Skip to content

README.md

Latest commit

 

History

History
118 lines (84 loc) · 3.41 KB

README.md

File metadata and controls

118 lines (84 loc) · 3.41 KB

Worm Circuitry Explorer

This repo is now a static Bun + Vite app for exploring the C. elegans connectome in the browser.

There is no Flask app, no Neo4j runtime, and no Python build step anymore. The graph dataset is generated at build time as public/data/connectome.v1.json, then loaded by the frontend as a static asset.

Requirements

  • Bun 1.3+
  • Node.js 20+
    Bun is the package manager, but the data builder and tests still run on Node for compatibility.

Local development

bun install
bun run build:data
bun run dev

Open the local Vite URL, usually http://localhost:5173.

Useful commands:

bun run test
bun run build
bun run preview

Notes:

  • bun run build:data regenerates public/data/connectome.v1.json.
  • bun run test and bun run build both regenerate the dataset first.
  • public/data/connectome.v1.json is ignored in git because it is a generated artifact.

Data pipeline

The data build entrypoint is:

  • scripts/build-data.mjs

The actual builder logic is here:

  • scripts/lib/build-connectome.mjs

The builder currently reads these raw inputs:

  • data/ChenVarshney/NeuronType.xls
  • data/ChenVarshney/NeuronFixedPoints.xls
  • data/ChenVarshney/NeuronConnect.csv
  • data/DynamicConnectome/celegans277/celegans277labels.csv
  • data/DynamicConnectome/celegans277/celegans277positions.csv
  • data/WormWeb/name_neurons.txt
  • data/Self/Sensors.tsv

The generated dataset contains:

  • Canonical nodes[] and edges[]
  • Group and receptor indexes
  • Presets used by the UI
  • Summary counts for the baseline graph

The builder preserves the legacy project rules that are still needed:

  • merge connectivity, type, group, sensor, and layout inputs
  • fill missing coordinates from symmetric left/right partner neurons
  • compute inD, outD, and D
  • include muscle nodes and NMJ edges from the muscle workbook
  • dedupe reciprocal electrical junction pairs for the initial full graph

Data provenance

Primary currently-used sources:

  • data/ChenVarshney/ Connectivity/type/muscle tables used by the current build. The folder includes its own SOURCES.txt.
  • data/WormWeb/ Neuron name/group/type mapping used by the current build. The folder includes its own SOURCES.txt.
  • data/Self/Sensors.tsv Local curated sensor annotations carried over from the legacy project.
  • data/DynamicConnectome/celegans277/ Layout labels and coordinates used for node positioning in the static app.

Historical reference data retained in the repo but not read by the current builder:

  • data/OpenWorm/
  • data/WormAtlas/

Those folders are still useful as provenance/reference material, but they are not part of the modernized build path.

Repo layout

  • index.html Static app shell.
  • src/ Modern browser code and app styles.
  • scripts/ Build-time data generation.
  • tests/ Node-based tests for the builder, query engine, and exporters.
  • public/static/ Small set of legacy CSS/font/image assets still used by the current static app.
  • data/ Raw source datasets and reference material.

What was removed

The old server-side/runtime stack has been removed from the working tree:

  • Flask templates
  • Python app/helper scripts
  • Neo4j-specific runtime code
  • legacy Sigma renderer assets
  • duplicated old static/ tree
  • generated GraphML export artifacts

If you need the original implementation details, use git history rather than keeping those files in the active tree.