ngregister

Start a local Neuroglancer instance with pre-configured shortcuts for custom registration actions:

• Translate the active layer at the cursor 🖱️ to the center 🎯 of the view with <t>.
• Place a translation landmark 📍 at the cursor 🖱️ (in the fixed volume) with <y>,
  and translate the active layer at the position of the cursor 🖱️ to the landmark 📍 with <shift+t>.
  You can delete manually the "__LANDMARK__" layer when done.
• Rotate from the center 🎯 of the view around the absolute axis $z$ (resp. $x$ and $y$) clockwise 🔃 with <i> (resp. <j> and <k>).
  Rotate counter-clockwise 🔄 with <shift+i>, <shift+j>, <shift+k>.

How to use

• Install dependencies using a virtual environment and the ./requirements.txt file.

Tip

To support the most recent Neuroglancer states which load resources using, e.g. gs://my-bucket/data.ome.zarr|zarr:, you will need the development version of Neuroglancer, that can be installed using:

pip install git+https://github.com/google/neuroglancer.git

You will require a build environment, including node and Python.h. To install node on your system, use nvm. If you use uv (uv pip ...), you should pick and install a uv-managed version of Python, as explained here.

• Start an interactive Python session with (blank state):

python -i ngregister.py

• Or start with a previous state from either a JSON file or a Neuroglancer URL:

python -i ngregister.py --url "https://neuroglancer-demo.appspot.com/#!..."

• Either save manually the JSON file from the web browser or get the Neuroglancer URL once you quit the interactive Python session.

Other related projects

As of neuroscales/ngtools#42, these features have been implemented in neuroscales/ngtools, which handles more complex use cases. HiPCTProject/ngregister can still be used as a standalone.

Warning

Keybindings are not yet uniformized between the two tools.