-
Notifications
You must be signed in to change notification settings - Fork 2
Annotation Framework Client
Using the client | Documentation
The following is adapted from FrameworkClientExamples.
Get an auth token from either of the following:
They both point to the same server. First one is provided for compatibilty with annotationframeworkclient.
The path prefix /<group> (/wclee in this example) is currently used for directing users to relevant resources in the backend.
In the future it will also be used to manage permissions for <group>.
For invalidating all existing tokens (for security), visit https://api.zetta.ai/auth/logout. This will discard all tokens in the system for the given user.
from annotationframeworkclient import FrameworkClient
auth_token = "<your_token>"
datastack_name = 'vnc_v0' # from https://api.zetta.ai/wclee/info/
client = FrameworkClient(
datastack_name,
server_address = "https://api.zetta.ai/wclee",
auth_token = auth_token
)
print(f"The image source is: {client.info.image_source()}")
> The image source is: precomputed://gs://zetta_lee_fly_vnc_001_precomputed/vnc1_full_v3align_2/realigned_v1
example_id = 495722871693421950
example_state = client.state.get_state_json(example_id)
example_state['layers'][0]['name'] = 'renamed'
new_id = client.state.upload_state_json(example_state)
new_state = client.state.get_state_json(new_id)
print(new_state['layers'][0]['name'])
> 'renamed'
url = client.state.build_neuroglancer_url(state_id=new_id)
print(url)
> https://neuromancer-seung-import.appspot.com/?json_url=https://api.zetta.ai/wclee/nglstate/api/v1/<new_state_id>
Available datastacks are listed on the InfoService portal.
New datastacks can be created by admins at admin portal. Permissiongroup and Tablemapping tabs on the page are not relevant at this time.
sv_id = 72835911972360756
client.chunkedgraph.get_root_id(supervoxel_id=sv_id)
root_id = 648518346677614825
leaves = client.chunkedgraph.get_leaves(root_id)
print(len(leaves))
> 52863
table_name = "test_table" # fails if already exists
client.annotation.create_table(
table_name = table_name,
schema_name = "microns_func_coreg",
description = "this is a test"
)
new_data = {
"type": "microns_func_coreg",
"pt": { "position": [1,2,3] },
"func_id": 0
}
client.annotation.post_annotation(table_name=table_name, data=[new_data])
The new table can be viewed at Annotation Engine portal.
example_schema = client.schema.schema_definition('microns_func_coreg')
example_schema['definitions']['FunctionalCoregistration']
This will output:
{'properties': {'func_id': {'title': 'func_id',
'type': 'number',
'format': 'integer',
'description': 'functional cell ID'},
'pt': {'type': 'object',
'$ref': '#/definitions/BoundSpatialPoint',
'description': 'location of cell body of functional cell'},
'valid': {'title': 'valid',
'type': ['boolean', 'null'],
'default': False,
'description': 'is this annotation valid'}},
'required': ['func_id', 'pt'],
'type': 'object',
'additionalProperties': True}