Feedback on Similarity Metrics function/UI

[nsheff]

Some feedback on the latest update.

It's really cool! I'm excited about this. A few suggestions for improvement:

• "Seqcol Similiarity Metrics" -- Can we think of a more specific or creative name?
• It is too slow. When someone runs a comparison, even of a seqcol with only 3 sequences, it actually freezes the entire server for 1-2 seconds. This will not be sustainable.
  1. why does the whole server seem to lock up when this is happening? Is there some part of the process that's not working async or something?
  2. even if it's using async correctly, this seems to be too much computation. What is the bottleneck? Is it possible to improve that?
     I am guessing that it's something simple and inefficient we are doing, because I don't think this should take that long. Try doing this on a larger seqcol. It should:
  3. happen async (so it doesn't take down other server operations in the meantime)
  4. Take much less computation overall
     Those are two independent things and both need to be addressed.

Visualization:

• The Strip Plot only seems to be showing one of the 5 metrics (names), for most of the sequences. Why? This probably needs a bit of information/explanation.
• In the strip plot, for many comparisons the points are all on top of each other. Is there a way to make them a bit easier to see in that case,
• instead of 'custom1' how about 'query_seqcol' ?
• The entire plot is a bit too large for my liking. I have to scroll to see it. Can you reduce the total size of the plot, espeically the height, by ~30%-50%?
• Overall the plots (all of them) have too large of Plot/whitespace and too small of Text/axis/legends
• In the summary table, make it more clear that it's a link. Either a link to a comparison report visual. or a blue hyperlink, or something.
• In the summary table, it's not clear why not all values are filled in. for the example one, most of them are empty.
• Change "Selected Digest" and "Compared Digest" to "seqcol A" and "seqcol B" -- I think this is more consistent with the terms we're using elsewhere.
• "Summary Table" -> "Seqcol Comparison Summary Table"
• SCIM says "b": "POSTed seqcol" and then has link to: https://dev.refget.pages.dev/collection/POSTed%20seqcol -- which doesn't make sense
  • can we actually use the digest of this?
  • If not, should probably at least disable that link.

Other semi-related minor issue I found:

• https://dev.refget.pages.dev/collection/POSTed%20seqcol showS:
  "Error fetching data from https://seqcolapi.databio.org/collection/POSTed seqcol?collated=false:
  Is the API service operating correctly at https://seqcolapi.databio.org?"
  Instead I think it should first say: "Unable to find a collection with digest; POSTed seqcol" -- to make it clear what this is trying to do.

[nsheff]

Any update on these issues for performance and visualization, @donaldcampbelljr and @sanghoonio ?

[donaldcampbelljr]

Draft PR on dev branch addresses some of the performance issues, adding human readable names, and ensuring all attributes are returned from the comparison (instead of null):
#45
Original PR for performance fix:
#43