1 |
# Copyright (c) 2004 by Intevation GmbH |
2 |
# Authors: |
3 |
# Martin Schulze <[email protected]> |
4 |
# |
5 |
# This program is free software; you can redistribute it and/or modify |
6 |
# it under the terms of the GNU General Public License as published by |
7 |
# the Free Software Foundation; either version 2 of the License, or |
8 |
# (at your option) any later version. |
9 |
# |
10 |
# This program is distributed in the hope that it will be useful, |
11 |
# but WITHOUT ANY WARRANTY; without even the implied warranty of |
12 |
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
13 |
# GNU General Public License for more details. |
14 |
# |
15 |
# You should have received a copy of the GNU General Public License |
16 |
# along with this program; if not, write to the Free Software |
17 |
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA |
18 |
|
19 |
""" |
20 |
Inspect WMS Capabilities for later processing. |
21 |
|
22 |
Information should only be retrieved with the proper get*() methods. |
23 |
|
24 |
class WMSCapabilitiesParser: |
25 |
__init__() |
26 |
|
27 |
grok(string) |
28 |
|
29 |
getTitle() |
30 |
getAbstract() |
31 |
getFees() |
32 |
getAccessConstraints() |
33 |
getFormats() |
34 |
getLayers() |
35 |
getSRS() |
36 |
|
37 |
getLayerTitle(layer) |
38 |
getLayerSRS(layer) |
39 |
getLayerLatLonBBox(layer) |
40 |
getLayerBBox(layer, srs) |
41 |
|
42 |
isQueryable(layer) |
43 |
|
44 |
get_srs_discrepancies() |
45 |
""" |
46 |
|
47 |
__version__ = "$Revision$" |
48 |
# $Source$ |
49 |
# $Id$ |
50 |
|
51 |
import xml.dom.minidom |
52 |
from xml.dom import Node |
53 |
|
54 |
from domutils import getElementsByName, getElementByName |
55 |
|
56 |
from Thuban import _ |
57 |
|
58 |
class WMSCapabilitiesParser: |
59 |
""" |
60 |
Thuban class to parse capabilities supplied as large string. |
61 |
|
62 |
This class provides methods to parse and retrieve particular |
63 |
information from the WMS capabilities XML. Information should |
64 |
only be extracted by the respective get*() methods. |
65 |
""" |
66 |
|
67 |
layers = None |
68 |
title = None |
69 |
abstract = None |
70 |
fees = None |
71 |
access = None |
72 |
formats = None |
73 |
srs_discrepancies = None |
74 |
error = None |
75 |
|
76 |
|
77 |
def __init__(self): |
78 |
""" |
79 |
Initialises an instance in this class. |
80 |
""" |
81 |
|
82 |
# Note that we must not initialise internal variables of the |
83 |
# class in a mutable way or it will be shared among all |
84 |
# instances. None is immutable, [] is not. |
85 |
self.error = [] |
86 |
|
87 |
|
88 |
def grok(self, data): |
89 |
""" |
90 |
Parses the XML response to a WMS GetCapabilities request. |
91 |
|
92 |
Internal datastructure of the class will be filled. |
93 |
Information should only be retrieved with the respective |
94 |
get*() methods. |
95 |
""" |
96 |
xml_dom = xml.dom.minidom.parseString(data) |
97 |
root = xml_dom.documentElement |
98 |
|
99 |
# Extract the title |
100 |
foo = getElementByName(getElementByName(root, 'Service'), 'Title') |
101 |
if foo: |
102 |
self.title = foo.childNodes[0].data |
103 |
|
104 |
# Extract the abstract |
105 |
foo = getElementByName(getElementByName(root, 'Service'), 'Abstract') |
106 |
if foo and len(foo.childNodes[0].data): |
107 |
self.abstract = foo.childNodes[0].data |
108 |
|
109 |
# Extract fees information |
110 |
foo = getElementByName(getElementByName(root, 'Service'), 'Fees') |
111 |
if foo and len(foo.childNodes[0].data) \ |
112 |
