passerelle/passerelle/apps/opengis/models.py

708 lines
28 KiB
Python

# passerelle - uniform access to multiple data sources and services
# Copyright (C) 2017 Entr'ouvert
#
# This program is free software: you can redistribute it and/or modify it
# under the terms of the GNU Affero General Public License as published
# by the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU Affero General Public License for more details.
#
# You should have received a copy of the GNU Affero General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
import html
import math
import xml.etree.ElementTree as ET
import pyproj
from django.contrib.postgres.fields import JSONField
from django.core.cache import cache
from django.db import models, transaction
from django.db.models import Q
from django.http import HttpResponse
from django.shortcuts import get_object_or_404
from django.template import Context, Template, TemplateSyntaxError
from django.urls import reverse
from django.utils.text import slugify
from django.utils.translation import gettext_lazy as _
from passerelle.base.models import BaseQuery, BaseResource
from passerelle.utils.api import endpoint
from passerelle.utils.conversion import num2deg, simplify
from passerelle.utils.jsonresponse import APIError
from passerelle.utils.templates import validate_template
def build_dict_from_xml(elem):
d = {}
for child in elem.find('.'):
if child.tag.startswith('{'):
continue
attribute_name = slugify(child.tag).replace('-', '_')
if child.text and child.text.strip():
d[attribute_name] = child.text.strip()
else:
d[attribute_name] = build_dict_from_xml(child)
return d
PROJECTIONS = (
('EPSG:2154', _('EPSG:2154 (Lambert-93)')),
('EPSG:3857', _('EPSG:3857 (WGS 84 / Pseudo-Mercator)')),
('EPSG:3945', _('EPSG:3945 (CC45)')),
('EPSG:4326', _('EPSG:4326 (WGS 84)')),
)
class OpenGIS(BaseResource):
category = _('Geographic information system')
wms_service_url = models.URLField(_('Web Map Service (WMS) URL'), max_length=256, blank=True)
wfs_service_url = models.URLField(_('Web Feature Service (WFS) URL'), max_length=256, blank=True)
query_layer = models.CharField(
_('Query Layer'),
max_length=256,
help_text=_('Corresponds to a WMS layer name and/or a WFS feature type.'),
blank=True,
)
projection = models.CharField(
_('GIS projection'), choices=PROJECTIONS, default='EPSG:3857', max_length=16
)
search_radius = models.IntegerField(_('Radius for point search'), default=5)
attributes_mapping = (
('road', ('road', 'road_name', 'street', 'street_name', 'voie', 'nom_voie', 'rue')),
('city', ('city', 'city_name', 'town', 'town_name', 'commune', 'nom_commune', 'ville', 'nom_ville')),
('house_number', ('house_number', 'number', 'numero', 'numero_voie', 'numero_rue')),
('postcode', ('postcode', 'postalCode', 'zipcode', 'codepostal', 'cp', 'code_postal', 'code_post')),
('country', ('country', 'country_name', 'pays', 'nom_pays')),
)
class Meta:
verbose_name = _('OpenGIS')
def get_capabilities(self, service_type, service_url):
if not service_url:
raise APIError('no %s URL declared' % service_type)
return self.requests.get(
service_url,
params={'request': 'GetCapabilities', 'service': service_type.upper()},
)
def get_service_version(self, service_type, service_url, renew=False):
cache_key = 'opengis-%s-%s-version' % (service_type, self.id)
if not renew:
service_version = cache.get(cache_key)
if service_version:
return service_version
response = self.get_capabilities(service_type, service_url)
element = ET.fromstring(response.content)
service_version = element.attrib.get('version')
# cache version number for an hour
cache.set(cache_key, service_version, 3600)
return service_version
def get_wms_service_version(self, renew=False):
return self.get_service_version('wms', self.wms_service_url, renew=renew)
def get_wfs_service_version(self, renew=False):
return self.get_service_version('wfs', self.wfs_service_url, renew=renew)
def get_output_format(self):
cache_key = 'opengis-%s-output-format' % self.id
output_format = cache.get(cache_key)
if output_format:
return output_format
response = self.get_capabilities('wfs', self.wfs_service_url)
element = ET.fromstring(response.content)
ns = {'ows': 'http://www.opengis.net/ows/1.1'}
formats = element.findall(
'.//ows:Operation[@name="GetFeature"]/'
'ows:Parameter[@name="outputFormat"]/'
'ows:AllowedValues/ows:Value',
ns,
)
for output_format in formats:
if 'json' in output_format.text.lower():
cache.set(cache_key, output_format.text, 3600)
return output_format.text
raise APIError('WFS server doesn\'t support json output format for GetFeature request')
def get_typename_label(self):
version_str = self.get_wfs_service_version()
version_tuple = tuple(int(x) for x in version_str.split('.'))
