def csw_global_dispatch(request):
"""pycsw wrapper"""
# this view should only operate if pycsw_local is the backend
# else, redirect to the URL of the non-pycsw_local backend
if settings.CATALOGUE['default'][
'ENGINE'] != 'geonode.catalogue.backends.pycsw_local':
return HttpResponseRedirect(settings.CATALOGUE['default']['URL'])
mdict = dict(settings.PYCSW['CONFIGURATION'], **CONFIGURATION)
env = request.META.copy()
env.update({
'local.app_root': os.path.dirname(__file__),
'REQUEST_URI': request.build_absolute_uri()
})
csw = server.Csw(mdict, env, version='2.0.2')
content = csw.dispatch_wsgi()
# pycsw 2.0 has an API break:
# pycsw < 2.0: content = xml_response
# pycsw >= 2.0: content = [http_status_code, content]
# deal with the API break
if isinstance(content, list): # pycsw 2.0+
content = content[1]
return HttpResponse(content, content_type=csw.contenttype)
def csw_view(request, catalog=None):
"""CSW dispatch view.
Wraps the WSGI call and allows us to tweak any django settings.
"""
env = request.META.copy()
env.update({
'local.app_root': os.path.dirname(__file__),
'REQUEST_URI': request.build_absolute_uri()
})
# pycsw prefers absolute urls, let's get them from the request.
url = request.build_absolute_uri()
PYCSW['server']['url'] = url
PYCSW['metadata:main']['provider_url'] = url
csw = server.Csw(PYCSW, env)
status, content = csw.dispatch_wsgi()
status_code = int(status[0:3])
response = HttpResponse(
content,
content_type=csw.contenttype,
status=status_code,
)
return response
def application(env, start_response):
"""WSGI wrapper"""
print(env['PATH_INFO'])
pycsw_root = get_pycsw_root_path(os.environ, env)
configuration_path = get_configuration_path(os.environ, env, pycsw_root)
env['local.app_root'] = pycsw_root
if 'HTTP_HOST' in env and ':' in env['HTTP_HOST']:
env['HTTP_HOST'] = env['HTTP_HOST'].split(':')[0]
csw = server.Csw(configuration_path, env)
status, contents = csw.dispatch_wsgi()
headers = {
'Content-Length': str(len(contents)),
'Content-Type': str(csw.contenttype)
}
if "gzip" in env.get("HTTP_ACCEPT_ENCODING", ""):
try:
compression_level = int(
csw.config.get("server", "gzip_compresslevel"))
contents, compress_headers = compress_response(
contents, compression_level)
headers.update(compress_headers)
except configparser.NoOptionError:
print(
"The client requested a gzip compressed response. However, "
"the server does not specify the 'gzip_compresslevel' option. "
"Returning an uncompressed response...")
except configparser.NoSectionError:
print('Could not load user configuration %s' % configuration_path)
start_response(status, list(headers.items()))
return [contents]
def csw_global_dispatch_by_catalog(request, catalog_slug):
"""pycsw wrapper"""
catalog = get_object_or_404(Catalog, slug=catalog_slug)
# TODO: Implement pycsw per catalog
if catalog:
pass
env = request.META.copy()
env.update({'local.app_root': os.path.dirname(__file__),
'REQUEST_URI': request.build_absolute_uri()})
csw = server.Csw(settings.PYCSW, env, version='2.0.2')
content = csw.dispatch_wsgi()
# pycsw 2.0 has an API break:
# pycsw < 2.0: content = xml_response
# pycsw >= 2.0: content = [http_status_code, content]
# deal with the API break
if isinstance(content, list): # pycsw 2.0+
content = content[1]
response = HttpResponse(content, content_type=csw.contenttype)
response['Access-Control-Allow-Origin'] = '*'
return response
def csw(request):
"""CSW WSGI wrapper"""
# serialize settings.CSW into SafeConfigParser
# object for interaction with pycsw
mdict = dict(settings.CSW, **CONFIGURATION)
# update server.url
server_url = '%s://%s%s' % \
(request.META['wsgi.url_scheme'],
request.META['HTTP_HOST'],
request.META['PATH_INFO'])
mdict['server']['url'] = server_url
env = request.META.copy()
env.update({
'local.app_root': os.path.dirname(__file__),
'REQUEST_URI': request.build_absolute_uri(),
})
csw = server.Csw(mdict, env)
content = csw.dispatch_wsgi()
return HttpResponse(content, content_type=csw.contenttype)
def csw(request):
"""CSW WSGI wrapper"""
# serialize settings.CSW into SafeConfigParser
# object for interaction with pycsw
mdict = dict(settings.CSW, **CONFIGURATION)
# TODO: pass just dict when pycsw supports it
config = SafeConfigParser()
