Lasso group of points

[meganap]

draw a lasso -get sample IDs, figure out of those have anything in common -get sample IDs, redo pcoa with only those samples

[antgonza]

+1

This functionality will be great + adding click on a sphere to find the name of the sample. Note, to redo PCoA will be out of the scope of this project as we have no "processing" backbone but should be in https://github.com/qiime/evident

[ElDeveloper]

We could just rewrite all those algorithms in javascript, right?

---

Here they implement lasso/group selection:

http://blog.tempt3d.com/2013/11/marquee-selection-with-threejs.html

[rob-knight]

Redoing pcoa would be REALLY good and the code is not hard. This is an extremely common analysis demand as you know, especially when you find you have to exclude mislabeled points etc.

Reassigning category labels and outputting new/corrected mapping file would also be really good.

On Aug 2, 2013, at 7:23 AM, Yoshiki Vázquez Baeza notifications@github.com<mailto:notifications@github.com> wrote:

We could just rewrite all those algorithms in javascript, right?

— Reply to this email directly or view it on GitHubhttps://github.com/qiime/emperor/issues/153#issuecomment-22005128.

[antgonza]

This is a list of concerns in case we want to do this in Emperor vs Evident:

• JavaScript floating number implementations in each browser and the 32/64bits representations of the numbers: http://stackoverflow.com/questions/588004/is-javascripts-floating-point-math-broken http://stackoverflow.com/questions/1458633/elegant-workaround-for-javascript-floating-point-number-problem
• The time to process each PCoA, as sometimes this can take a lot of time, it will mean that the browser will send a kill signal. This may be avoid by the new implementations that Sophie, Sam and I are working on.
• We will need to store the distance matrices, which are huge, on JavaScript. This is not a problem if using Evident.

[rob-knight]

That's a good point, maybe focus should be on doing this via evident not direct in emperor.

On Aug 3, 2013, at 10:31 AM, "Antonio Gonzalez" notifications@github.com<mailto:notifications@github.com> wrote:

This is a list of concerns in case we want to do this in Emperor vs Evident:

• JavaScript floating number implementations in each browser and the 32/64bits representations of the numbers: http://stackoverflow.com/questions/588004/is-javascripts-floating-point-math-broken http://stackoverflow.com/questions/1458633/elegant-workaround-for-javascript-floating-point-number-problem
• The time to process each PCoA, as sometimes this can take a lot of time, it will mean that the browser will send a kill signal. This may be avoid by the new implementations that Sophie, Sam and I are working on.
• We will need to store the distance matrices, which are huge, on JavaScript. This is not a problem if using Evident.

— Reply to this email directly or view it on GitHubhttps://github.com/qiime/emperor/issues/153#issuecomment-22055722.