def populate_datamart(): print 'populate datamart' adminlevel_REG = AdminLevelType.get(u'REG') from functools import partial import pyproj from shapely.ops import transform project = partial(pyproj.transform, pyproj.Proj(init='epsg:4326'), pyproj.Proj(init='epsg:3857')) shape = MultiPolygon([Polygon([(0, 0), (0, 1), (1, 1), (1, 0), (0, 0)])]) reprojected = transform(project, shape) geometry = from_shape(reprojected, 3857) div = AdministrativeDivision( **{ 'code': 30, 'leveltype_id': adminlevel_REG.id, 'name': u'Administrative division level 3' }) div.geom = geometry div.hazardcategories = [] DBSession.add(div) DBSession.flush()
def upscale_hazardcategories(target_adminlevel_mnemonic): # identify the admin divisions matching the given level: admindivs = DBSession.query(AdministrativeDivision)\ .join(AdminLevelType) \ .filter(AdminLevelType.mnemonic == target_adminlevel_mnemonic) hazardtypes = DBSession.query(HazardType) # for each of these admin divisions and hazard types, # identify the highest HazardCategory among their children: for admindiv in admindivs: # identify max(level) for each hazardtype across children for hazardtype in hazardtypes: hazardcategory = DBSession.query(HazardCategory) \ .join((AdministrativeDivision, HazardCategory.administrativedivisions)) \ .join(HazardType) \ .join(HazardLevel) \ .filter(HazardType.id == hazardtype.id) \ .filter(AdministrativeDivision.parent == admindiv) \ .order_by(HazardLevel.order.asc()).first() if hazardcategory: # find the highest hazardlevel for all children admindivs print '[upscaling] admindiv {} inherits hazardlevel {} for {}'\ .format(admindiv.code, hazardcategory.hazardlevel_id, hazardcategory.hazardtype.mnemonic) admindiv.hazardcategories.append(hazardcategory) DBSession.add(admindiv)
def populate_preprocessed(type): hazardset_id = u'preprocessed' hazardtype = HazardType.get(type) print 'Populating hazardset {}'.format(hazardset_id) hazardset = HazardSet( id=hazardset_id, hazardtype=hazardtype, local=False, data_lastupdated_date=datetime.now(), metadata_lastupdated_date=datetime.now()) DBSession.add(hazardset) layer = Layer( hazardlevel=None, mask=False, return_period=None, data_lastupdated_date=datetime.now(), metadata_lastupdated_date=datetime.now(), geonode_id=new_geonode_id(), download_url='test', calculation_method_quality=5, scientific_quality=1, local=False, downloaded=True ) hazardset.layers.append(layer) hazardset.complete = True DBSession.flush()
def populate_notpreprocessed(type, unit): hazardset_id = u'notpreprocessed' hazardtype = HazardType.get(type) hazardtype_settings = settings['hazard_types'][hazardtype.mnemonic] print 'Populating hazardset {}'.format(hazardset_id) hazardset = HazardSet( id=hazardset_id, hazardtype=hazardtype, local=False, data_lastupdated_date=datetime.now(), metadata_lastupdated_date=datetime.now()) DBSession.add(hazardset) return_periods = hazardtype_settings['return_periods'] for level in (u'HIG', u'MED', u'LOW'): hazardlevel = HazardLevel.get(level) level_return_periods = return_periods[level] if isinstance(level_return_periods, list): return_period = level_return_periods[0] else: return_period = level_return_periods layer = Layer( hazardlevel=hazardlevel, mask=False, return_period=return_period, hazardunit=unit, data_lastupdated_date=datetime.now(), metadata_lastupdated_date=datetime.now(), geonode_id=new_geonode_id(), download_url='test', calculation_method_quality=5, scientific_quality=1, local=False, downloaded=True ) hazardset.layers.append(layer) layer = Layer( hazardlevel=None, mask=True, return_period=hazardtype_settings['mask_return_period'], hazardunit=unit, data_lastupdated_date=datetime.now(), metadata_lastupdated_date=datetime.now(), geonode_id=new_geonode_id(), download_url='test', calculation_method_quality=5, scientific_quality=1, local=False, downloaded=True ) hazardset.layers.append(layer) hazardset.complete = True DBSession.flush()
def add_resource(hazard_type, resource): categories = DBSession.query(HazardCategory) \ .join(HazardType) \ .filter(HazardType.mnemonic == hazard_type) for category in categories: association = HazardCategoryFurtherResourceAssociation(order=1) association.hazardcategory = category resource.hazardcategory_associations.append(association) DBSession.add(resource)