for section, options in mdict.iteritems():
config.add_section(section)
for k, v in options.iteritems():
config.set(section, k, v)
# update server.url
server_url = '%s://%s%s' % \
(request.META['wsgi.url_scheme'],
request.META['HTTP_HOST'],
request.META['PATH_INFO'])
config.set('server', 'url', server_url)
env = request.META.copy()
env.update({
'local.app_root': os.path.dirname(__file__),
'REQUEST_URI': request.build_absolute_uri(),
})
csw = server.Csw(config, env)
content = csw.dispatch_wsgi()
return HttpResponse(content, content_type=csw.contenttype)
def csw_global_dispatch(request):
"""pycsw wrapper"""
# this view should only operate if pycsw_local is the backend
# else, redirect to the URL of the non-pycsw_local backend
if (settings.CATALOGUE['default']['ENGINE'] !=
'geonode.catalogue.backends.pycsw_local'):
return HttpResponseRedirect(settings.CATALOGUE['default']['URL'])
# serialize pycsw settings into SafeConfigParser
# object for interaction with pycsw
# TODO: pass just dict when pycsw supports it
mdict = dict(settings.PYCSW['CONFIGURATION'], **CONFIGURATION)
config = SafeConfigParser()
for section, options in mdict.iteritems():
config.add_section(section)
for option, value in options.iteritems():
config.set(section, option, value)
env = request.META.copy()
env.update({
'local.app_root': os.path.dirname(__file__),
'REQUEST_URI': request.build_absolute_uri(),
})
csw = server.Csw(config, env)
content = csw.dispatch_wsgi()
return HttpResponse(content, content_type=csw.contenttype)
def application(env, start_response):
"""WSGI wrapper"""
config = 'default.cfg'
if 'PYCSW_CONFIG' in env:
config = env['PYCSW_CONFIG']
if env['QUERY_STRING'].lower().find('config') != -1:
for kvp in env['QUERY_STRING'].split('&'):
if kvp.lower().find('config') != -1:
config = kvp.split('=')[1]
if not os.path.isabs(config):
config = os.path.join(app_path, config)
if 'HTTP_HOST' in env and ':' in env['HTTP_HOST']:
env['HTTP_HOST'] = env['HTTP_HOST'].split(':')[0]
env['local.app_root'] = app_path
csw = server.Csw(config, env)
gzip = False
if ('HTTP_ACCEPT_ENCODING' in env
and env['HTTP_ACCEPT_ENCODING'].find('gzip') != -1):
# set for gzip compressed response
gzip = True
# set compression level
if csw.config.has_option('server', 'gzip_compresslevel'):
gzip_compresslevel = \
int(csw.config.get('server', 'gzip_compresslevel'))
else:
gzip_compresslevel = 0
contents = csw.dispatch_wsgi()
headers = {}
if gzip and gzip_compresslevel > 0:
import gzip
buf = StringIO()
gzipfile = gzip.GzipFile(mode='wb',
fileobj=buf,
compresslevel=gzip_compresslevel)
gzipfile.write(contents)
gzipfile.close()
contents = buf.getvalue()
headers['Content-Encoding'] = 'gzip'
headers['Content-Length'] = str(len(contents))
headers['Content-Type'] = csw.contenttype
status = '200 OK'
start_response(status, headers.items())
return [contents]
def test_suites(test_identifier, use_xml_canonicalisation,
save_results_directory, configuration, request_method,
request_data, expected_result, normalize_identifier_fields):
"""Test suites.
This function is automatically parametrized by pytest as a result of the
``conftest:pytest_generate_tests`` function. The input parameters are thus
supplied by pytest as a result of discovering and parsing the existing
test suites located under ``tests/functionaltests/suites``.
Parameters
----------
configuration: SafeConfigParser
The configuration to use with the pycsw server instance under test
request_method: str
The HTTP method of the request. Either GET or POST.
request_data: str
Either the path to the request file, for POST requests, or the request
parameters, for GET requests
expected_result: str
Path to the file that holds the expected result
normalize_identifier_fields: bool
Whether to normalize the identifier fields in responses. This
parameter is used only in the 'harvesting' and 'manager' suites
use_xml_canonicalisation: bool
Whether to compare results with their expected values by using xml
canonicalisation or simply by doing a diff.
save_results_directory: str
Path to a directory where to test results should be saved to. A value
of ``None`` (the default) means that results will not get saved to
disk.
