Closed vincerubinetti closed 5 months ago
Name | Link |
---|---|
Latest commit | ca6de0124b08c54d012d739df28f38d4b0be1730 |
Latest deploy log | https://app.netlify.com/sites/gene-plexus/deploys/661d5b923addcc00085ea403 |
Deploy Preview | https://deploy-preview-26--gene-plexus.netlify.app |
Preview on mobile | Toggle QR Code...Use your smartphone camera to open QR code link. |
To edit notification comments on pull requests, go to your Netlify site configuration.
I forgot to open this PR yesterday. I wanted to open it today so I didn't forget and to get it off my mind, but I'd recommend waiting until Monday to review.
@ChristopherMancuso You can run Faisal's run_local.sh
script to test the entire stack out locally.
I was able to look at this with run_local.sh
and look over the function commit updates.
It is nice to have pulled the info not he function input and outputs out into a readme. For that though, there is only a README.md
in the ml function folder and noting for the id conversion. And in the ML README there are two sections called gpz_convert_ids
. If Im following what is going on, should the top part in the ML README.md be put in a READ me in the id_convert
folder?
I mistakenly had that in the ml folder, see the last commit where I move it up to the folder above it.
Please also make sure the notes I wrote are correct. There are probably some gaps that could be filled, like a top level comment above df_probs
and df_sim
.
I just checked over that readme file and committed some updates.
Also, I did try training in one species and testing in another and that didn't work. Probably not for this PR to worry about, right?
Also, I did try training in one species and testing in another and that didn't work
How do you mean? Did you do this through the API directly? The frontend keeps them the same species for now, as we discussed.
I did it on the front end in run_local.sh
. I had forgot you said the API was restricting to the species being the same in each of the buttons for now.
I added one more commit, please check it. I included the inputs for /ml
back in the output. This way, the results.json
that we get back from the server is a "complete package", and when a user saves it as json and reuploads it, it will populate the "inputs" section in the UI and remind them what the job parameters were.
Note: the deploy preview here won't work since it will be hitting the live cloud functions which are not updated with the changes in this PR yet. If you want to test just the frontend on the Netlify preview, put
?mock=true
at the end of the url and it will use fake data instead of a real api.