deppen8/linked-scatterplot-table-download-poc

PoC for linking a Bokeh scatterplot selection to a table with a download button

Bokeh scatter linked to table with download - proof of concept

This repo contains a minimal proof-of-concept for linking a scatterplot selection to a table in Bokeh. It also contains a "Download" button that triggers download.js which will save the current table as a CSV file.

To run locally:

$ cd path/to/repo
$ bokeh serve --show app/

Requirements

Created with bokeh==2.3.0 and Python 3.9.2. Not tested with other versions.

linked-scatterplot-table-download-poc's Issues

update table when resetting to all points selected

Hi, I've been looking for code like this for a while now - thanks for sharing! I am adapting this for a standalone bokeh html file. When I un-select a subset of points so the graph goes back to "all" points, the cb_obj.indices is an empty list. So I updated the source1.selected.js_on_change callback like so:

for (var i = 0; i < inds.length; i++) {
  d2['x'].push(d1['x'][inds[i]])
  d2['y'].push(d1['y'][inds[i]])
  d2['label'].push(d1['label'][inds[i]])
}

becomes

if (inds.length==0) {
  for (var i = 0; i < d1['x'].length; i++) {
    d2['x'].push(d1['x'][i])
    d2['y'].push(d1['y'][i])
    d2['label'].push(d1['label'][i])
  }
} else {
  for (var i = 0; i < inds.length; i++) {
    d2['x'].push(d1['x'][inds[i]])
    d2['y'].push(d1['y'][inds[i]])
    d2['label'].push(d1['label'][inds[i]])
  }
}