"""
request_environment = _prepare_wsgi_test_environment(
request_method, request_data)
pycsw_server = server.Csw(rtconfig=configuration, env=request_environment)
encoding = "utf-8"
status, raw_contents = pycsw_server.dispatch_wsgi()
contents = raw_contents.decode(encoding)
with codecs.open(expected_result, encoding=encoding) as fh:
expected = fh.read()
normalized_result = _normalize(
contents, normalize_identifiers=normalize_identifier_fields)
if use_xml_canonicalisation:
print("Comparing results using XML canonicalization...")
matches_expected = _compare_with_xml_canonicalisation(
normalized_result, expected)
else:
print("Comparing results using diffs...")
matches_expected = _compare_without_xml_canonicalisation(
normalized_result, expected)
if not matches_expected and use_xml_canonicalisation:
print("expected: {0}".format(expected))
print("response: {0}".format(normalized_result))
if save_results_directory is not None:
_save_test_result(save_results_directory, normalized_result,
test_identifier, encoding)
assert matches_expected
def csw_global_dispatch(request, url=None, catalog_id=None):
"""pycsw wrapper"""
msg = None
# test for authentication and authorization
if any(word in request.body for word in ['Harvest ', 'Transaction ']):
if not _is_authenticated():
msg = 'Not authenticated'
if not _is_authorized():
msg = 'Not authorized'
if msg is not None:
template = loader.get_template('search/csw-2.0.2-exception.xml')
context = RequestContext(request, {'exception_text': msg})
return HttpResponseForbidden(template.render(context),
content_type='application/xml')
env = request.META.copy()
# TODO: remove this workaround
# HH should be able to pass env['wsgi.input'] without hanging
# details at https://github.com/cga-harvard/HHypermap/issues/94
if request.method == 'POST':
from StringIO import StringIO
env['wsgi.input'] = StringIO(request.body)
env.update({
'local.app_root': os.path.dirname(__file__),
'REQUEST_URI': request.build_absolute_uri()
})
# if this is a catalog based CSW, then update settings
if url is not None:
settings.REGISTRY_PYCSW['server']['url'] = url
if catalog_id is not None:
settings.REGISTRY_PYCSW['repository'][
'filter'] = 'catalog_id = %d' % catalog_id
csw = server.Csw(settings.REGISTRY_PYCSW, env)
content = csw.dispatch_wsgi()
# pycsw 2.0 has an API break:
# pycsw < 2.0: content = xml_response
# pycsw >= 2.0: content = [http_status_code, content]
# deal with the API break
if isinstance(content, list): # pycsw 2.0+
content = content[1]
response = HttpResponse(content, content_type=csw.contenttype)
# TODO: Fix before 1.0 release. CORS should not be enabled blindly like this.
response['Access-Control-Allow-Origin'] = '*'
return response
def _csw_local_dispatch(self, keywords=None, start=0, limit=10, bbox=None, identifier=None):
"""
HTTP-less CSW
"""
# serialize pycsw settings into SafeConfigParser
# object for interaction with pycsw
mdict = dict(settings.PYCSW['CONFIGURATION'], **CONFIGURATION)
config = SafeConfigParser()
for section, options in mdict.iteritems():
config.add_section(section)
for option, value in options.iteritems():
config.set(section, option, value)
# fake HTTP environment variable
os.environ['QUERY_STRING'] = ''
# init pycsw
csw = server.Csw(config)
# fake HTTP method
csw.requesttype = 'POST'
# fake HTTP request parameters
if identifier is None: # it's a GetRecords request
formats = []
for f in self.catalogue.formats:
formats.append(METADATA_FORMATS[f][0])
csw.kvp = {
'elementsetname': 'full',
'typenames': formats,
'resulttype': 'results',
'constraintlanguage': 'CQL_TEXT',
'constraint': 'csw:AnyText like "%%%s%%"' % keywords,
'outputschema': 'http://www.isotc211.org/2005/gmd',
'constraint': None,
'startposition': start,
'maxrecords': limit
}
response = csw.getrecords()
else: # it's a GetRecordById request
csw.kvp = {
'id': [identifier],
'outputschema': 'http://www.isotc211.org/2005/gmd',
}
# FIXME(Ariel): Remove this try/except block when pycsw deals with
# empty geometry fields better.
# https://gist.github.com/ingenieroariel/717bb720a201030e9b3a
try:
response = csw.getrecordbyid()
except ReadingError:
return []
return etree.tostring(response)
def test_suites(configuration, request_method, request_data, expected_result,
normalize_identifier_fields):
"""Test suites.
