dvidtools

Python tools to fetch data from DVID servers.

Find the documentation here.

Want to query a neuPrint server instead? Check out neuprint-python.

What can dvidtools do for you?

• get/set user bookmarks
• get/set neuron annotations (names)
• download precomputed meshes, skeletons (SWCs) and ROIs
• generate meshes or skeletons from scratch
• get basic neuron info (# of voxels/synapses)
• fetch synapses
• fetch connectivity (adjacency matrix, connectivity table)
• retrieve labels (TODO, to split, etc)
• map positions to body IDs
• detect potential open ends (based on a script by Stephen Plaza)

Install

Make sure you have Python 3 (3.8 or later), pip. Then run this:

pip3 install dvidtools

To install the dev version straight from Github:

pip3 install git+https://github.com/flyconnectome/dvid_tools@master

Optional dependencies

Necessary dependencies will be installed automatically.

If you plan to use the tip detector with classifier-derived confidence, you will also need sciki-learn:

pip3 install scikit-learn

For from-scratch skeletonization you need to install skeletor:

pip3 install skeletor

Examples

Please see the documentation for examples.

Testing

For testing you need to have two environment variables set: DVID_TEST_SERVER and DVID_TEST_NODE. These should point to a DVID server/node that contain the Janelia hemibrain dataset. Then run:

$ pytest -v