Export your work¶
Everything you do in BigClust — recomputed embeddings, clusterings, refreshed meta data, manual refinements — lives in the session. This page is how it leaves.
What is where¶
| You want | Do this | You get |
|---|---|---|
| The selected IDs | Selection → Copy to Clipboard → IDs (Cmd+C) | Newline-separated IDs |
| The selected rows | Selection → Copy to Clipboard → Meta Data | The meta rows for the selection |
| The whole meta table | Export → Meta Data → To CSV | <project>_meta_data.csv |
| The whole meta table, pasteable | Export → Meta Data → To Clipboard | Tab-separated, ready for a spreadsheet |
| A shareable interactive plot | Export → Embedding → To Plotly | A single self-contained HTML file |
| The same, with side panels | Export → Embedding → To Dashboard | HTML with selection-linked panels |
| A cluster assignment | Cluster tab → Export | CSV with id and bigclust_cluster |
| A filtered meta subset | Meta Data Explorer → Save | selected_meta.csv |
| A feature ranking | Feature Comparison → export | feature_ranking.csv |
| Everything, reopenable | Export → Project | A complete project directory |
A full project snapshot¶
Export → Project is the one that matters. Pick a destination folder and
BigClust writes a complete, self-contained project there: the info file as it
currently stands in memory, plus every data file it references, downloading
remote ones as it goes.
That "as it currently stands in memory" is the point. A snapshot captures:
- Meta data you refreshed from a backend or merged from a file, which otherwise vanishes at exit.
- Meta source definitions you saved, and their
last_updatedstamps. - The project as a whole, so a slow remote project becomes a fast local one.
Open the exported directory from then on and you pick up where you left off.
The neuroglancer block is preserved verbatim rather than being downloaded —
meshes and segmentation stay remote, which is what keeps a snapshot to a sensible
size.
If the destination exists and is not empty, you are asked before anything is overwritten. A download that fails partway removes its partial file rather than leaving a truncated one behind.
A snapshot does not capture everything
Cluster assignments and recomputed embeddings are not included — they are session state, not project files. Export those separately (Cluster tab → Export) if you want them.
Sharing a plot with someone who doesn't have BigClust¶
Export → Embedding → To Plotly writes one HTML file containing your current embedding, with the colouring and labels you have set up. It opens in any browser, zooms and pans, and shows hover info. Plotly itself is loaded from a CDN, so the file is small but needs a network connection to render.
To Dashboard adds panels linked to the selection, so the recipient can pick points and see what they are.
This is the right way to put a figure in a slide deck or send a colleague something to look at. It is not a substitute for sharing the project directory — the HTML is a static picture of one embedding, with no data behind it.
Reproducing a session¶
There is no single "save my session" file. To pick up exactly where you left off, the reliable combination is:
- Export → Project — the data, including any refreshed meta.
- Cluster tab → Export — the cluster assignments, reloadable with the folder button next to Run.
- Note the Embeddings tab settings you used, including the
Random seed:. With the seed fixed, a recompute is reproducible.
Next: work in several views.