This function is automatically parametrized by pytest as a result of the
``conftest:pytest_generate_tests`` function. The input parameters are thus
supplied by pytest as a result of discovering and parsing the existing
test suites located under ``tests/functionaltests/suites``.
Parameters
----------
configuration: SafeConfigParser
The configuration to use with the pycsw server instance under test
request_method: str
The HTTP method of the request. Either GET or POST.
request_data: str
Either the path to the request file, for POST requests, or the request
parameters, for GET requests
expected_result: str
Path to the file that holds the expected result
normalize_identifier_fields: bool
Whether to normalize the identifier fields in responses. This
parameter is used only in the 'harvesting' and 'manager' suites
"""
request_environment = _prepare_wsgi_test_environment(request_method,
request_data)
pycsw_server = server.Csw(rtconfig=configuration, env=request_environment)
encoding = "utf-8"
status, raw_contents = pycsw_server.dispatch_wsgi()
contents = raw_contents.decode(encoding)
with codecs.open(expected_result, encoding=encoding) as fh:
expected = fh.read()
normalized_result = _normalize(
contents,
normalize_identifiers=normalize_identifier_fields
)
try:
matches_expected = _test_xml_result(normalized_result, expected)
except etree.XMLSyntaxError:
# the file is either not XML (perhaps JSON?) or malformed
matches_expected = _test_json_result(normalized_result, expected)
except etree.C14NError:
print("XML canonicalization has failed. Trying to compare result "
"with expected using difflib")
matches_expected = _test_xml_diff(normalized_result, expected)
if not matches_expected:
#print("expected: {0}".format(expected.encode(encoding)))
#print("response: {0}".format(normalized_result.encode(encoding)))
print("expected: {0}".format(expected))
print("response: {0}".format(normalized_result))
assert matches_expected
def post_response(self, request_body):
"""
Utility function to retrieve POST response from PyCSW
Keyword Parameters:
request_body -- Bytes representing the body of the HTTP POST
Exceptions:
TempRepoLocked -- raised when PyCSW repo is under construction
by another thread or process & is not available for this call.
"""
pycsw_config = self.get_config()
self._build_repo_if_missing()
csw = server.Csw(pycsw_config, self.env)
csw.requesttype = 'POST'
csw.request = request_body
return csw.dispatch()
def csw_global_dispatch(request, url=None, catalog_id=None):
"""pycsw wrapper"""
if request.user.is_authenticated(): # turn on CSW-T
settings.REGISTRY_PYCSW['manager']['transactions'] = 'true'
env = request.META.copy()
# TODO: remove this workaround
# HH should be able to pass env['wsgi.input'] without hanging
# details at https://github.com/cga-harvard/HHypermap/issues/94
if request.method == 'POST':
from StringIO import StringIO
env['wsgi.input'] = StringIO(request.body)
env.update({
'local.app_root': os.path.dirname(__file__),
'REQUEST_URI': request.build_absolute_uri()
})
# if this is a catalog based CSW, then update settings
if url is not None:
settings.REGISTRY_PYCSW['server']['url'] = url
if catalog_id is not None:
settings.REGISTRY_PYCSW['repository'][
'filter'] = 'catalog_id = %d' % catalog_id
csw = server.Csw(settings.REGISTRY_PYCSW, env)
content = csw.dispatch_wsgi()
# pycsw 2.0 has an API break:
# pycsw < 2.0: content = xml_response
# pycsw >= 2.0: content = [http_status_code, content]
# deal with the API break
if isinstance(content, list): # pycsw 2.0+
content = content[1]
response = HttpResponse(content, content_type=csw.contenttype)
# TODO: Fix before 1.0 release. CORS should not be enabled blindly like this.
response['Access-Control-Allow-Origin'] = '*'
return response
def csw_global_dispatch(request):
"""pycsw wrapper"""
# this view should only operate if pycsw_local is the backend
# else, redirect to the URL of the non-pycsw_local backend
if settings.CATALOGUE['default']['ENGINE'] != 'geonode.catalogue.backends.pycsw_local':
return HttpResponseRedirect(settings.CATALOGUE['default']['URL'])
mdict = dict(settings.PYCSW['CONFIGURATION'], **CONFIGURATION)
env = request.META.copy()
env.update({'local.app_root': os.path.dirname(__file__),
'REQUEST_URI': request.build_absolute_uri()})
csw = server.Csw(mdict, env)
content = csw.dispatch_wsgi()
return HttpResponse(content, content_type=csw.contenttype)
def csw_view(request, catalog=None):
"""CSW dispatch view.