def process_outputs(): print "Decision Tree running..." # first of all, remove all records # in the datamart table linking admin divs with hazard categories: hazardcategory_administrativedivision_table.delete() # identify the admin level for which we run the decision tree: # (REG)ion aka admin level 2 dt_level = DBSession.query(AdminLevelType)\ .filter(AdminLevelType.mnemonic == u'REG').one() # then, identify the unique (admindiv, hazardtype) tuples # contained in the Output table: admindiv_hazardtype_tuples = ( DBSession.query(AdministrativeDivision, HazardType).distinct() .join(Output) .join(HazardSet) .join(HazardType) # the following should not be necessary in production # because only the lowest admin levels should be inserted # in the Output table: .filter(AdministrativeDivision.leveltype_id == dt_level.id) # not necessary, but practical for debugging: .order_by(AdministrativeDivision.code) ) # for each tuple, identify the most relevant HazardSet # in the light of the criteria that we all agreed on (cf Decision Tree): for (admindiv, hazardtype) in admindiv_hazardtype_tuples: (hazardset, output) = DBSession.query(HazardSet, Output) \ .join(Output) \ .filter(Output.admin_id == admindiv.id) \ .filter(HazardSet.hazardtype_id == hazardtype.id) \ .order_by(HazardSet.calculation_method_quality.desc(), HazardSet.scientific_quality.desc(), HazardSet.local.desc(), HazardSet.data_lastupdated_date.desc()).first() print "[decision tree] admindiv {} gets hazardlevel {} from {} for {}"\ .format(admindiv.code, output.hazardlevel_id, hazardset.id, hazardtype.mnemonic) # find the relevant HazardCategory for the current hazardtype # and the hazardset's hazardlevel hazardcategory = DBSession.query(HazardCategory) \ .filter(HazardCategory.hazardtype_id == hazardtype.id) \ .filter(HazardCategory.hazardlevel_id == output.hazardlevel_id) \ .one() # append new hazardcategory to current admin div: admindiv.hazardcategories.append(hazardcategory) DBSession.add(admindiv) # UpScaling level2 (REG)ion -> level1 (PRO)vince upscale_hazardcategories(u'PRO') # UpScaling level1 (PRO)vince -> level0 (COU)ntry upscale_hazardcategories(u'COU') transaction.commit()
def populate_datamart(): print 'populate datamart' adminlevel_REG = AdminLevelType.get(u'REG') shape = MultiPolygon([ Polygon([(0, 0), (0, 1), (1, 1), (1, 0), (0, 0)]) ]) geometry = from_shape(shape, 4326) div = AdministrativeDivision(**{ 'code': 30, 'leveltype_id': adminlevel_REG.id, 'name': u'Administrative division level 3' }) div.geom = geometry div.hazardcategories = [] DBSession.add(div) DBSession.flush()
def populate_processing(): print 'populate processing' hazardset_id = u'test' hazardtype = HazardType.get(u'EQ') hazardtype_settings = settings['hazard_types'][hazardtype.mnemonic] print 'Populating hazardset {}'.format(hazardset_id) hazardset = HazardSet() hazardset.id = hazardset_id hazardset.hazardtype = hazardtype hazardset.local = False hazardset.data_lastupdated_date = datetime.now() hazardset.metadata_lastupdated_date = datetime.now() DBSession.add(hazardset) return_periods = hazardtype_settings['global']['return_periods'] unit = hazardtype_settings['thresholds'].keys()[0] for level in (u'HIG', u'MED', u'LOW'): hazardlevel = HazardLevel.get(level) return_period = return_periods[level] layer = Layer() layer.title = "{}-{}".format(id, return_period) layer.hazardlevel = hazardlevel layer.return_period = return_period layer.hazardunit = unit layer.data_lastupdated_date = datetime.now() layer.metadata_lastupdated_date = datetime.now() layer.geonode_id = new_geonode_id() layer.download_url = 'test' layer.calculation_method_quality = 5 layer.scientific_quality = 1 layer.local = False layer.downloaded = True hazardset.layers.append(layer) DBSession.flush() hazardset.complete = True DBSession.flush()