if version_tuple <= (1, 1, 0):
return 'typename'
else:
return 'typenames'
def check_status(self):
if self.wms_service_url:
response = self.requests.get(
self.wms_service_url, params={'service': 'WMS', 'request': 'GetCapabilities'}
)
response.raise_for_status()
if self.wfs_service_url:
response = self.requests.get(
self.wfs_service_url, params={'service': 'WFS', 'request': 'GetCapabilities'}
)
response.raise_for_status()
def build_get_features_params(self, typename=None, property_name=None, cql_filter=None, xml_filter=None):
params = {
'version': self.get_wfs_service_version(),
'service': 'WFS',
'request': 'GetFeature',
self.get_typename_label(): typename or self.query_layer,
'outputFormat': self.get_output_format(),
}
if property_name:
params['propertyName'] = property_name
if cql_filter:
params['cql_filter'] = cql_filter
if xml_filter:
params['filter'] = xml_filter
return params
@endpoint(
perm='can_access',
description=_('Get features'),
parameters={
'type_names': {
'description': _('Type of feature to query. Defaults to globally defined query layer'),
'example_value': 'feature',
},
'property_name': {'description': _('Property to list'), 'example_value': 'nom_commune'},
'cql_filter': {
'description': _('CQL filter applied to the query'),
'example_value': 'commune=\'Paris\'',
},
'filter_property_name': {
'description': _('Property by which to filter'),
'example_value': 'voie',
},
'q': {'description': _('Filter value'), 'example_value': 'rue du chateau'},
'case_insensitive': {
'description': _('Enables case-insensitive search'),
},
'xml_filter': {
'description': _('Filter applied to the query'),
'example_value': '<Filter><PropertyIsEqualTo><PropertyName>typeparking'
'</PropertyName></PropertyIsEqualTo></Filter>',
},
},
)
def features(
self,
request,
property_name,
type_names=None,
cql_filter=None,
filter_property_name=None,
q=None,
case_insensitive=False,
xml_filter=None,
**kwargs,
):
if cql_filter:
if filter_property_name and q:
if 'case-insensitive' in kwargs or case_insensitive:
operator = 'ILIKE'
else:
operator = 'LIKE'
cql_filter += ' AND %s %s \'%%%s%%\'' % (filter_property_name, operator, q)
params = self.build_get_features_params(type_names, property_name, cql_filter, xml_filter)
response = self.requests.get(self.wfs_service_url, params=params)
data = []
try:
response = response.json()
except ValueError:
self.handle_opengis_error(response)
# if handle_opengis_error did not raise an error, we raise a generic one
raise APIError('OpenGIS Error: unparsable error', data={'content': repr(response.content[:1024])})
for feature in response['features']:
feature['text'] = feature['properties'].get(property_name)
data.append(feature)
return {'data': data}
def handle_opengis_error(self, response):
try:
root = ET.fromstring(response.content)
except ET.ParseError:
return None
if root.tag != '{http://www.opengis.net/ows/1.1}ExceptionReport':
return None
exception = root.find('{http://www.opengis.net/ows/1.1}Exception')
exception_code = exception.attrib.get('exceptionCode')
if exception is None:
return None
