-
Notifications
You must be signed in to change notification settings - Fork 2
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Could you specify the details to install foldy in windows? #18
Comments
Hi Jiyao, Thank you for asking, I would love to help get the development environment working on Windows. Aside: I rewrote this section of the README for clarify, but it doesn't really address how to use Windows / WSL yet. I have not used WSL before, but from what I read, it seems like the WSL file system is pretty separate from the Windows environment. I would advocate doing everything in WSL. You can do the I also filed #19 to run the frontend in docker, so that you don't have to deal with conda and NPM. If it is hard to get everything running in WSL, I can get to #19 sooner. Best, |
Hi Jiyao, #19 is now complete, and the instructions have been updated. Now you just need git and docker to run Foldy locally, so I believe you can use either Windows Powershell or WSL. The instructions are:
Please let me know how it goes! |
@Jacoberts This is very helpful. I installed docker as you specified before and installed git as described at https://gitforwindows.org/.Then I ran the three steps you specified in window power shell. Finally I submitted my job at my local site: http://localhost:3000/. I thought my local computer is used to predict the structure. But it said "queued". Where the computation is done? |
Hi Jiyao, We only provide full deployment instructions for the cloud. I'm so sorry for this confusion, this is not clear from the README, which I'll update to clarify this point. The "local / development" deployment is meant to evaluate and develop the frontend and backend - it does not run the workers (alphafold, diffdock, etc). All data that is exposed (eg, the structure you are seeing) is precomputed, just for testing purposes. This is noted briefly in the README but could be more emphasized. We could certainly add support local deployment for folks with big GPUs and big disks, and I'm starting to think that would improve the value of Foldy! The demands for a local deployment are twofold: GPUs and large databases (~3TB in total). We originally developed for the user that didn't have good hardware, so we targeted cloud-first. But for users who have that hardware & disk space, we could make the local deployment fully functional. I think it would be a small change to make Foldy work on a large desktop, and I could evaluate it this week. Would that be useful to you? Best, |
Hi Jacob, I did notice that users can not run Foldy in the local environment as mentioned in README. 3TB is a little more than a regular desktop. I was trying to set up the production environment and found it non-trivial to set it up. As you mentioned, it will greatly increase the value of Foldy if users can run Foldy in their local desktop with enough space and GPU. Many computational labs have such ability. By the way, are you going to enable users to run ESMFold in Foldy in the near future? Users can get most AlphaFold structures from AlphaFOld UniProt databse, but may be interested in getting ESMFold structures. Thanks a lot for all your help! |
Hi Jiyao, Adding support for ESMFoldy is a great idea. I especially like it because it is easier to install on a standard machine. Supporting ESMFold could be a great way to make the local deployment of Foldy useful. I've created #20 for that feature request. Best, |
I am trying to install the local deployment in my pc since my MacOS is version 10 and can not install docker. After I installed docker, I installed WSL. After I install conda, I can run conda commands.
My question is whether I should run git and pip in linux/WSL, or conda command window.
The text was updated successfully, but these errors were encountered: