword values. |
---|
66 | |
---|
67 | A description of MARS keywords/arguments and examples of their |
---|
68 | values can be found here: |
---|
69 | https://software.ecmwf.int/wiki/display/UDOC/\ |
---|
70 | Identification+keywords#Identificationkeywords-class |
---|
71 | |
---|
72 | Attributes |
---|
73 | ---------- |
---|
74 | server : ECMWFService or ECMWFDataServer |
---|
75 | This is the connection to the ECMWF data servers. |
---|
76 | |
---|
77 | public : int |
---|
78 | Decides which Web API Server version is used. |
---|
79 | |
---|
80 | marsclass : str, optional |
---|
81 | Characterisation of dataset. |
---|
82 | |
---|
83 | dataset : str, optional |
---|
84 | For public datasets there is the specific naming and parameter |
---|
85 | dataset which has to be used to characterize the type of |
---|
86 | data. |
---|
87 | |
---|
88 | type : str, optional |
---|
89 | Determines the type of fields to be retrieved. |
---|
90 | |
---|
91 | levtype : str, optional |
---|
92 | Denotes type of level. |
---|
93 | |
---|
94 | levelist : str, optional |
---|
95 | Specifies the required levels. |
---|
96 | |
---|
97 | repres : str, optional |
---|
98 | Selects the representation of the archived data. |
---|
99 | |
---|
100 | date : str, optional |
---|
101 | Specifies the Analysis date, the Forecast base date or |
---|
102 | Observations date. |
---|
103 | |
---|
104 | resol : str, optional |
---|
105 | Specifies the desired triangular truncation of retrieved data, |
---|
106 | before carrying out any other selected post-processing. |
---|
107 | |
---|
108 | stream : str, optional |
---|
109 | Identifies the forecasting system used to generate the data. |
---|
110 | |
---|
111 | area : str, optional |
---|
112 | Specifies the desired sub-area of data to be extracted. |
---|
113 | |
---|
114 | time : str, optional |
---|
115 | Specifies the time of the data in hours and minutes. |
---|
116 | |
---|
117 | step : str, optional |
---|
118 | Specifies the forecast time step from forecast base time. |
---|
119 | |
---|
120 | expver : str, optional |
---|
121 | The version of the dataset. |
---|
122 | |
---|
123 | number : str, optional |
---|
124 | Selects the member in ensemble forecast run. |
---|
125 | |
---|
126 | accuracy : str, optional |
---|
127 | Specifies the number of bits per value to be used in the |
---|
128 | generated GRIB coded fields. |
---|
129 | |
---|
130 | grid : str, optional |
---|
131 | Specifies the output grid which can be either a Gaussian grid |
---|
132 | or a Latitude/Longitude grid. |
---|
133 | |
---|
134 | gaussian : str, optional |
---|
135 | This parameter is deprecated and should no longer be used. |
---|
136 | Specifies the desired type of Gaussian grid for the output. |
---|
137 | |
---|
138 | target : str, optional |
---|
139 | Specifies a file into which data is to be written after |
---|
140 | retrieval or manipulation. |
---|
141 | |
---|
142 | param : str, optional |
---|
143 | Specifies the meteorological parameter. |
---|
144 | ''' |
---|
145 | |
---|
146 | def __init__(self, server, public, marsclass="EA", dataset="", type="", |
---|
147 | levtype="", levelist="", repres="", date="", resol="", |
---|
148 | stream="", area="", time="", step="", expver="1", |
---|
149 | number="", accuracy="", grid="", gaussian="", target="", |
---|
150 | param=""): |
---|
151 | '''Initialises the instance of the MarsRetrieval class and |
---|
152 | defines and assigns a set of the necessary retrieval parameters |
---|
153 | for the FLEXPART input data. |
---|
154 | A description of MARS keywords/arguments, their dependencies |
---|
155 | on each other and examples of their values can be found here: |
---|
156 | |
---|
157 | https://software.ecmwf.int/wiki/display/UDOC/MARS+keywords |
---|
158 | |
---|
159 | Parameters |
---|
160 | ---------- |
---|
161 | server : ECMWFService or ECMWFDataServer |
---|
162 | This is the connection to the ECMWF data servers. |
---|
163 | It is needed for the pythonic access of ECMWF data. |
---|
164 | |
---|
165 | public : int |