def populate_db(): config = ConfigParser.ConfigParser() config.read(local_settings_path) db_url = config.get('app:main', 'sqlalchemy.url') from sqlalchemy import create_engine engine = create_engine(db_url) from ..scripts.initializedb import populate_db as populate populate(engine, drop_all=True) shape = MultiPolygon([ Polygon([(0, 0), (0, 1), (1, 1), (1, 0), (0, 0)]) ]) geometry = from_shape(shape, 4326) div_level_1 = AdministrativeDivision(**{ 'code': 10, 'leveltype_id': 1, 'name': u'Division level 1' }) div_level_1.geom = geometry DBSession.add(div_level_1) div_level_2 = AdministrativeDivision(**{ 'code': 20, 'leveltype_id': 2, 'name': u'Division level 2' }) div_level_2.parent_code = div_level_1.code div_level_2.geom = geometry DBSession.add(div_level_2) shape = MultiPolygon([ Polygon([(0, 0), (0, 1), (.5, 1), (.5, 0), (0, 0)]) ]) geometry = from_shape(shape, 4326) div_level_3_1 = AdministrativeDivision(**{ 'code': 30, 'leveltype_id': 3, 'name': u'Division level 3 - 1' }) div_level_3_1.parent_code = div_level_2.code div_level_3_1.geom = geometry div_level_3_1.hazardcategories = [] shape = MultiPolygon([ Polygon([(.5, 0), (.5, 1), (1, 1), (1, 0), (.5, 0)]) ]) geometry = from_shape(shape, 4326) div_level_3_2 = AdministrativeDivision(**{ 'code': 31, 'leveltype_id': 3, 'name': u'Division level 3 - 2' }) div_level_3_2.parent_code = div_level_2.code div_level_3_2.geom = geometry div_level_3_2.hazardcategories = [] category_eq_hig = HazardCategory(**{ 'general_recommendation': u'General recommendation for EQ HIG', }) category_eq_hig.hazardtype = DBSession.query(HazardType) \ .filter(HazardType.mnemonic == u'EQ').one() category_eq_hig.hazardlevel = DBSession.query(HazardLevel) \ .filter(HazardLevel.mnemonic == u'HIG').one() associate_admindiv_hazardcategory(div_level_3_1, category_eq_hig, u'test') associate_admindiv_hazardcategory(div_level_3_2, category_eq_hig, u'test') climate_rec = ClimateChangeRecommendation() climate_rec.text = u'Climate change recommendation' climate_rec.administrativedivision = div_level_1 climate_rec.hazardtype = DBSession.query(HazardType) \ .filter(HazardType.mnemonic == u'EQ').one() DBSession.add(climate_rec) technical_rec = TechnicalRecommendation(**{ 'text': u'Recommendation #1 for earthquake, applied to' ' hazard categories HIG, MED and LOW' }) association = HazardCategoryTechnicalRecommendationAssociation(order=1) association.hazardcategory = category_eq_hig technical_rec.hazardcategory_associations.append(association) DBSession.add(technical_rec) technical_rec = TechnicalRecommendation(**{ 'text': u'Educational web resources on earthquakes and' ' seismic hazard' }) association = HazardCategoryTechnicalRecommendationAssociation(order=1) association.hazardcategory = category_eq_hig technical_rec.hazardcategory_associations.append(association) DBSession.add(technical_rec) category_fl_med = HazardCategory(**{ 'general_recommendation': u'General recommendation for FL MED', }) category_fl_med.hazardtype = DBSession.query(HazardType) \ .filter(HazardType.mnemonic == u'FL').one() category_fl_med.hazardlevel = DBSession.query(HazardLevel) \ .filter(HazardLevel.mnemonic == u'MED').one() associate_admindiv_hazardcategory(div_level_3_1, category_fl_med, u'test') DBSession.add(div_level_3_1) associate_admindiv_hazardcategory(div_level_3_2, category_fl_med, u'test') DBSession.add(div_level_3_2) further_resource = FurtherResource(**{ 'text': u'Educational web resources on earthquakes and' + ' seismic hazard', 'url': u'http://earthquake.usgs.gov/learn/?source=sitemap' }) association = HazardCategoryFurtherResourceAssociation(order=1) association.hazardcategory = category_eq_hig further_resource.hazardcategory_associations.append(association) DBSession.add(further_resource) # Add further resource for one division only further_resource = FurtherResource(**{ 'text': u'Further resource specific to division level 3 - 2', 'url': u'http://domain.com/the/document/url.txt' }) association = HazardCategoryFurtherResourceAssociation(order=2) association.hazardcategory = category_eq_hig association.administrativedivision = div_level_3_2 further_resource.hazardcategory_associations.append(association) DBSession.add(further_resource)
def populate_datamart(): shape = Polygon([(0, 0), (0, 1), (1, 1), (1, 0), (0, 0)]) geometry = from_shape(shape, 3857) div_level_1 = AdministrativeDivision(**{ 'code': 10, 'leveltype_id': 1, 'name': u'Division level 1' }) div_level_1.geom = geometry DBSession.add(div_level_1) div_level_2 = AdministrativeDivision(**{ 'code': 20, 'leveltype_id': 2, 'name': u'Division level 2' }) div_level_2.parent_code = div_level_1.code div_level_2.geom = geometry DBSession.add(div_level_2) div_level_3_0 = AdministrativeDivision(**{ 'code': 30, 'leveltype_id': 3, 'name': u'Division level 3 - 1' }) div_level_3_0.parent_code = div_level_2.code div_level_3_0.geom = geometry div_level_3_0.hazardcategories = [] DBSession.add(div_level_3_0) div_level_3_1 = AdministrativeDivision(**{ 'code': 31, 'leveltype_id': 3, 'name': u'Division level 3 - 1' }) div_level_3_1.parent_code = div_level_2.code div_level_3_1.geom = geometry div_level_3_1.hazardcategories = [] DBSession.add(div_level_3_1) div_level_3_2 = AdministrativeDivision(**{ 'code': 32, 'leveltype_id': 3, 'name': u'Division level 3 - 2' }) div_level_3_2.parent_code = div_level_2.code div_level_3_2.geom = geometry div_level_3_2.hazardcategories = [] DBSession.add(div_level_3_2) div_level_3_3 = AdministrativeDivision(**{ 'code': 33, 'leveltype_id': 3, 'name': u'Division level 3 - 3' }) div_level_3_3.parent_code = div_level_2.code div_level_3_3.geom = geometry div_level_3_3.hazardcategories = [] DBSession.add(div_level_3_3) div_level_3_4 = AdministrativeDivision(**{ 'code': 34, 'leveltype_id': 3, 'name': u'Division level 3 - 4' }) div_level_3_4.parent_code = div_level_2.code div_level_3_4.geom = geometry div_level_3_4.hazardcategories = [] DBSession.add(div_level_3_4) div_level_3_5 = AdministrativeDivision(**{ 'code': 35, 'leveltype_id': 3, 'name': u'Division level 3 - 5' }) div_level_3_5.parent_code = div_level_2.code div_level_3_5.geom = geometry div_level_3_5.hazardcategories = [] DBSession.add(div_level_3_5) div_level_3_6 = AdministrativeDivision(**{ 'code': 36, 'leveltype_id': 3, 'name': u'Division level 3 - 6' }) div_level_3_6.parent_code = div_level_2.code div_level_3_6.geom = geometry div_level_3_6.hazardcategories = [] DBSession.add(div_level_3_6) DBSession.flush()
def populate_processing(): hazardtypeEQ = HazardType.get(u'EQ') hazardtypeFL = HazardType.get(u'FL') hazardset_id = u'hazardset1' print 'Populating hazardset {}'.format(hazardset_id) hazardset1 = HazardSet() hazardset1.id = hazardset_id hazardset1.hazardtype = hazardtypeEQ hazardset1.local = False hazardset1.calculation_method_quality = 0 hazardset1.data_lastupdated_date = datetime.now() hazardset1.metadata_lastupdated_date = datetime.now() hazardset1.complete = True hazardset1.layers.extend(make_layers()) DBSession.add(hazardset1) hazardset_id = u'hazardset1b' print 'Populating hazardset {}'.format(hazardset_id) hazardset1b = HazardSet() hazardset1b.id = hazardset_id hazardset1b.hazardtype = hazardtypeFL hazardset1b.local = True hazardset1b.calculation_method_quality = 3 hazardset1b.data_lastupdated_date = datetime.now() hazardset1b.metadata_lastupdated_date = datetime.now() hazardset1b.complete = True hazardset1b.layers.extend(make_layers()) DBSession.add(hazardset1b) # create hazardsets 2, 3 ... 5 hazardset_id = u'hazardset2' # calculation_method_quality = 1 / scientific_quality = 0 print 'Populating hazardset {}'.format(hazardset_id) hazardset2 = HazardSet() hazardset2.id = hazardset_id hazardset2.hazardtype = hazardtypeEQ hazardset2.local = True hazardset2.calculation_method_quality = 1 hazardset2.scientific_quality = 0 hazardset2.data_lastupdated_date = datetime.now() hazardset2.metadata_lastupdated_date = datetime.now() hazardset2.complete = True hazardset2.layers.extend(make_layers()) DBSession.add(hazardset2) hazardset_id = u'hazardset3' # date = 2015-01-01 / global / # calculation_method_quality = 1 / scientific_quality = 2 print 'Populating hazardset {}'.format(hazardset_id) hazardset3 = HazardSet() hazardset3.id = hazardset_id hazardset3.hazardtype = hazardtypeEQ hazardset3.local = False hazardset3.calculation_method_quality = 1 hazardset3.scientific_quality = 2 hazardset3.data_lastupdated_date = datetime(2015, 1, 1, 0, 0) hazardset3.metadata_lastupdated_date = datetime.now() hazardset3.complete = True hazardset3.layers.extend(make_layers()) DBSession.add(hazardset3) hazardset_id = u'hazardset4' # date = 2015-01-01 / local / # calculation_method_quality = 1 / scientific_quality = 2 print 'Populating hazardset {}'.format(hazardset_id) hazardset4 = HazardSet() hazardset4.id = hazardset_id hazardset4.hazardtype = hazardtypeEQ hazardset4.local = True hazardset4.calculation_method_quality = 1 hazardset4.scientific_quality = 2 hazardset4.data_lastupdated_date = datetime(2015, 1, 1, 0, 0) hazardset4.metadata_lastupdated_date = datetime.now() hazardset4.complete = True hazardset4.layers.extend(make_layers()) DBSession.add(hazardset4) hazardset_id = u'hazardset5' # date = now() / local / # calculation_method_quality = 1 / scientific_quality = 2 print 'Populating hazardset {}'.format(hazardset_id) hazardset5 = HazardSet() hazardset5.id = hazardset_id hazardset5.hazardtype = hazardtypeEQ hazardset5.local = True hazardset5.calculation_method_quality = 1 hazardset5.scientific_quality = 2 hazardset5.data_lastupdated_date = datetime.now() hazardset5.metadata_lastupdated_date = datetime.now() hazardset5.complete = True hazardset5.layers.extend(make_layers()) DBSession.add(hazardset5) hazardset_id = u'hazardset6' # date = now() / global / # calculation_method_quality = 1 / scientific_quality = 2 print 'Populating hazardset {}'.format(hazardset_id) hazardset6 = HazardSet() hazardset6.id = hazardset_id hazardset6.hazardtype = hazardtypeEQ hazardset6.local = False hazardset6.calculation_method_quality = 1 hazardset6.scientific_quality = 2 hazardset6.data_lastupdated_date = datetime.now() hazardset6.metadata_lastupdated_date = datetime.now() hazardset6.complete = True hazardset6.layers.extend(make_layers()) DBSession.add(hazardset6) # populate output table # admin div (code 30) has only one hazardset for EQ admin30 = DBSession.query(AdministrativeDivision)\ .filter(AdministrativeDivision.code == 30).one() # => test outcome = hazardset1 output1 = Output() output1.coverage_ratio = 10 output1.hazardset = hazardset1 output1.administrativedivision = admin30 output1.hazardlevel = HazardLevel.get(u'HIG') DBSession.add(output1) # admin div (code 30) also has another hazardset # but this one is for FL # => test outcome = hazardset1b output1b = Output() output1b.coverage_ratio = 11 output1b.hazardset = hazardset1b output1b.administrativedivision = admin30 output1b.hazardlevel = HazardLevel.get(u'NPR') DBSession.add(output1b) # admin div (code 31) has 2 hazardsets, # one with a higher calculation_method_quality # => test outcome = hazardset2 admin31 = DBSession.query(AdministrativeDivision)\ .filter(AdministrativeDivision.code == 31).one() output2 = Output() output2.coverage_ratio = 20 output2.hazardset = hazardset1 # calculation_method_quality = 0 output2.administrativedivision = admin31 output2.hazardlevel = HazardLevel.get(u'MED') DBSession.add(output2) output3 = Output() output3.coverage_ratio = 30 output3.hazardset = hazardset2 # calculation_method_quality = 1 output3.administrativedivision = admin31 output3.hazardlevel = HazardLevel.get(u'LOW') DBSession.add(output3) # admin div (code 32) has 2 hazardsets, # both share the same calculation_method_quality, # one with a higher scientific_quality # => test outcome = hazardset3 admin32 = DBSession.query(AdministrativeDivision)\ .filter(AdministrativeDivision.code == 32).one() output4 = Output() output4.coverage_ratio = 40 output4.hazardset = hazardset2 # calculation_method_quality = 1 / scientific_quality = 0 output4.administrativedivision = admin32 output4.hazardlevel = HazardLevel.get(u'MED') DBSession.add(output4) output5 = Output() output5.coverage_ratio = 50 output5.hazardset = hazardset3 # calculation_method_quality = 1 / scientific_quality = 2 output5.administrativedivision = admin32 output5.hazardlevel = HazardLevel.get(u'LOW') DBSession.add(output5) # admin div (code 33) has 2 hazardsets, # both share the same ratings, one is global, one local # => test outcome = hazardset4 admin33 = DBSession.query(AdministrativeDivision)\ .filter(AdministrativeDivision.code == 33).one() output6 = Output() output6.coverage_ratio = 60 output6.hazardset = hazardset3 # global / calculation_method_quality = 1 / scientific_quality = 2 output6.administrativedivision = admin33 output6.hazardlevel = HazardLevel.get(u'MED') DBSession.add(output6) output7 = Output() output7.coverage_ratio = 70 output7.hazardset = hazardset4 # local / calculation_method_quality = 1 / scientific_quality = 2 output7.administrativedivision = admin33 output7.hazardlevel = HazardLevel.get(u'LOW') DBSession.add(output7) # admin div (code 34) has 2 hazardsets, # both share the same ratings, are local, one is more recent # => test outcome = hazardset5 admin34 = DBSession.query(AdministrativeDivision)\ .filter(AdministrativeDivision.code == 34).one() output8 = Output() output8.coverage_ratio = 80 output8.hazardset = hazardset4 # date = 2015-01-01 / local / # calculation_method_quality = 1 / scientific_quality = 2 output8.administrativedivision = admin34 output8.hazardlevel = HazardLevel.get(u'MED') DBSession.add(output8) output9 = Output() output9.coverage_ratio = 90 output9.hazardset = hazardset5 # date = now() / local / # calculation_method_quality = 1 / scientific_quality = 2 output9.administrativedivision = admin34 output9.hazardlevel = HazardLevel.get(u'LOW') DBSession.add(output9) # admin div (code 35) has 2 hazardsets, # both share the same ratings, are global, one is more recent # => test outcome = hazardset6 admin35 = DBSession.query(AdministrativeDivision)\ .filter(AdministrativeDivision.code == 35).one() output10 = Output() output10.coverage_ratio = 95 output10.hazardset = hazardset3 # date = 2015-01-01 / global / # calculation_method_quality = 1 / scientific_quality = 2 output10.administrativedivision = admin35 output10.hazardlevel = HazardLevel.get(u'MED') DBSession.add(output10) output11 = Output() output11.coverage_ratio = 99 output11.hazardset = hazardset6 # date = now() / global / # calculation_method_quality = 1 / scientific_quality = 2 output11.administrativedivision = admin35 output11.hazardlevel = HazardLevel.get(u'LOW') DBSession.add(output11) DBSession.flush()
def import_recommendations(): engine = engine_from_config(settings, 'sqlalchemy.') DBSession.configure(bind=engine) with transaction.manager: DBSession.query(HazardCategoryTechnicalRecommendationAssociation) \ .delete() DBSession.query(TechnicalRecommendation).delete() # First load general recommendations with open('data/general_recommendations.csv', 'rb') as csvfile: recommendations = csv.reader(csvfile, delimiter=',') for row in recommendations: hazardcategory = DBSession.query(HazardCategory) \ .join(HazardLevel) \ .join(HazardType) \ .filter(HazardLevel.mnemonic == row[1]) \ .filter(HazardType.mnemonic == row[0]) \ .one() hazardcategory.general_recommendation = row[2] DBSession.add(hazardcategory) categories = [] for type in [u'EQ', u'FL', u'CY', u'TS', u'CF', u'VA', u'DG']: for level in [u'HIG', u'MED', u'LOW', u'VLO']: hazardcategory = DBSession.query(HazardCategory) \ .join(HazardLevel) \ .join(HazardType) \ .filter(HazardLevel.mnemonic == level) \ .filter(HazardType.mnemonic == type) \ .one() categories.append(hazardcategory) # Then technical recommendations hctra = HazardCategoryTechnicalRecommendationAssociation with open('data/technical_recommendations.csv', 'rb') as csvfile: recommendations = csv.reader(csvfile, delimiter=',') next(recommendations, None) # skip the headers for row in recommendations: technical_rec = TechnicalRecommendation(**{ 'text': row[0] }) associations = technical_rec.hazardcategory_associations # the other columns are hazard category (type / level) for col_index in range(1, 28): value = row[col_index] if value is not '' and value is not 'Y': association = hctra(order=value) association.hazardcategory = categories[col_index - 1] associations.append(association) DBSession.add(technical_rec) # Climate change recommendations DBSession.query(ClimateChangeRecommendation).delete() # hazard types and corresponding columns hazard_types = [ (u'FL', 6), (u'EQ', 7), (u'CY', 8), (u'CF', 9), (u'DG', 10), (u'TS', 11), (u'VA', 12), (u'LS', 13), ] with open('data/climate_change_recommendations.csv', 'rb') as csvfile: countries = csv.reader(csvfile, delimiter=',') next(countries, None) # skip the headers for row in countries: division = DBSession.query(AdministrativeDivision) \ .filter(AdministrativeDivision.code == row[1]) \ .one_or_none() if not division: continue for hazard_type, column in hazard_types: text = row[column] if text == 'NA': continue climate_rec = ClimateChangeRecommendation() climate_rec.text = row[column] climate_rec.administrativedivision = division climate_rec.hazardtype = DBSession.query(HazardType) \ .filter(HazardType.mnemonic == hazard_type).one() DBSession.add(climate_rec)
def harvest_layer(object, dry_run=False): title = object['title'] ''' typename = urllib.unquote( object['distribution_url'].split('/').pop()) # print title ''' hazardset_id = object['hazard_set'] if not hazardset_id: logger.info('{} - hazard_set is empty'.format(title)) return False hazard_type = object['hazard_type'] if not hazard_type: logger.warning('{} - hazard_type is empty'.format(title)) return False hazardtype = hazardtype_from_geonode(hazard_type) if hazardtype is None: logger.warning('{} - hazard_type not supported: {}' .format(title, hazard_type)) return False type_settings = settings['hazard_types'][hazardtype.mnemonic] preprocessed = 'values' in type_settings ''' csw_wkt_geometry = object['csw_wkt_geometry'] bounds = wkt_loads(csw_wkt_geometry).bounds # print ' bounds :', bounds # TODO: minimum global bbox should be defined in settings minimum_global_bounds = (-175, -45, 175, 45) from shapely.geometry import box local = not box(bounds).contains(box(minimum_global_bounds)) ''' local = 'GLOBAL' not in hazardset_id ''' local = ( bounds[0] > -175 or bounds[1] > -45 or bounds[2] < 175 or bounds[3] < 45) ''' mask = False if preprocessed is True: hazardlevel = None hazard_unit = None if object['hazard_period']: logger.info('{} - Has a return period'.format(title)) return False hazard_period = None else: hazard_period = int(object['hazard_period']) hazardlevel = None for level in (u'LOW', u'MED', u'HIG'): return_periods = type_settings['return_periods'][level] if isinstance(return_periods, list): if (hazard_period >= return_periods[0] and hazard_period <= return_periods[1]): hazardlevel = HazardLevel.get(level) break else: if hazard_period == return_periods: hazardlevel = HazardLevel.get(level) if ('mask_return_period' in type_settings and hazard_period == type_settings['mask_return_period']): mask = True if hazardlevel is None and not mask: logger.info('{} - No corresponding hazard_level'.format(title)) return False hazard_unit = object['hazard_unit'] if hazard_unit == '': logger.info('{} - hazard_unit is empty'.format(title)) return False if object['srid'] != 'EPSG:4326': logger.info('{} - srid is different from "EPSG:4326"' .format(title)) return False data_update_date = object['data_update_date'] if not data_update_date: logger.warning('{} - data_update_date is empty'.format(title)) # TODO: Restore bypassed constraint to get Volcanic data # return False data_update_date = datetime.now() metadata_update_date = object['metadata_update_date'] if not metadata_update_date: logger.warning('{} - metadata_update_date is empty'.format(title)) # return False metadata_update_date = datetime.now() calculation_method_quality = object['calculation_method_quality'] if not calculation_method_quality: logger.warning('{} - calculation_method_quality is empty' .format(title)) return False calculation_method_quality = int(float(calculation_method_quality)) scientific_quality = object['scientific_quality'] if not scientific_quality: logger.warning('{} - scientific_quality is empty'.format(title)) return False scientific_quality = int(float(scientific_quality)) download_url = object['download_url'] if not download_url: logger.warning('{} - download_url is empty'.format(title)) return False hazardset = DBSession.query(HazardSet).get(hazardset_id) if hazardset is None: logger.info('{} - Create new hazardset {}' .format(title, hazardset_id)) hazardset = HazardSet() hazardset.id = hazardset_id hazardset.hazardtype = hazardtype hazardset.data_lastupdated_date = data_update_date hazardset.metadata_lastupdated_date = metadata_update_date DBSession.add(hazardset) else: # print ' Hazardset {} found'.format(hazardset_id) pass layer = DBSession.query(Layer) \ .filter(Layer.hazardset_id == hazardset_id) if hazardlevel is not None: layer = layer.filter(Layer.hazardlevel_id == hazardlevel.id) if mask: layer = layer.filter(Layer.mask.is_(True)) layer = layer.first() if layer is None: layer = Layer() logger.info('{} - Create new Layer {}'.format(title, title)) DBSession.add(layer) else: if object['id'] == layer.geonode_id: # TODO: metadata change return False if hazard_period > layer.return_period: logger.info('{} - Use preferred return period {}' .format(title, layer.return_period)) return False logger.info('{} - Replace layer for level {}' .format(title, hazardlevel.mnemonic)) layer.hazardset = hazardset layer.hazardlevel = hazardlevel layer.mask = mask layer.return_period = hazard_period layer.hazardunit = hazard_unit layer.data_lastupdated_date = data_update_date layer.metadata_lastupdated_date = metadata_update_date layer.geonode_id = object['id'] layer.download_url = download_url # TODO: retrieve quality attributes layer.calculation_method_quality = calculation_method_quality layer.scientific_quality = scientific_quality layer.local = local DBSession.flush() return True