exception_text = exception.find('{http://www.opengis.net/ows/1.1}ExceptionText')
if exception_text is None:
return None
content = exception_text.text
content = html.unescape(content)
raise APIError('OpenGIS Error: %s' % exception_code or 'unknown code', data={'text': content})
def convert_coordinates(self, lon, lat, reverse=False):
lon, lat = float(lon), float(lat)
if self.projection != 'EPSG:4326':
wgs84 = pyproj.Proj(init='EPSG:4326')
target_projection = pyproj.Proj(init=self.projection)
if reverse:
# pylint: disable=unpacking-non-sequence
lon, lat = pyproj.transform( # pylint: disable=unpacking-non-sequence
target_projection, wgs84, lon, lat
)
else:
lon, lat = pyproj.transform( # pylint: disable=unpacking-non-sequence
wgs84, target_projection, lon, lat
)
return lon, lat
def get_bbox(self, lon1, lat1, lon2, lat2):
if self.projection == 'EPSG:4326':
# send as is but invert coordinates
return '%.6f,%.6f,%.6f,%.6f' % (lat1, lon1, lat2, lon2)
wgs84 = pyproj.Proj(init='EPSG:4326')
target_projection = pyproj.Proj(init=self.projection)
# pylint: disable=unpacking-non-sequence
x1, y1 = pyproj.transform(wgs84, target_projection, lon1, lat1)
# pylint: disable=unpacking-non-sequence
x2, y2 = pyproj.transform(wgs84, target_projection, lon2, lat2)
return '%.6f,%.6f,%.6f,%.6f' % (x1, y1, x2, y2)
@endpoint(
perm='can_access',
description=_('Get feature info'),
parameters={
'lat': {'description': _('Latitude'), 'example_value': '45.79689'},
'lon': {'description': _('Longitude'), 'example_value': '4.78414'},
'query_layer': {'description': _('Defaults to globally defined query layer')},
},
)
def feature_info(self, request, lat, lon, query_layer=None):
try:
lat, lon = float(lat), float(lon)
except ValueError:
raise APIError('Bad coordinates format')
bbox = self.get_bbox(lon - 0.0001, lat - 0.0001, lon + 0.0001, lat + 0.0001)
params = {
'version': '1.3.0',
'service': 'WMS',
'request': 'GetFeatureInfo',
'info_format': 'application/vnd.ogc.gml',
'styles': '',
'i': '24',
'J': '24', # pixel in the middle of
'height': '50',
'WIDTH': '50', # a 50x50 square
'crs': self.projection,
'layers': query_layer or self.query_layer,
'query_layers': query_layer or self.query_layer,
'bbox': bbox,
}
response = self.requests.get(self.wms_service_url, params=params)
element = ET.fromstring(response.content)
return {'err': 0, 'data': build_dict_from_xml(element)}
# https://carton.entrouvert.org/hydda-tiles/16/33650/23378.pn
@endpoint(
name='tile',
pattern=r'^(?P<zoom>\d+)/(?P<tile_x>\d+)/(?P<tile_y>\d+).png',
description=_('Get Map Tile'),
example_pattern='{zoom}/{tile_x}/{tile_y}.png',
parameters={
'zoom': {'description': _('Zoom Level'), 'example_value': '16'},
'tile_x': {'description': _('X Coordinate'), 'example_value': '33650'},
'tile_y': {'description': _('Y Coordinate'), 'example_value': '23378'},
'query_layer': {'description': _('Defaults to globally defined query layer')},
},
)
def tile(self, request, zoom, tile_x, tile_y, query_layer=None):
zoom = int(zoom)
tile_x = int(tile_x)
tile_y = int(tile_y)
# lower left
ll_lon, ll_lat = num2deg(tile_x, tile_y + 1, zoom)
# upper right
ur_lon, ur_lat = num2deg(tile_x + 1, tile_y, zoom)
bbox = self.get_bbox(ll_lon, ll_lat, ur_lon, ur_lat)
params = {
'version': '1.3.0',
'service': 'wMS',
'request': 'getMap',
'layers': query_layer or self.query_layer,
'styles': '',
'format': 'image/png',
'transparent': 'false',
'height': '256',
'width': '256',
'crs': self.projection,
'bbox': bbox,
}
response = self.requests.get(self.wms_service_url, params=params, cache_duration=300)
return HttpResponse(response.content, content_type='image/png')
@endpoint(
perm='can_access',
description=_('Get feature info'),
parameters={
'lat': {'description': _('Latitude'), 'example_value': '45.79689'},
'lon': {'description': _('Longitude'), 'example_value': '4.78414'},
'type_names': {
'description': _('Type of feature to query. Defaults to globally defined query layer'),
'example_value': 'feature',
},
},
)
def reverse(self, request, lat, lon, type_names=None, **kwargs):
lon, lat = self.convert_coordinates(lon, lat)
cql_filter = 'DWITHIN(the_geom,Point(%.6f %.6f),%s,meters)' % (lon, lat, self.search_radius)
params = self.build_get_features_params(typename=type_names, cql_filter=cql_filter)
response = self.requests.get(self.wfs_service_url, params=params)
if not response.ok:
raise APIError('Webservice returned status code %s' % response.status_code)
closest_feature = {}
min_delta = None
for feature in response.json().get('features'):
if not feature['geometry']['type'] == 'Point':
continue # skip unknown
lon_diff = abs(float(lon) - float(feature['geometry']['coordinates'][0]))
lat_diff = abs(float(lat) - float(feature['geometry']['coordinates'][1]))
# compute hypotenuse til the point
delta = math.sqrt(lon_diff * lon_diff + lat_diff * lat_diff)
if min_delta is None:
min_delta = delta
# choose the shortest
if delta <= min_delta:
closest_feature = feature
if closest_feature:
result = {}
point_lon = closest_feature['geometry']['coordinates'][0]
point_lat = closest_feature['geometry']['coordinates'][1]
point_lon, point_lat = self.convert_coordinates(point_lon, point_lat, reverse=True)
result['lon'] = "%.6f" % point_lon
result['lat'] = "%.6f" % point_lat
result['address'] = {}
for attribute, properties in self.attributes_mapping:
for field in properties:
if closest_feature['properties'].get(field):
result['address'][attribute] = str(closest_feature['properties'][field])
break
return result
raise APIError('Unable to geocode')
@endpoint(
name='query',
description=_('Query'),
pattern=r'^(?P<query_slug>[\w:_-]+)/$',
perm='can_access',
parameters={
'bbox': {
'description': _(
'Only include results inside bounding box. Expected coordinates '
'format is lonmin,latmin,lonmax,latmax'
),
},
'circle': {
'description': _(
'Only include results inside circle. Expected coordinates '
'format is center_longitude,center_latitude,radius ("radius" '
'being a distance in meters)'
),
},
'q': {
'description': _('Text search for specified properties'),
},
'property': {
'description': _(
'Filter on any GeoJSON property value. If a feature has an "id" property, '
'format will be property:id=value.'