---|
166 | Decides which Web API version is used: |
---|
167 | 0: member-state users and full archive access |
---|
168 | 1: public access and limited access to the public server and |
---|
169 | datasets. Needs the parameter dataset. |
---|
170 | Default is "0" and for member-state users. |
---|
171 | |
---|
172 | marsclass : str, optional |
---|
173 | Characterisation of dataset. E.g. EI (ERA-Interim), |
---|
174 | E4 (ERA40), OD (Operational archive), EA (ERA5). |
---|
175 | Default is the ERA5 dataset "EA". |
---|
176 | |
---|
177 | dataset : str, optional |
---|
178 | For public datasets there is the specific naming and parameter |
---|
179 | dataset which has to be used to characterize the type of |
---|
180 | data. Usually there is less data available, either in times, |
---|
181 | domain or parameter. |
---|
182 | Default is an empty string. |
---|
183 | |
---|
184 | type : str, optional |
---|
185 | Determines the type of fields to be retrieved. |
---|
186 | Selects between observations, images or fields. |
---|
187 | Examples for fields: Analysis (an), Forecast (fc), |
---|
188 | Perturbed Forecast (pf), Control Forecast (cf) and so on. |
---|
189 | Default is an empty string. |
---|
190 | |
---|
191 | levtype : str, optional |
---|
192 | Denotes type of level. Has a direct implication on valid |
---|
193 | levelist values! |
---|
194 | E.g. model level (ml), pressure level (pl), surface (sfc), |
---|
195 | potential vorticity (pv), potential temperature (pt) |
---|
196 | and depth (dp). |
---|
197 | Default is an empty string. |
---|
198 | |
---|
199 | levelist : str, optional |
---|
200 | Specifies the required levels. It has to have a valid |
---|
201 | correspondence to the selected levtype. |
---|
202 | Examples: model level: 1/to/137, pressure levels: 500/to/1000 |
---|
203 | Default is an empty string. |
---|
204 | |
---|
205 | repres : str, optional |
---|
206 | Selects the representation of the archived data. |
---|
207 | E.g. sh - spherical harmonics, gg - Gaussian grid, |
---|
208 | ll - latitude/longitude, ... |
---|
209 | Default is an empty string. |
---|
210 | |
---|
211 | date : str, optional |
---|
212 | Specifies the Analysis date, the Forecast base date or |
---|
213 | Observations date. Valid formats are: |
---|
214 | Absolute as YYYY-MM-DD or YYYYMMDD. |
---|
215 | Default is an empty string. |
---|
216 | |
---|
217 | resol : str, optional |
---|
218 | Specifies the desired triangular truncation of retrieved data, |
---|
219 | before carrying out any other selected post-processing. |
---|
220 | The default is automatic truncation (auto), by which the lowest |
---|
221 | resolution compatible with the value specified in grid is |
---|
222 | automatically selected for the retrieval. |
---|
223 | Users wanting to perform post-processing from full spectral |
---|
224 | resolution should specify Archived Value (av). |
---|
225 | The following are examples of existing resolutions found in |
---|
226 | the archive: 63, 106, 159, 213, 255, 319, 399, 511, 799 or 1279. |
---|
227 | This keyword has no meaning/effect if the archived data is |
---|
228 | not in spherical harmonics representation. |
---|
229 | The best selection can be found here: |
---|
230 | https://software.ecmwf.int/wiki/display/UDOC/\ |
---|
231 | Retrieve#Retrieve-Truncationbeforeinterpolation |
---|
232 | Default is an empty string. |
---|
233 | |
---|
234 | stream : str, optional |
---|
235 | Identifies the forecasting system used to generate the data. |
---|
236 | E.g. oper (Atmospheric model), enfo (Ensemble forecats), ... |
---|
237 | Default is an empty string. |
---|
238 | |
---|
239 | area : str, optional |
---|
240 | Specifies the desired sub-area of data to be extracted. |
---|
241 | Areas can be defined to wrap around the globe. |
---|
242 | |
---|
243 | Latitude values must be given as signed numbers, with: |
---|
244 | north latitudes (i.e. north of the equator) |
---|
245 | being positive (e.g: 40.5) |
---|
246 | south latitutes (i.e. south of the equator) |
---|