def process_hazardset(hazardset, force=False): print hazardset.id chrono = datetime.datetime.now() last_percent = 0 level_VLO = HazardLevel.get(u'VLO') if hazardset is None: raise ProcessException('HazardSet {} does not exist.' .format(hazardset.id)) if hazardset.processed: if force: hazardset.processed = False else: raise ProcessException('HazardSet {} has already been processed.' .format(hazardset.id)) # clean previous outputs DBSession.query(Output) \ .filter(Output.hazardset_id == hazardset.id) \ .delete() DBSession.flush() hazardtype = hazardset.hazardtype hazardtype_settings = settings['hazard_types'][hazardtype.mnemonic] thresholds = hazardtype_settings['thresholds'] project = partial( pyproj.transform, pyproj.Proj(init='epsg:3857'), pyproj.Proj(init='epsg:4326')) layers = {} for level in (u'HIG', u'MED', u'LOW'): hazardlevel = HazardLevel.get(level) layer = DBSession.query(Layer) \ .filter(Layer.hazardset_id == hazardset.id) \ .filter(Layer.hazardlevel_id == hazardlevel.id) \ .one() layers[level] = layer with rasterio.drivers(): with rasterio.open(layers['HIG'].path()) as src_hig, \ rasterio.open(layers['MED'].path()) as src_med, \ rasterio.open(layers['LOW'].path()) as src_low: readers = {} readers['HIG'] = src_hig readers['MED'] = src_med readers['LOW'] = src_low polygon_hig = polygonFromBounds(src_hig.bounds) polygon_med = polygonFromBounds(src_med.bounds) polygon_low = polygonFromBounds(src_low.bounds) polygon = cascaded_union(( polygon_hig, polygon_med, polygon_low)) adminlevel_REG = AdminLevelType.get(u'REG') admindivs = DBSession.query(AdministrativeDivision) \ .filter(AdministrativeDivision.leveltype_id == adminlevel_REG.id) \ if hazardset.local: admindivs = admindivs \ .filter( func.ST_Transform(AdministrativeDivision.geom, 4326) .intersects( func.ST_GeomFromText(polygon.wkt, 4326))) \ .filter(func.ST_Intersects( func.ST_Transform(AdministrativeDivision.geom, 4326), func.ST_GeomFromText(polygon.wkt, 4326))) current = 0 outputs = 0 total = admindivs.count() for admindiv in admindivs: # print ' ', admindiv.id, admindiv.code, admindiv.name current += 1 if admindiv.geom is None: print ' ', ('{}-{} has null geometry' .format(admindiv.code, admindiv.name)) continue reprojected = transform( project, to_shape(admindiv.geom)) output = Output() output.hazardset = hazardset output.administrativedivision = admindiv output.hazardlevel = None # TODO: calculate coverage ratio output.coverage_ratio = 100 for level in (u'HIG', u'MED', u'LOW'): layer = layers[level] src = readers[level] if not reprojected.intersects(polygon): continue window = src.window(*reprojected.bounds) data = src.read(1, window=window, masked=True) if data.shape[0] * data.shape[1] == 0: continue threshold = thresholds[layer.hazardunit] positive_data = (data > threshold).astype(rasterio.uint8) division = features.rasterize( ((g, 1) for g in [reprojected]), out_shape=data.shape, transform=src.window_transform(window), all_touched=True) masked = numpy.ma.masked_array(positive_data, mask=~division.astype(bool)) if str(numpy.max(masked)) == str(numpy.ma.masked): break else: if output.hazardlevel is None: output.hazardlevel = level_VLO if numpy.max(masked) > 0: output.hazardlevel = layer.hazardlevel break if output.hazardlevel is not None: # print ' hazardlevel :', output.hazardlevel.mnemonic DBSession.add(output) outputs += 1 percent = int(100.0 * current / total) if percent % 10 == 0 and percent != last_percent: print ' ... processed {}%'.format(percent) last_percent = percent pass hazardset.processed = True DBSession.flush() transaction.commit() print ('Successfully processed {} divisions, {} outputs generated in {}' .format(total, outputs, datetime.datetime.now() - chrono))