and foo.childNodes[0].data.lower() != 'none': |
113 |
self.fees = foo.childNodes[0].data |
114 |
|
115 |
# Extract access information |
116 |
foo = getElementByName(getElementByName(root, 'Service'), |
117 |
'AccessConstraints') |
118 |
if foo and len(foo.childNodes[0].data) \ |
119 |
and foo.childNodes[0].data.lower() != 'none': |
120 |
self.access = foo.childNodes[0].data |
121 |
|
122 |
foo = getElementByName(getElementByName( |
123 |
root, 'Capability'), 'Request') |
124 |
|
125 |
# Need to distinguish between Map and GetMap for v1.0 and v1.1 |
126 |
bar = getElementByName(foo, 'GetMap') |
127 |
if bar: |
128 |
# WMS 1.1 |
129 |
foo = getElementsByName(bar, 'Format') |
130 |
self.formats = map((lambda i: i.childNodes[0].data), foo) |
131 |
else: |
132 |
# WMS 1.0 |
133 |
foo = getElementByName(getElementByName( |
134 |
foo, 'Map'), 'Format') |
135 |
for node in foo.childNodes: |
136 |
if node.nodeType == Node.ELEMENT_NODE: |
137 |
try: |
138 |
self.formats.append(node.nodeName) |
139 |
except AttributeError: |
140 |
self.formats = [node.nodeName] |
141 |
|
142 |
# Extract layer names |
143 |
self.layers = [] |
144 |
self.peekLayers(getElementByName(getElementByName( |
145 |
root, 'Capability'), 'Layer'), -1) |
146 |
|
147 |
xml_dom.unlink() |
148 |
|
149 |
|
150 |
def peekLayers(self, top, parent): |
151 |
""" |
152 |
Inspect the provided DOM fragment referenced as top. |
153 |
|
154 |
This method will inspect all included layers and traverse the |
155 |
tree recursively in order to fill the internal datastructure. |
156 |
|
157 |
Note that SRS other than EPSG:* are not yet supported, |
158 |
especially there is no support for AUTO and NONE. |
159 |
""" |
160 |
|
161 |
index = len (self.layers) |
162 |
self.layers.append({}) |
163 |
self.layers[index]['parent'] = parent |
164 |
|
165 |
for foo in top.attributes.keys(): |
166 |
if foo == 'queryable': |
167 |
self.layers[index]['queryable'] \ |
168 |
= int(top.attributes.get(foo).nodeValue) |
169 |
|
170 |
foo = getElementByName(top, 'Title') |
171 |
if foo and len(foo.childNodes[0].data): |
172 |
self.layers[index]['title'] = foo.childNodes[0].data |
173 |
else: |
174 |
# A <Title> is required for each layer, <name> is optional |
175 |
# See OGC 01-068r3, 7.1.4.5.1 and 7.1.4.5.2 |
176 |
self.error.append(_("No title found for layer #%d") % index) |
177 |
|
178 |
foo = getElementByName(top, 'Name') |
179 |
if foo and len(foo.childNodes[0].data): |
180 |
self.layers[index]['name'] = foo.childNodes[0].data |
181 |
|
182 |
# These values are only used for an integrity check |
183 |
for foo in getElementsByName(top, 'SRS'): |
184 |
for srs in foo.childNodes[0].data.split(" "): |
185 |
if srs[0:5] == 'EPSG:': |
186 |
srs = srs[5:] |
187 |
try: |
188 |
int(srs) |
189 |
try: |
190 |
self.layers[index]['srs'].append(srs) |
191 |
except KeyError: |
192 |
self.layers[index]['srs'] = [srs] |
193 |
except ValueError: |
194 |
if srs[0:4].upper() == 'AUTO' \ |
195 |
or srs[0:4].upper() == 'NONE': |
196 |
try: |
197 |
self.layers[index]['_srs_'].append(srs) |
198 |
except KeyError: |
199 |
self.layers[index]['_srs_'] = [srs] |
200 |
else: |
201 |
self.error.append(_("SRS '%s' is not numerical and not" |
202 |
" AUTO/NONE in layer '%s'") \ |
203 |
% (srs, self.layers[index]['title'])) |
204 |
|
205 |
foo = getElementByName(top, 'LatLonBoundingBox') |
206 |
if foo is not None: |
207 |
self.layers[index]['llbbox'] = {} |
208 |
for corner in ['minx', 'miny', 'maxx', 'maxy']: |
209 |