Wraps the WSGI call and allows us to tweak any django settings.
"""
# PUT creates catalog, DELETE removes catalog.
# GET and POST are managed by pycsw.
if catalog and request.META['REQUEST_METHOD'] == 'PUT':
message = create_index(catalog)
return HttpResponse(message, status=200)
if catalog and request.META['REQUEST_METHOD'] == 'DELETE':
message, status = delete_index(catalog)
return HttpResponse(message, status=status)
env = request.META.copy()
env.update({
'local.app_root': os.path.dirname(__file__),
'REQUEST_URI': request.build_absolute_uri()
})
# pycsw prefers absolute urls, let's get them from the request.
url = request.build_absolute_uri()
PYCSW['server']['url'] = url
PYCSW['metadata:main']['provider_url'] = url
# Enable CSW-T when a catalog is defined in the
if catalog:
PYCSW['manager']['transactions'] = 'true'
csw = server.Csw(PYCSW, env)
status, content = csw.dispatch_wsgi()
status_code = int(status[0:3])
response = HttpResponse(
content,
content_type=csw.contenttype,
status=status_code,
)
return response
def get_response(self, request_url, query_string_parameters):
"""
Utility function to retrieve GET response from PyCSW
Keyword Parameters:
request_url -- String representation of the full-URL for this
HTTP GET request
query_string_parameters -- Dictionary, representing the
'request_url' query string key/value pairs
Exceptions:
TempRepoLocked -- raised when PyCSW repo is under construction
by another thread or process & is not available for this call.
"""
pycsw_config = self.get_config()
self._build_repo_if_missing()
csw = server.Csw(pycsw_config, self.env)
csw.requesttype = 'GET'
csw.request = request_url
csw.kvp = query_string_parameters
return csw.dispatch()
def csw_global_dispatch(request):
"""pycsw wrapper"""
msg = None
# test for authentication and authorization
if any(word in request.body for word in ['Harvest ', 'Transaction ']):
if not _is_authenticated():
msg = 'Not authenticated'
if not _is_authorized():
msg = 'Not authorized'
if msg is not None:
template = loader.get_template('search/csw-2.0.2-exception.xml')
context = RequestContext(request, {
'exception_text': msg
})
return HttpResponseForbidden(template.render(context), content_type='application/xml')
env = request.META.copy()
env.update({'local.app_root': os.path.dirname(__file__),
'REQUEST_URI': request.build_absolute_uri()})
csw = server.Csw(settings.PYCSW, env, version='2.0.2')
content = csw.dispatch_wsgi()
# pycsw 2.0 has an API break:
# pycsw < 2.0: content = xml_response
# pycsw >= 2.0: content = [http_status_code, content]
# deal with the API break
if isinstance(content, list): # pycsw 2.0+
content = content[1]
response = HttpResponse(content, content_type=csw.contenttype)
response['Access-Control-Allow-Origin'] = '*'
return response
def csw(self, *args, **kwargs):
"""Wrapper around pycsw for dispatching CSW requests"""
# Cycle through the CONFIGURATION dictionary and push the params into the config object
config = SafeConfigParser()
for section, options in CONFIGURATION.iteritems():
config.add_section(section)
for k, v in options.iteritems():
config.set(section, k, v)
# Calculate and insert the server URL into the config
server_url = 'http://%s%s' % \
(kwargs["environ"]['HTTP_HOST'],
kwargs["environ"]['PATH_INFO'])
config.set('server', 'url', server_url)
# Make a copy of the WSGI request environment and add parameters
env = kwargs["environ"].copy()
query = kwargs["environ"]["QUERY_STRING"]
if query != "":
absolute_uri = "%s?%s" % (server_url, query)
else:
absolute_uri = server_url
env.update({
'local.app_root': os.path.dirname(__file__),
'REQUEST_URI': absolute_uri
})
# Create an instance of pycsw's CSW class to handle the request
csw = server.Csw(config, env)
# Run the request
content = csw.dispatch_wsgi()
# Set the response Content-type, and return the result
kwargs["pylons"].response.content_type = csw.contenttype
return content
def csw_global_dispatch(request):
"""pycsw wrapper"""
# this view should only operate if pycsw_local is the backend
# else, redirect to the URL of the non-pycsw_local backend
if settings.CATALOGUE['default']['ENGINE'] != 'geonode.catalogue.backends.pycsw_local':
return HttpResponseRedirect(settings.CATALOGUE['default']['URL'])
mdict = dict(settings.PYCSW['CONFIGURATION'], **CONFIGURATION)