247 | being negative (e.g: -50.5) |
---|
248 | Longtitude values must be given as signed numbers, with: |
---|
249 | east longitudes (i.e. east of the 0 degree meridian) |
---|
250 | being positive (e.g: 35.0) |
---|
251 | west longitudes (i.e. west of the 0 degree meridian) |
---|
252 | being negative (e.g: -20.5) |
---|
253 | |
---|
254 | E.g.: North/West/South/East |
---|
255 | Default is an empty string. |
---|
256 | |
---|
257 | time : str, optional |
---|
258 | Specifies the time of the data in hours and minutes. |
---|
259 | Valid values depend on the type of data: Analysis time, |
---|
260 | Forecast base time or First guess verification time |
---|
261 | (all usually at synoptic hours: 00, 06, 12 and 18 ). |
---|
262 | Observation time (any combination in hours and minutes is valid, |
---|
263 | subject to data availability in the archive). |
---|
264 | The syntax is HHMM or HH:MM. If MM is omitted it defaults to 00. |
---|
265 | Default is an empty string. |
---|
266 | |
---|
267 | step : str, optional |
---|
268 | Specifies the forecast time step from forecast base time. |
---|
269 | Valid values are hours (HH) from forecast base time. It also |
---|
270 | specifies the length of the forecast which verifies at |
---|
271 | First Guess time. |
---|
272 | E.g. 1/3/6-hourly |
---|
273 | Default is an empty string. |
---|
274 | |
---|
275 | expver : str, optional |
---|
276 | The version of the dataset. Each experiment is assigned a |
---|
277 | unique code (version). Production data is assigned 1 or 2, |
---|
278 | and experimental data in Operations 11, 12 ,... |
---|
279 | Research or Member State's experiments have a four letter |
---|
280 | experiment identifier. |
---|
281 | Default is "1". |
---|
282 | |
---|
283 | number : str, optional |
---|
284 | Selects the member in ensemble forecast run. (Only then it |
---|
285 | is necessary.) It has a different meaning depending on |
---|
286 | the type of data. |
---|
287 | E.g. Perturbed Forecasts: specifies the Ensemble forecast member |
---|
288 | Default is an empty string. |
---|
289 | |
---|
290 | accuracy : str, optional |
---|
291 | Specifies the number of bits per value to be used in the |
---|
292 | generated GRIB coded fields. |
---|
293 | A positive integer may be given to specify the preferred number |
---|
294 | of bits per packed value. This must not be greater than the |
---|
295 | number of bits normally used for a Fortran integer on the |
---|
296 | processor handling the request (typically 32 or 64 bit). |
---|
297 | Within a compute request the accuracy of the original fields |
---|
298 | can be passed to the result field by specifying accuracy=av. |
---|
299 | Default is an empty string. |
---|
300 | |
---|
301 | grid : str, optional |
---|
302 | Specifies the output grid which can be either a Gaussian grid |
---|
303 | or a Latitude/Longitude grid. MARS requests specifying |
---|
304 | grid=av will return the archived model grid. |
---|
305 | |
---|
306 | Lat/Lon grid: The grid spacing needs to be an integer |
---|
307 | fraction of 90 degrees e.g. grid = 0.5/0.5 |
---|
308 | |
---|
309 | Gaussian grid: specified by a letter denoting the type of |
---|
310 | Gaussian grid followed by an integer (the grid number) |
---|
311 | representing the number of lines between the Pole and Equator, |
---|
312 | e.g. |
---|
313 | grid = F160 - full (or regular) Gaussian grid with |
---|
314 | 160 latitude lines between the pole and equator |
---|
315 | grid = N320 - ECMWF original reduced Gaussian grid with |
---|
316 | 320 latitude lines between the pole and equator, |
---|
317 | see Reduced Gaussian Grids for grid numbers used at ECMWF |
---|
318 | grid = O640 - ECMWF octahedral (reduced) Gaussian grid with |
---|
319 | 640 latitude lines between the pole and equator |
---|
320 | Default is an empty string. |
---|
321 | |
---|
322 | gaussian : str, optional |
---|
323 | This parameter is deprecated and should no longer be used. |
---|
324 | Specifies the desired type of Gaussian grid for the output. |
---|