),
'optional': True,
},
},
show=False,
)
def query(self, request, query_slug, bbox=None, circle=None, q=None, **kwargs):
if bbox and circle:
raise APIError('bbox and circle parameters are mutually exclusive')
for param in kwargs:
if not param.startswith('property'):
raise APIError('extra parameter: %s' % param)
query = get_object_or_404(Query, resource=self, slug=query_slug)
return query.q(request, bbox, q, circle, **kwargs)
def export_json(self):
d = super().export_json()
d['queries'] = [query.export_json() for query in self.queries.all()]
return d
@classmethod
def import_json_real(cls, overwrite, instance, d, **kwargs):
queries = d.pop('queries', [])
instance = super().import_json_real(overwrite, instance, d, **kwargs)
new = []
if instance and overwrite:
Query.objects.filter(resource=instance).delete()
for query in queries:
q = Query.import_json(query)
q.resource = instance
new.append(q)
Query.objects.bulk_create(new)
return instance
def daily(self):
super().daily()
self.update_queries()
def update_queries(self, query_id=None):
queries = self.queries.all()
if query_id:
queries = queries.filter(pk=query_id)
for query in queries:
try:
query.update_cache()
except APIError:
self.logger.warning('failed to cache results for query %s', query)
def create_query_url(self):
return reverse('opengis-query-new', kwargs={'slug': self.slug})
def save(self, *args, **kwargs):
super().save(*args, **kwargs)
if self.queries.exists():
self.add_job('update_queries')
class Query(BaseQuery):
resource = models.ForeignKey(
to=OpenGIS, on_delete=models.CASCADE, related_name='queries', verbose_name=_('Resource')
)
typename = models.CharField(_('Feature type'), max_length=256)
filter_expression = models.TextField(_('XML filter'), blank=True)
indexing_template = models.TextField(
verbose_name=_('Indexing template'), blank=True, validators=[validate_template]
)
computed_properties = JSONField(blank=True, default=dict)
delete_view = 'opengis-query-delete'
edit_view = 'opengis-query-edit'
def as_endpoint(self):
resource_endpoint = self.resource.query.endpoint_info
endpoint = super().as_endpoint(path=resource_endpoint.name)
# use parameters added by resource endpoint
endpoint.func = resource_endpoint.func
endpoint.show_undocumented_params = False
endpoint.parameters = resource_endpoint.parameters
return endpoint
def q(self, request, bbox, q, circle, **kwargs):
features = self.features.all()
if not features.exists():
raise APIError(
'No data. (maybe not synchronized yet? retry in a few minutes.)', extra_dict={'features': []}
)
filters = {}
for lookup, value in kwargs.items():
lookup = lookup.replace('property', 'data__properties').replace(':', '__')
converted_value = self.cast_str(value)
if converted_value:
lookup += '__in'
filters[lookup] = [value, converted_value]
else:
filters[lookup] = value
features = features.filter(**filters)
lonmin, latmin, lonmax, latmax = None, None, None, None
if bbox:
try:
lonmin, latmin, lonmax, latmax = (float(x) for x in bbox.split(','))
except (ValueError, AttributeError):
raise APIError(
'Invalid bbox parameter, it must be a comma separated list of floating point numbers.'
)
if circle:
try:
center_lon, center_lat, radius = (float(x) for x in circle.split(','))
except (ValueError, AttributeError):
raise APIError(
'Invalid circle parameter, it must be a comma separated list of '
'floating point numbers.'
