Todo

Data Handling

In a typical use case, the user has obtained simulation data from and wishes to visualize this The data obtained can be of data corresponding to currents from every compartment of every cell or partial data corresponding to voltage traces from soma. The data can also be discrete data like spike events corresponding to the cell. Typical data sizes that we would want to visualize would be of the order of 10’s of Gb. Moogli must serve this purpose. Some of the issues for this are as follows.

1. Parallel reading of the data.
2. Reading chunks of data - google maps like buffering beforehand. But only if the file sizes are large.
3. Transposing the data - for visualization we’d need data values at each instant.
4. Digitizing and reading off a colormap
5. Binning and sliding window, in some cases user may want to bin chunks of data, or may want to visualize only the max in a sliding window.

File formats

To address the issue of large data sets generated - it would be best to use an existing standard format. The following can be considered and must be explored in finer detail

1. neohdf5
2. neuronvisio’s hdf5 - custom hierarchy
3. neurohdf
4. custom format inspired by moose moogli or neuroconstruct.
5. NetCDF etc?

Color mapping

1. Should use standard colormaps - perhaps use matplotlib cm. The color bar must be generated at the bottom of the visualizer - also must indicate ticks and labels.
2. There must be a schema to fade out the color slowly.
3. Alpha can also be used as spike event visualization in case of many neurons cluttering the scene.

Picking

It is in the interest to get this to work as Moogli can be the potential default visualizer for any existing simulators. There are issues to fix in this regard, as using VBO’s means, having to do colorpicking instead of ray tracing or glName - which happens to be slow. See issues.

Demos

Scripts

1. Moose demo
2. Neuron demo
3. Nest demo [point neurons depicted as points] - perhaps network connections can be shown.
4. Brian demo [point neurons depicted as points]
5. LFPy

Videos

1. Demonstrating simple case of taking snapshots of the scenes of interest.
2. How to make animation videos - of movies. Camera movements recording formats etc.