self.layers[index]['llbbox'][corner] \ |
210 |
= foo.attributes.get(corner).nodeValue |
211 |
|
212 |
boxes = getElementsByName(top, 'BoundingBox') |
213 |
if boxes != []: |
214 |
self.layers[index]['bbox'] = {} |
215 |
for foo in boxes: |
216 |
srs = foo.attributes.get('SRS').nodeValue |
217 |
if srs[0:5] == 'EPSG:': |
218 |
srs = srs[5:] |
219 |
self.layers[index]['bbox'][srs] = {} |
220 |
for corner in ['minx', 'miny', 'maxx', 'maxy']: |
221 |
self.layers[index]['bbox'][srs][corner] \ |
222 |
= foo.attributes.get(corner).nodeValue |
223 |
|
224 |
# Traverse subsidiary layers |
225 |
sublayer = getElementsByName(top, 'Layer') |
226 |
for l in sublayer: |
227 |
self.peekLayers(l, index) |
228 |
|
229 |
|
230 |
def getTitle(self): |
231 |
""" |
232 |
Returns the main title of the WMS object. |
233 |
|
234 |
If no title is provided in the capabilities, an empty string |
235 |
is returned. |
236 |
""" |
237 |
if self.title is None: |
238 |
return '' |
239 |
else: |
240 |
return self.title |
241 |
|
242 |
|
243 |
def getAbstract(self): |
244 |
""" |
245 |
Returns the abstract of the WMS object. |
246 |
|
247 |
If no abstract is provided in the capabilities, an empty |
248 |
string is returned. |
249 |
""" |
250 |
if self.abstract is None: |
251 |
return '' |
252 |
else: |
253 |
return self.abstract |
254 |
|
255 |
|
256 |
def getFees(self): |
257 |
""" |
258 |
Returns the fees information of the WMS object. |
259 |
|
260 |
If no information is provided in the capabilities or if it is |
261 |
set to 'none', an empty string is returned. |
262 |
""" |
263 |
if self.fees is None: |
264 |
return '' |
265 |
else: |
266 |
return self.fees |
267 |
|
268 |
|
269 |
def getAccessConstraints(self): |
270 |
""" |
271 |
Returns information about access constraints for the WMS object. |
272 |
|
273 |
If no information is provided in the capabilities or if it is |
274 |
set to 'none', an empty string is returned. |
275 |
""" |
276 |
if self.access is None: |
277 |
return '' |
278 |
else: |
279 |
return self.access |
280 |
|
281 |
|
282 |
def getFormats(self): |
283 |
""" |
284 |
Returns a list of supported output formats. |
285 |
|
286 |
These are used in the GetMap request. This method will |
287 |
default to 'image/jpeg' if no format is recognised in XML |
288 |
Capabilities, assuming that JPEG will always be supported on |
289 |
the server side. |
290 |
""" |
291 |
if self.formats is None: |
292 |
return ['image/jpeg'] |
293 |
else: |
294 |
return self.formats |
295 |
|
296 |
def getLayers(self): |
297 |
""" |
298 |
Returns a list of layer names. |
299 |
|
300 |
Only named layers will be returned, since a layer may have a |
301 |
title but doesn't have to have a name associated to it as |
302 |
well. If no layers were found, an empty list is returned. |
303 |
""" |
304 |
result = [] |
305 |
for layer in self.layers: |
306 |
if 'name' in layer: |
307 |
result.append(layer['name']) |
308 |
|
309 |
return result |
310 |
|
311 |
|
312 |
def getSRS(self): |
313 |
""" |
314 |
Returns the root list of spatial reference systems (SRS). |
315 |
|
316 |
This list is taken from the root layer. Those SRS are common |
317 |
to all subsidiary layers. If no SRS are common to all layers, |
318 |
an empty list is returned. If no layers were detected, an |
319 |
empty list is returned as well. |
320 |
""" |
321 |
if len(self.layers) == 0: |
322 |
return [] |
323 |
|
324 |
# index 0 denotes the root layer by design |
325 |
if 'srs' in self.layers[0].keys(): |