325 | Valid Gaussian grids are quasi-regular (reduced) or regular. |
---|
326 | Keyword gaussian can only be specified together with |
---|
327 | keyword grid. Gaussian without grid has no effect. |
---|
328 | Default is an empty string. |
---|
329 | |
---|
330 | target : str, optional |
---|
331 | Specifies a file into which data is to be written after |
---|
332 | retrieval or manipulation. Path names should always be |
---|
333 | enclosed in double quotes. The MARS client supports automatic |
---|
334 | generation of multiple target files using MARS keywords |
---|
335 | enclosed in square brackets [ ]. If the environment variable |
---|
336 | MARS_MULTITARGET_STRICT_FORMAT is set to 1 before calling mars, |
---|
337 | the keyword values will be used in the filename as shown by |
---|
338 | the ecCodes GRIB tool grib_ls -m, e.g. with |
---|
339 | MARS_MULTITARGET_STRICT_FORMAT set to 1 the keywords time, |
---|
340 | expver and param will be formatted as 0600, 0001 and 129.128 |
---|
341 | rather than 600, 1 and 129. |
---|
342 | Default is an empty string. |
---|
343 | |
---|
344 | param : str, optional |
---|
345 | Specifies the meteorological parameter. |
---|
346 | The list of meteorological parameters in MARS is extensive. |
---|
347 | Their availability is directly related to their meteorological |
---|
348 | meaning and, therefore, the rest of directives specified |
---|
349 | in the MARS request. |
---|
350 | Meteorological parameters can be specified by their |
---|
351 | GRIB code (param=130), their mnemonic (param=t) or |
---|
352 | full name (param=temperature). |
---|
353 | The list of parameter should be seperated by a "/"-sign. |
---|
354 | E.g. 130/131/133 |
---|
355 | Default is an empty string. |
---|
356 | |
---|
357 | Return |
---|
358 | ------ |
---|
359 | |
---|
360 | ''' |
---|
361 | |
---|
362 | self.server = server |
---|
363 | self.public = public |
---|
364 | self.marsclass = marsclass |
---|
365 | self.dataset = dataset |
---|
366 | self.type = type |
---|
367 | self.levtype = levtype |
---|
368 | self.levelist = levelist |
---|
369 | self.repres = repres |
---|
370 | self.date = date |
---|
371 | self.resol = resol |
---|
372 | self.stream = stream |
---|
373 | self.area = area |
---|
374 | self.time = time |
---|
375 | self.step = step |
---|
376 | self.expver = expver |
---|
377 | self.number = number |
---|
378 | self.accuracy = accuracy |
---|
379 | self.grid = grid |
---|
380 | self.gaussian = gaussian |
---|
381 | self.target = target |
---|
382 | self.param = param |
---|
383 | |
---|
384 | return |
---|
385 | |
---|
386 | |
---|
387 | def display_info(self): |
---|
388 | '''Prints all class attributes and their values to the |
---|
389 | standard output. |
---|
390 | |
---|
391 | Parameters |
---|
392 | ---------- |
---|
393 | |
---|
394 | Return |
---|
395 | ------ |
---|
396 | |
---|
397 | ''' |
---|
398 | # Get all class attributes and their values as a dictionary |
---|
399 | attrs = vars(self).copy() |
---|
400 | |
---|
401 | # iterate through all attributes and print them |
---|
402 | # with their corresponding values |
---|
403 | for item in attrs.items(): |
---|
404 | if item[0] in ['server', 'public']: |
---|
405 | pass |
---|
406 | else: |
---|
407 | print(item[0] + ': ' + str(item[1])) |
---|
408 | |
---|
409 | return |
---|
410 | |
---|
411 | |
---|
412 | def print_infodata_csv(self, inputdir, request_number): |
---|
413 | '''Write all request parameter in alpabetical order into a "csv" file. |
---|
414 | |
---|
415 | Parameters |
---|
416 | ---------- |
---|
417 | inputdir : str |
---|
418 | The path where all data from the retrievals are stored. |
---|
419 | |
---|
420 | request_number : int |
---|
421 | Number of mars requests for flux and non-flux data. |
---|
422 | |
---|
423 | Return |
---|
424 | ------ |
---|
425 | |
---|
426 | ''' |
---|
427 | |
---|
428 | # Get all class attributes and their values as a dictionary |
---|
429 | attrs = vars(self).copy() |
---|
430 | del attrs['server'] |
---|
431 | del attrs['public'] |