access_token = None
if request and request.user:
access_token = get_or_create_token(request.user)
if access_token and access_token.is_expired():
access_token = None
absolute_uri = ('%s' % request.build_absolute_uri())
query_string = ('%s' % request.META['QUERY_STRING'])
env = request.META.copy()
if access_token and not access_token.is_expired():
env.update({'access_token': access_token.token})
if 'access_token' not in query_string:
absolute_uri = ('%s&access_token=%s' % (absolute_uri, access_token.token))
query_string = ('%s&access_token=%s' % (query_string, access_token.token))
env.update({'local.app_root': os.path.dirname(__file__),
'REQUEST_URI': absolute_uri,
'QUERY_STRING': query_string})
# Save original filter before doing anything
mdict_filter = mdict['repository']['filter']
try:
# Filter out Layers not accessible to the User
authorized_ids = []
if request.user:
profiles = Profile.objects.filter(username=str(request.user))
else:
profiles = Profile.objects.filter(username="******")
if profiles:
authorized = list(
get_objects_for_user(
profiles[0],
'base.view_resourcebase').values('id'))
layers = ResourceBase.objects.filter(
id__in=[d['id'] for d in authorized])
if layers:
authorized_ids = [d['id'] for d in authorized]
if len(authorized_ids) > 0:
authorized_layers = "(" + (", ".join(str(e)
for e in authorized_ids)) + ")"
authorized_layers_filter = "id IN " + authorized_layers
mdict['repository']['filter'] += " AND " + authorized_layers_filter
if request.user and request.user.is_authenticated:
mdict['repository']['filter'] = "({}) OR ({})".format(mdict['repository']['filter'],
authorized_layers_filter)
else:
authorized_layers_filter = "id = -9999"
mdict['repository']['filter'] += " AND " + authorized_layers_filter
# Filter out Documents and Maps
if 'ALTERNATES_ONLY' in settings.CATALOGUE['default'] and settings.CATALOGUE['default']['ALTERNATES_ONLY']:
mdict['repository']['filter'] += " AND alternate IS NOT NULL"
# Filter out Layers belonging to specific Groups
is_admin = False
if request.user:
is_admin = request.user.is_superuser if request.user else False
if not is_admin and settings.GROUP_PRIVATE_RESOURCES:
groups_ids = []
if request.user and request.user.is_authenticated:
for group in request.user.groups.all():
groups_ids.append(group.id)
group_list_all = []
try:
group_list_all = request.user.group_list_all().values('group')
except BaseException:
pass
for group in group_list_all:
if isinstance(group, dict):
if 'group' in group:
groups_ids.append(group['group'])
else:
groups_ids.append(group.id)
public_groups = GroupProfile.objects.exclude(
access="private").values('group')
for group in public_groups:
if isinstance(group, dict):
if 'group' in group:
groups_ids.append(group['group'])
else:
groups_ids.append(group.id)
if len(groups_ids) > 0:
groups = "(" + (", ".join(str(e) for e in groups_ids)) + ")"
groups_filter = "(group_id IS NULL OR group_id IN " + groups + ")"
mdict['repository']['filter'] += " AND " + groups_filter
else:
groups_filter = "group_id IS NULL"
mdict['repository']['filter'] += " AND " + groups_filter
csw = server.Csw(mdict, env, version='2.0.2')
content = csw.dispatch_wsgi()
# pycsw 2.0 has an API break:
# pycsw < 2.0: content = xml_response
# pycsw >= 2.0: content = [http_status_code, content]
# deal with the API break
if isinstance(content, list): # pycsw 2.0+
content = content[1]
spaces = {'csw': 'http://www.opengis.net/cat/csw/2.0.2',
'dc': 'http://purl.org/dc/elements/1.1/',
'dct': 'http://purl.org/dc/terms/',
'gmd': 'http://www.isotc211.org/2005/gmd',
'gml': 'http://www.opengis.net/gml',
'ows': 'http://www.opengis.net/ows',
'xs': 'http://www.w3.org/2001/XMLSchema',
'xsi': 'http://www.w3.org/2001/XMLSchema-instance',
'ogc': 'http://www.opengis.net/ogc',
'gco': 'http://www.isotc211.org/2005/gco',
'gmi': 'http://www.isotc211.org/2005/gmi'}
for prefix, uri in spaces.items():
ET.register_namespace(prefix, uri)
if access_token and not access_token.is_expired():
tree = dlxml.fromstring(content)
for online_resource in tree.findall(
'*//gmd:CI_OnlineResource', spaces):
try:
linkage = online_resource.find('gmd:linkage', spaces)
for url in linkage.findall('gmd:URL', spaces):
if url.text:
if '?' not in url.text:
url.text += "?"
