Minor suggestions and improvements

[xileF1337]

Hi Anda and Stefan,

Thanks for your work on the drforna code base. I played around with it a bit and it looks promising. I have a question and a small list of suggestions which, in my oppinion, woud improve your tool even more. Of course, they reflect my personal view and if you don't agree with me, feel free to simply close this issue.

The question first: is there a way to pass a local file to be loaded directly with the URL? So I could, for example, write the simulation output of some BarMap run to some temp file and spit out a link for the user to click on that immediately loads to performed simulation?

The list of suggested improvements:

• allow resizing and even hiding the structure plotting pane
  • resizing would allow to give more space to the table if needed
  • hiding would massively improve the performance when only analyzing the table
• add a customizable population filter
  • only show states with population >= x, where x can be adjusted by the user
  • would improve performance as well as visual clarity
• enlarge individual structure plots when clicking them
  • currently, small structures are barely visible. yeah, one could look at the table, but
    thats not as intuitive I think
  • I think of some pop-up / hovering window as with usual image galleries
• add visible feedback for button clicks
  • currently, exactly nothing happens when clicking -- have I clicked the button properly or did I miss it? Is the app hanging? Give me some feedback, please!
• reduce overall "jumpyness"
  • the timeline jumps when moving the cursor over it, on my screen between time 2.0 and 2.5 of the example ("grow") seq simulation
  • don't make the word "Structure" in the table header move during simulation
  • plot the values with a fixed format (e.g. 4 digits for ID, 4 decimal digits for occupancies etc.). There is still an issue open (Visual glitch: Structure list vs. page scrolling  #26 [edit: corrected issue link]) concerning the IDs, so I guess this is known and considered already.
  • improve the structure list as discussed in Improve robustness of the input parser. #21
• small glitch: there is a strange peak on top of the 'A' when displaying the first nucleotides (which are then drawn really huge)

[bad-ants-fleet]

Hi Felix,
I see this issue only now, sorry. To your first question, I do not think that this is possible (for security reasons). You would need to have a local server and handle the file stuff from there. I might be wrong on this though. If you learn about a solution, let me know. It sounds like a nice feature.

Some of your feedback has been resolved, e.g. the population filter, enlarging, etc. Maybe you can take another look to see what you would still improve?

I am aware of your suggestions concerning "ID column jumping". However, I don't like the solution of reserving the space for all future IDs to come. Some Kinfold simulations will end up with IDs in the 10-thousands. The dead space in the beginning does not look nice. On the other hand, jumping seems no problem for me, so this "issue" is currently on hold.

Rubstness of the input parser would indeed be nice, and probably not difficult to do. Maybe I'll do it, or maybe someone else jumps in fixing this? There is a "parse_drf" function that can easily be replaced by something more robust.