Skip to content

Tutorials

Six of them, in order. Each is short and stands alone, but they build.

Everything shown is a real output from a real run — if a number looks oddly specific, that is because it is.

1. Your first query

Open a dataset, pull an edge list, and learn to read the provenance stamped on every frame you get back.

2. Selecting neurons

The ids() mini-language: IDs, types, regexes, annotation columns, and how to find out what a dataset will even accept.

3. Connectivity

Edges, adjacency matrices, individual synapses, per-ROI breakdowns and predicted transmitters.

4. Morphology

Skeletons and meshes as navis neurons, and where they come from on each backend.

5. Versions and root IDs

The one that will save you. Root IDs change when neurons are edited; this is how you stop that silently corrupting your analysis.

6. One script, two datasets

Write it once, run it on a fly and on a mouse. What genuinely transfers, and what does not.

Before you start

You will need working credentials — see Credentials. The tutorials use FlyWire and hemibrain, both of which need only a token you can get yourself in about a minute.

import connecto as cn

cn.auth_status()   # should say OK for cave and neuprint