Add new hierachical clustering tool based on scipy #6214
Conversation
|
One thing this version does not support atm is multiple heights or n_clusters values, which, on the command line would result in multi-column tabular output with the cluster assignments for all different settings. |
|
We should maybe have column names for the cluster assignment output. Let me add this still before merging. |
| @@ -0,0 +1,6 @@ | |||
| sample cluster_id_h18.0 | |||
There was a problem hiding this comment.
why is this here h18.0 a float and below an integer? Is anything wired going on?
Any both files are identical besides the header...
There was a problem hiding this comment.
h18.0 is the height at which the tree got cut in order to produce the cluster IDs below, and height can be a float.
The results are otherwise identical because I chose --height and -n so that both tests give the same result. So before I added now the column names only a single test file was needed.
* Add new hierachical clustering tool * Describe input format in tool help * Add .shed.yml * Add edam information * Fix flake8 issues * Add note about test data origin * Try to fix tests * Update tools/clustering_from_distmat/.shed.yml * Write column names to cluster assignment output --------- Co-authored-by: Björn Grüning <[email protected]>
A tool like this was requested by @dbrites
FOR CONTRIBUTOR: