A simple wrapper for interacting with the ArcGIS Online REST API.
This project is geared towards developers who want to get data into the ArcGIS Online as easily as possible. The REST API can be pretty confusing if you aren't familiar with ArcGIS jargon and the ArcGIS API for Python is cumbersome for cloud-based programs, so we created this simple wrapper to abstract some of the difficulties with creating resources and adding data.
First you will need an ArcGIS Online account. If you plan on using token authentication, rather than a username and password, you will need an app registered on the Developer portal.
Then you will need to know a thing or two about ArcGIS.
- "Feature services allow you to serve features over the Internet and provide the symbology to use when displaying the features."
- A feature service can store many feature layers.
- "A feature layer is a grouping of similar geographic features, for example, buildings, parcels, cities, roads, and earthquake epicenters."
- A feature layer can store many geometric features.
- "Feature classes are collections of common features, each having the same spatial representation, such as points, lines, or polygons, and a common set of attribute columns."
pip install simple-arcgis-wrapper
import simple_arcgis_wrapper as saw
You will need a registered app and tokens obtained through the OAuth flow. Check out this link to learn more about setting up OAuth.
api = saw.ArcgisAPI(
access_token='ACCESS_TOKEN', # access token obtained from user
refresh_token='REFRESH_TOKEN', # refresh token obtained from user
username='USERNAME', # username obtained from user
client_id='CLIENT_ID' # your OAuth app's client ID
)
If you just want to add data to your own account you can use this authentication scheme. Use this for one-off tasks, as this scheme will only be valid for 1 hour. Use OAuth tokens for longer-lived operations.
api = saw.ArcgisAPI.fromusernamepassword(
username='username',
password='password'
)
service = api.services.create_feature_service('NAME', 'DESCRIPTION')
# service is a FeatureService object
print(service.id, service.name, service.url)
A feature layer stores your features, so you need to define the layer type and any additional fields.
fields = saw.fields.Fields()
fields.add_field('Date', saw.fields.DateField)
fields.add_field('Name', saw.fields.StringField)
fields.add_field('Altitude', saw.fields.DoubleField)
layer = api.services.create_feature_layer(
layer_type='point', # point, line, or polygon
name='NAME',
description='DESCRIPTION',
feature_service_url=service.url,
fields=fields,
x_min=10.0, y_min=10.0, # min bounding box parameters
x_max=20.0, y_max=20.0, # max bounding box parameters
wkid=4326 # well-known ID spatial reference
)
# layer is a FeatureLayer object
print(layer.id, layer.name, layer.url)
A table is like a feature layer except it doesn't have coordinates.
fields = saw.fields.Fields()
fields.add_field('Date', saw.fields.DateField)
fields.add_field('Name', saw.fields.StringField)
fields.add_field('Email', saw.fields.StringField)
table = api.services.create_table(
name="TABLE_NAME",
description="TABLE_DESCRIPTION",
feature_service_url=feature_service.url,
fields=fields
)
# table is a Table object
print(table.id, table.name, table.url)
# attribute keys must match the layer's fields
attributes = {
'Date': '2020-01-01 15:30:45',
'Name': 'John Doe',
'Altitude': 12.5
}
success = api.services.add_point(
lon=10.0,
lat=20.0,
attributes=attributes
layer_id=layer.id,
feature_service_url=service.url
)
print(success) # True or False
attributes = {
"Date": "2020-01-01 15:30:45",
"Name": "John Doe",
"DeviceId": "abc123",
}
p1 = {
'lon': 10.0, 'lat': 20.0,
'Date': "2020-01-01 12:12:12",
'Name': 'John Doe',
'DeviceId': 'abc123'
}
p2 = {
'lon': 10.0, 'lat': 20.0,
'Date': "2020-01-01 12:12:12",
'Name': 'John Doe',
'DeviceId': 'abc123'
}
adds = api.services.add_points(
points=[p1, p2],
layer_id=layer.id,
feature_service_url=service.url
)
# adds is a dict where key is object ID and value is success
for object_id, success in adds.items():
print(object_id, success)
Get a feature service by passing the exact name of the service.
other_service = self.api.services.get_feature_service('OTHER_NAME')
Get a layer from a feature service by looking up it's ID or exact name.
layer_by_id = api.services.get_feature_layer(service.url, layer_id=0)
layer_by_name = api.services.get_feature_layer(service.url, layer_name="other layer")
You can get features from a feature layer by passing an SQL 92 formatted where clause as described here. Specify the attributes you want returned with the out_fields argument.
Pro tip: return all features with where="1=1"
Only point features supported right now.
point_features = api.services.get_features(
where="DeviceId = 'abc123'",
layer_id=layer.id,
feature_service_url=service.url,
out_fields=['OBJECTID']
)
# point is a PointFeature object
point = point_features[0]
print(point.id, point.x, point.y)
Only updating the service's title supported right now.
success = api.services.update_feature_service(
feature_service_id=service.id,
title="New Title"
)
print(success) # True or False
Update multiple features by passing a list of tuples which contain an object ID, attributes to update and geometry respectively. If you want to update features based on a where clause, first get the features you want as described above.
If you are not updating attributes or geometry, pass None.
updates_list = [
(0, {"Name": "John Doe II"}, {"x": 10.1, "y": 20.1}),
(1, None, {"x": 10.3, "y": 20.2}),
(1, {"Name": "John Doe II"}, None),
]
updates = api.services.update_features(
updates=updates_list,
layer_id=layer.id,
feature_service_url=service.url
)
for object_id, success in updates.items():
print(object_id, success)
Update multiple table rows by passing a list of tuples which contain an object ID and attributes to update. You can use a where clause to filter the affected rows.
updates_list = [
(0, {"Name": "John Doe II", "Email": "johndoe2@example.com"}),
(1, {"Name": "John Doe II", "Email": "johndoe2@example.com"})
]
updates = api.services.update_table_rows(
updates=updates_list,
table_id=table.id,
feature_service_url=service.url
)
for object_id, success in updates.items():
print(object_id, success)
Delete features by passing an SQL 92 where clause.
deletes = api.services.delete_features(
where="DeviceId = 'abc123'",
layer_id=layer.id,
feature_service_url=service.url
)
for object_id, success in deletes.items():
print(object_id, success)
api.services.delete_feature_layers([layer.id], service.url)
api.services.delete_feature_service(service.id)
Invalid arguments to ArcGIS may result in an error. You can catch them with ArcGISException which includes the message returned from ArcGIS.
try:
# try to create a duplicate feature service
api.services.create_feature_service('NAME', 'DESCRIPTION')
except saw.exceptions.ArcGISException as e:
print(e)
Before testing, configure the following environment variables:
- ARCGIS_ACCESS_TOKEN
- ARCGIS_REFRESH_TOKEN
- ARCGIS_CLIENT_ID
- ARCGIS_USERNAME
- ARCGIS_PASSWORD
By running the tests you will incur a small charge to your account.
python -m unittest tests/test*.py
- Casey Slaught - Lead developer - Caracal
This project is licensed under the MIT License - see the LICENSE for details