else:
url.text += "&"
url.text += ("access_token=%s" % (access_token.token))
url.set('updated', 'yes')
except BaseException:
pass
content = ET.tostring(tree, encoding='utf8', method='xml')
finally:
# Restore original filter before doing anything
mdict['repository']['filter'] = mdict_filter
return HttpResponse(content, content_type=csw.contenttype)
def csw_global_dispatch(request, layer_filter=None, config_updater=None):
"""pycsw wrapper"""
# this view should only operate if pycsw_local is the backend
# else, redirect to the URL of the non-pycsw_local backend
if settings.CATALOGUE['default'][
'ENGINE'] != 'geonode.catalogue.backends.pycsw_local':
return HttpResponseRedirect(settings.CATALOGUE['default']['URL'])
mdict = dict(settings.PYCSW['CONFIGURATION'], **CONFIGURATION)
mdict = config_updater(mdict) if config_updater else mdict
access_token = None
if request and request.user:
access_token = get_or_create_token(request.user)
if access_token and access_token.is_expired():
access_token = None
absolute_uri = f'{request.build_absolute_uri()}'
query_string = f"{request.META['QUERY_STRING']}"
env = request.META.copy()
if access_token and not access_token.is_expired():
env.update({'access_token': access_token.token})
if 'access_token' not in query_string:
absolute_uri = f'{absolute_uri}&access_token={access_token.token}'
query_string = f'{query_string}&access_token={access_token.token}'
env.update({
'local.app_root': os.path.dirname(__file__),
'REQUEST_URI': absolute_uri,
'QUERY_STRING': query_string
})
# Save original filter before doing anything
mdict_filter = mdict['repository']['filter']
try:
# Filter out Layers not accessible to the User
authorized_ids = []
if request.user:
profiles = get_user_model().objects.filter(
username=str(request.user))
else:
profiles = get_user_model().objects.filter(
username="******")
if profiles:
authorized = list(
get_objects_for_user(profiles[0],
'base.view_resourcebase').values('id'))
layers = ResourceBase.objects.filter(
id__in=[d['id'] for d in authorized])
if layer_filter and layers:
layers = layer_filter(layers)
if layers:
authorized_ids = [d.id for d in layers]
if len(authorized_ids) > 0:
authorized_layers = "(" + (", ".join(
str(e) for e in authorized_ids)) + ")"
authorized_layers_filter = "id IN " + authorized_layers
mdict['repository']['filter'] += " AND " + authorized_layers_filter
if request.user and request.user.is_authenticated:
mdict['repository'][
'filter'] = f"({mdict['repository']['filter']}) OR ({authorized_layers_filter})"
else:
authorized_layers_filter = "id = -9999"
mdict['repository']['filter'] += " AND " + authorized_layers_filter
# Filter out Documents and Maps
if 'ALTERNATES_ONLY' in settings.CATALOGUE[
'default'] and settings.CATALOGUE['default']['ALTERNATES_ONLY']:
mdict['repository']['filter'] += " AND alternate IS NOT NULL"
# Filter out Layers belonging to specific Groups
is_admin = False
if request.user:
is_admin = request.user.is_superuser if request.user else False
if not is_admin and settings.GROUP_PRIVATE_RESOURCES:
groups_ids = []
if request.user and request.user.is_authenticated:
for group in request.user.groups.all():
groups_ids.append(group.id)
group_list_all = []
try:
group_list_all = request.user.group_list_all().values(
'group')
except Exception:
pass
for group in group_list_all:
if isinstance(group, dict):
if 'group' in group:
groups_ids.append(group['group'])
else:
groups_ids.append(group.id)
public_groups = GroupProfile.objects.exclude(