---|
432 | |
---|
433 | # open a file to store all requests to |
---|
434 | with open(os.path.join(inputdir, |
---|
435 | _config.FILE_MARS_REQUESTS), 'a') as f: |
---|
436 | f.write(str(request_number) + ', ') |
---|
437 | f.write(', '.join(str(attrs[key]) |
---|
438 | for key in sorted(attrs.keys()))) |
---|
439 | f.write('\n') |
---|
440 | |
---|
441 | return |
---|
442 | |
---|
443 | def data_retrieve(self): |
---|
444 | '''Submits a MARS retrieval. Depending on the existence of |
---|
445 | ECMWF Web-API or CDS API it is submitted via Python or a |
---|
446 | subprocess in the Shell. The parameter for the mars retrieval |
---|
447 | are taken from the defined class attributes. |
---|
448 | |
---|
449 | Parameters |
---|
450 | ---------- |
---|
451 | |
---|
452 | Return |
---|
453 | ------ |
---|
454 | |
---|
455 | ''' |
---|
456 | # Get all class attributes and their values as a dictionary |
---|
457 | attrs = vars(self).copy() |
---|
458 | |
---|
459 | # eliminate unnecessary attributes from the dictionary attrs |
---|
460 | del attrs['server'] |
---|
461 | del attrs['public'] |
---|
462 | |
---|
463 | # exchange parameter name for marsclass |
---|
464 | mclass = attrs.get('marsclass') |
---|
465 | del attrs['marsclass'] |
---|
466 | attrs['class'] = mclass |
---|
467 | |
---|
468 | # prepare target variable as needed for the Web API or CDS API mode |
---|
469 | # within the dictionary for full access |
---|
470 | # as a single variable for public access |
---|
471 | target = attrs.get('target') |
---|
472 | if not int(self.public): |
---|
473 | del attrs['target'] |
---|
474 | print('target: ' + target) |
---|
475 | |
---|
476 | # find all keys without a value and convert all other values to strings |
---|
477 | empty_keys = [] |
---|
478 | for key, value in attrs.items(): |
---|
479 | if value == '': |
---|
480 | empty_keys.append(str(key)) |
---|
481 | else: |
---|
482 | attrs[key] = str(value) |
---|
483 | |
---|
484 | # delete all empty parameter from the dictionary |
---|
485 | for key in empty_keys: |
---|
486 | del attrs[key] |
---|
487 | |
---|
488 | attrs['ppengine'] = 'emos' |
---|
489 | |
---|
490 | # MARS request via Python script |
---|
491 | if self.server: |
---|
492 | try: |
---|
493 | if cds_api and isinstance(self.server, cdsapi.Client): |
---|
494 | print('RETRIEVE ERA5 WITH CDS API!') |
---|
495 | self.server.retrieve(_config.CDS_DATASET, |
---|
496 | attrs, target) |
---|
497 | elif ec_api and isinstance(self.server, ecmwfapi.ECMWFDataServer): |
---|
498 | print('RETRIEVE PUBLIC DATA (NOT ERA5)!') |
---|
499 | self.server.retrieve(attrs) |
---|
500 | elif ec_api and isinstance(self.server, ecmwfapi.ECMWFService): |
---|
501 | print('EXECUTE NON-PUBLIC RETRIEVAL (NOT ERA5)!') |
---|
502 | self.server.execute(attrs, target) |
---|
503 | else: |
---|
504 | print('ERROR:') |
---|
505 | print('No match for Web API instance!') |
---|
506 | raise IOError |
---|
507 | except Exception as e: |
---|
508 | print('\n\nMARS Request failed!') |
---|
509 | print(e) |
---|
510 | tb = sys.exc_info()[2] |
---|
511 | print(traceback.format_exc()) |
---|
512 | sys.exit() |
---|
513 | |
---|
514 | # MARS request via call in shell |
---|
515 | else: |
---|
516 | request_str = 'ret' |
---|
517 | for key, value in attrs.items(): |
---|
518 | request_str = request_str + ',' + key + '=' + str(value) |
---|
519 | request_str += ',target="' + target + '"' |
---|
520 | p = subprocess.Popen(['mars', '-e'], |
---|
521 | stdin=subprocess.PIPE, |
---|
522 | stdout=subprocess.PIPE, |
---|
523 | stderr=subprocess.PIPE, |
---|
524 | bufsize=1) |
---|
525 | pout = p.communicate(input=request_str.encode())[0] |
---|
526 | print(pout.decode()) |
---|
527 | |
---|
528 | if 'Some errors reported' in pout.decode(): |
---|
529 | print('MARS Request failed - please check request') |
---|
530 | raise IOError |
---|
531 | elif os.stat(target).st_size == 0: |
---|
532 | print('MARS Request returned no data - please check request') |
---|
533 | raise IOError |
---|
534 | |
---|
535 | return |
---|