)
coords = self.get_bbox_containing_circle(center_lon, center_lat, radius)
lonmin, latmin, lonmax, latmax = coords
if lonmin is not None:
# adjust lonmin, latmin, lonmax, latmax to make sure min are min and max are max
lonmin, lonmax = min(lonmin, lonmax), max(lonmin, lonmax)
latmin, latmax = min(latmin, latmax), max(latmin, latmax)
features = features.filter(
Q(bbox_lat2__isnull=True, lon__gte=lonmin, lon__lte=lonmax, lat__gte=latmin, lat__lte=latmax)
| Q(
bbox_lat2__isnull=False, # geometry != Point
lon__lte=lonmax,
bbox_lon2__gte=lonmin,
lat__lte=latmax,
bbox_lat2__gte=latmin,
)
)
if q:
features = features.filter(text__search=simplify(q))
data = {'type': 'FeatureCollection', 'name': self.typename}
if circle:
results = []
for feature in features:
if feature.bbox_lat2: # not a point
results.append(feature.data)
else:
distance = self.get_coords_distance(feature.lat, feature.lon, center_lat, center_lon)
if distance < radius:
results.append(feature.data)
data['features'] = results
else:
data['features'] = list(features.values_list('data', flat=True))
return data
@staticmethod
def get_bbox_containing_circle(lon, lat, r):
earth_radius = 6378137
offset_in_radian = 180 / math.pi * (r / earth_radius)
lonmin = lon - offset_in_radian
lonmax = lon + offset_in_radian
latmin = lat - offset_in_radian
latmax = lat + offset_in_radian
return lonmin, latmin, lonmax, latmax
@staticmethod
def get_coords_distance(lat, lon, lat0, lon0):
# simplest distance approximation https://www.mkompf.com/gps/distcalc.html
deglen = 111300
x = lat - lat0
lat = (lat + lat0) / 2 * 0.01745
y = (lon - lon0) * math.cos(lat)
return deglen * math.sqrt(x * x + y * y)
@staticmethod
def cast_str(s):
if s.lower() == 'true':
return True
if s.lower() == 'false':
return False
try:
return int(s)
except ValueError:
try:
return float(s)
except ValueError:
return None
def update_cache(self):
data = self.resource.features(None, None, type_names=self.typename, xml_filter=self.filter_expression)
features = []
templates = {}
if self.indexing_template:
templates['text'] = Template(self.indexing_template)
for key, tplt in (self.computed_properties or {}).items():
try:
templates['computed_property_%s' % key] = Template(tplt)
except TemplateSyntaxError:
pass
def add_coordinates(coordinates):
nonlocal min_lat, min_lon, max_lat, max_lon
if not coordinates:
return
if not isinstance(coordinates[0], (float, int)):
for child in coordinates:
add_coordinates(child)
return
# position
lon, lat = coordinates
if min_lat is None or lat < min_lat:
min_lat = lat
if max_lat is None or lat > max_lat:
max_lat = lat
if min_lon is None or lon < min_lon:
min_lon = lon
if max_lon is None or lon > max_lon:
max_lon = lon
for feature in data['data']:
geometry = feature.get('geometry') or {}
if not geometry:
continue
if geometry.get('type') == 'Point':
try:
lon, lat = geometry['coordinates']
except (KeyError, TypeError):
self.resource.logger.warning('invalid coordinates in geometry: %s', geometry)
continue
lon2, lat2 = None, None
else:
# define bbox, lat/lon as min values, bbox_lat2/bbox_lon2 as max values
min_lat, min_lon, max_lat, max_lon = None, None, None, None
add_coordinates(geometry['coordinates'])
lat, lon, lat2, lon2 = min_lat, min_lon, max_lat, max_lon
text = ''
if self.indexing_template:
context = Context(feature.get('properties', {}))
text = simplify(templates['text'].render(context))
context = Context(feature)
for key in self.computed_properties or {}:
if not templates.get('computed_property_%s' % key):
continue
if not feature.get('properties'):
feature['properties'] = {}
feature['properties'][key] = templates['computed_property_%s' % key].render(context)
features.append(
FeatureCache(
query=self,
lat=lat,
lon=lon,
bbox_lat2=lat2,
bbox_lon2=lon2,
text=text,
data=feature,
)
)
with transaction.atomic():
self.features.all().delete()
FeatureCache.objects.bulk_create(features)
def save(self, *args, **kwargs):
super().save(*args, **kwargs)
self.resource.add_job('update_queries', query_id=self.pk)
class FeatureCache(models.Model):
query = models.ForeignKey(
to=Query, on_delete=models.CASCADE, related_name='features', verbose_name=_('Query')
)
lat = models.FloatField()
lon = models.FloatField()
bbox_lat2 = models.FloatField(blank=True, null=True)
bbox_lon2 = models.FloatField(blank=True, null=True)
text = models.CharField(max_length=2048)
data = JSONField()