access="private").values('group')
for group in public_groups:
if isinstance(group, dict):
if 'group' in group:
groups_ids.append(group['group'])
else:
groups_ids.append(group.id)
if len(groups_ids) > 0:
groups = "(" + (", ".join(str(e) for e in groups_ids)) + ")"
groups_filter = "(group_id IS NULL OR group_id IN " + groups + ")"
mdict['repository']['filter'] += " AND " + groups_filter
else:
groups_filter = "group_id IS NULL"
mdict['repository']['filter'] += " AND " + groups_filter
csw = server.Csw(mdict, env, version='2.0.2')
content = csw.dispatch_wsgi()
# pycsw 2.0 has an API break:
# - pycsw < 2.0: content = xml_response
# - pycsw >= 2.0: content = [http_status_code, content]
# deal with the API break
if isinstance(content, list): # pycsw 2.0+
content = content[1]
finally:
# Restore original filter before doing anything
mdict['repository']['filter'] = mdict_filter
return HttpResponse(content, content_type=csw.contenttype)
GZIP = False
if os.environ.has_key('PYCSW_CONFIG'):
CONFIG = os.environ['PYCSW_CONFIG']
if os.environ['QUERY_STRING'].lower().find('config') != -1:
for kvp in os.environ['QUERY_STRING'].split('&'):
if kvp.lower().find('config') != -1:
CONFIG = kvp.split('=')[1]
if (os.environ.has_key('HTTP_ACCEPT_ENCODING')
and os.environ['HTTP_ACCEPT_ENCODING'].find('gzip') != -1):
# set for gzip compressed response
GZIP = True
# get runtime configuration
CSW = server.Csw(CONFIG)
# set compression level
if CSW.config.has_option('server', 'gzip_compresslevel'):
GZIP_COMPRESSLEVEL = \
int(CSW.config.get('server', 'gzip_compresslevel'))
else:
GZIP_COMPRESSLEVEL = 0
# go!
OUTP = CSW.dispatch_cgi()
sys.stdout.write("Content-Type:%s\r\n" % CSW.contenttype)
if GZIP and GZIP_COMPRESSLEVEL > 0:
import gzip
def _csw_local_dispatch(self,
keywords=None,
start=0,
limit=10,
bbox=None,
identifier=None):
"""
HTTP-less CSW
"""
# serialize pycsw settings into SafeConfigParser
# object for interaction with pycsw
mdict = dict(settings.PYCSW['CONFIGURATION'], **CONFIGURATION)
if 'server' in settings.PYCSW['CONFIGURATION']:
# override server system defaults with user specified directives
mdict['server'].update(settings.PYCSW['CONFIGURATION']['server'])
config = SafeConfigParser()
for section, options in mdict.iteritems():
config.add_section(section)
for option, value in options.iteritems():
config.set(section, option, value)
# fake HTTP environment variable
os.environ['QUERY_STRING'] = ''
# init pycsw
csw = server.Csw(config, version='2.0.2')
# fake HTTP method
csw.requesttype = 'GET'
# fake HTTP request parameters
if identifier is None: # it's a GetRecords request
formats = []
for f in self.catalogue.formats:
formats.append(METADATA_FORMATS[f][0])
csw.kvp = {
'service': 'CSW',
'version': '2.0.2',
'elementsetname': 'full',
'typenames': formats,
'resulttype': 'results',
'constraintlanguage': 'CQL_TEXT',
'outputschema': 'http://www.isotc211.org/2005/gmd',
'constraint': None,
'startposition': start,
'maxrecords': limit
}
response = csw.getrecords()
else: # it's a GetRecordById request
csw.kvp = {
'service': 'CSW',
'version': '2.0.2',
'request': 'GetRecordById',
'id': identifier,
'outputschema': 'http://www.isotc211.org/2005/gmd',
}
# FIXME(Ariel): Remove this try/except block when pycsw deals with
# empty geometry fields better.
# https://gist.github.com/ingenieroariel/717bb720a201030e9b3a
try:
response = csw.dispatch()
except ReadingError:
return []
if isinstance(response, list): # pycsw 2.0+
response = response[1]
return response