326 |
return self.layers[0]['srs'] |
327 |
|
328 |
|
329 |
def getLayerTitle(self, name): |
330 |
""" |
331 |
Returns the title of the named layer. |
332 |
|
333 |
If no such title or no such layer exists, an empty string is |
334 |
returned. |
335 |
""" |
336 |
for layer in self.layers: |
337 |
if 'name' in layer: |
338 |
if layer['name'] == name: |
339 |
if 'title' in layer: |
340 |
return layer['title'] |
341 |
|
342 |
return '' |
343 |
|
344 |
|
345 |
def getLayerSRS(self, name): |
346 |
""" |
347 |
Returns a list of spacial reference systems (SRS). |
348 |
|
349 |
The SRS are specified by the European Petroleum Survey Group |
350 |
(EPSG). There should be at least one SRS per layer, though. |
351 |
The prefix 'EPSG:' will be stripped. If none exists, an empty |
352 |
list is returned. |
353 |
|
354 |
The specification [OGC 01-068r3] says about inheritance of |
355 |
SRS: |
356 |
|
357 |
- Every layer is available in one or more SRS (or in an |
358 |
undefined SRS) |
359 |
|
360 |
- Geographic information whose SRS is undefined |
361 |
(e.g. digitised historical maps) shall use 'NONE' |
362 |
(case-sensitive). This implementation does not support |
363 |
this. |
364 |
|
365 |
- Every layer shall have at least one SRS element that is |
366 |
either stated explicitly or inherited from a parent layer. |
367 |
|
368 |
- The root layer element shall include a sequence of zero or |
369 |
more SRS elements listing all SRS which are common for to |
370 |
all subsidiary layers. |
371 |
|
372 |
- Layers may optionally add to the global SRS list, or to the |
373 |
list inherited from a parent layer. |
374 |
|
375 |
- A server which has the ability to transform data to |
376 |
different SRSes may choose not to provide an explicit |
377 |
BoundingBox for every possible SRS available for each Layer. |
378 |
Thus the list of <SRS> elements are authoritative. |
379 |
|
380 |
This implementation returns the list of SRS for the given |
381 |
layer, calculated by looking at BoundingBoxes defined in the |
382 |
named layer and all layers higher in the hierarchy up to the |
383 |
root layer which weren't overwritten. |
384 |
""" |
385 |
for i in range(len(self.layers)): |
386 |
if 'name' in self.layers[i]: |
387 |
if self.layers[i]['name'] == name: |
388 |
pivot = i |
389 |
break |
390 |
else: |
391 |
return [] |
392 |
|
393 |
result = [] |
394 |
while pivot != -1: |
395 |
if 'srs' in self.layers[pivot]: |
396 |
for srs in self.layers[pivot]['srs']: |
397 |
if srs not in result: |
398 |
result.append(srs) |
399 |
pivot = self.layers[pivot]['parent'] |
400 |
|
401 |
return result |
402 |
|
403 |
|
404 |
def getLayerLatLonBBox(self, name): |
405 |
""" |
406 |
Returns a dictionary of the LatLonBoundingBox. |
407 |
|
408 |
... for the named layer if it exists. The SRS is always |
409 |
EPSG:4326 per convention. There should be at least one, |
410 |
though, inherited from the root layer at least. If none |
411 |
exists, the value None is returned. |
412 |
""" |
413 |
for layer in self.layers: |
414 |
if 'name' in layer: |
415 |
if layer['name'] == name: |
416 |
if 'llbbox' in layer: |
417 |
return layer['llbbox'] |
418 |
else: |
419 |
# No LatLonBoundingBox found in current layer, |
420 |
# so traverse the hierarchy upwards and check |
421 |
# again until there is one. |
422 |
pivot = layer |
423 |
while pivot['parent'] != -1: |
424 |
pivot = self.layers[pivot['parent']] |
425 |
if 'llbbox' in pivot: |
426 |
return pivot['llbbox'] |
427 |
|
428 |
return None |
429 |
|
430 |
|
431 |
def getLayerBBox(self, name, srs): |
432 |
""" |
433 |
Returns a dictionary of the BoundingBox. |
434 |
|
435 |
If no such BoundingBox exists, None is returned. |
436 |
|
437 |
The specification [OGC 01-068r3] says about BoundingBoxes: |
438 |
|
439 |
- Layers may have zero or more BoundingBox elements what are |
440 |
either stated explicitly or inherited from a parent layer. |
441 |
|
442 |
- A layer may have multiple BoundingBox elements, but each one |
443 |
shall state a different SRS. |
444 |
|
445 |
- A layer inherits any BoundingBoxes defined by its |
446 |
parents. |
447 |
|
448 |
- A BoundingBox inherited from the parent layer for a |
449 |
particular SRS is replaced by any declaration for the same |
450 |
SRS in the current layer. |
451 |
|
452 |
- A BoundingBox in the child layer for a new SRS which is not |
453 |
already declared by the parent, is added to the list of |
454 |
BoundingBoxes for the child layer. |
455 |
|
456 |
- A single layer shall not contain more than one BoundingBox |
457 |
element for the same SRS. |
458 |
""" |
459 |
for layer in self.layers: |
460 |
if 'name' in layer: |
461 |
if layer['name'] == name: |
462 |
if 'bbox' in layer: |
463 |
if srs in layer['bbox']: |
464 |
return layer['bbox'][srs] |
465 |
|
466 |
# No BoundingBox for the given SRS found in |
467 |
# current layer, so traverse the hierarchy upwards |
468 |
# and check again until there is one. |
469 |
pivot = layer |
470 |
while pivot['parent'] != -1: |
471 |
pivot = self.layers[pivot['parent']] |
472 |
if 'bbox' in pivot: |
473 |
if srs in pivot['bbox']: |
474 |
return pivot['bbox'][srs] |
475 |
|
476 |
# No matching BBox found, let's see if it was EPSG:4326 |
477 |
if srs == '4326': |
478 |
return self.getLayerLatLonBBox(name) |
479 |
|
480 |
return None |
481 |
|
482 |
|
483 |
def isQueryable(self, name): |
484 |
""" |
485 |
Returns the value of the queryable attribute of a layer. |
486 |
|
487 |
This attribute denotes whether this layer can be queried |
488 |
through the GetFeatureInfo request. The default value is 0. |
489 |
|
490 |
The specification [OGC 01-068r3] this attribute can be |
491 |
inherited. |
492 |
""" |
493 |
|
494 |
for layer in self.layers: |
495 |
if 'name' in layer: |
496 |
if layer['name'] == name: |
497 |
try: |
498 |
return layer['queryable'] |
499 |
except KeyError: |
500 |
# No attribute in this layer, so traverse the |
501 |
# hierarchy upwards |
502 |
pivot = layer |
503 |
while pivot['parent'] != -1: |
504 |
pivot = self.layers[pivot['parent']] |
505 |
if 'queryable' in pivot: |
506 |
return pivot['queryable'] |
507 |
return 0 |
508 |
|
509 |
|
510 |
def get_srs_discrepancies(self): |
511 |
""" |
512 |
Returns a list of layer identifications where the denoted SRS |
513 |
values differ from the calculated ones through BoundingBox |
514 |
elements. |
515 |
""" |
516 |
return self.srs_discrepancies |
517 |
|
518 |
|
519 |
if __name__ == "__main__": |
520 |
print "This module cannot be executed standalone." |
521 |
|
522 |
import os |
523 |
|
524 |
sample = "test/sample.xml" |
525 |
try: |
526 |
f = open(sample, "r") |
527 |
except IOError: |
528 |
try: |
529 |
f = open(os.path.dirname(__file__) + "/" + sample, "r") |
530 |
except IOError: |
531 |
print "Cannot open %s for reading" % sample |
532 |
|
533 |
if f is not None: |
534 |
sample = f.read(); |
535 |
f.close() |
536 |
|
537 |
ina = WMSCapabilitiesParser() |
538 |
ina.grok(sample) |