search

jonorthwash / ud-annotatrix

GNU General Public License v3.0
61 stars 49 forks source link

GSoC 2019 Progress Updates #372

Open jonorthwash opened 5 years ago

jonorthwash commented 5 years ago

This issue is for tracking progress throughout the rest of GSoC 2019.

jonorthwash commented 5 years ago

For #356, push the latest commits that fix the indentation and spelling issues, and we'll close the PR.

Going forward, let's do one PR per GitHub issue.

jonorthwash commented 5 years ago

GitHub integration is one of the top priorities for the summer. Let's focus on that, and fixing smaller regressions and minor annoyances.

jonorthwash commented 5 years ago

Can you open a GitHub issue for each thing in the roadmap? #357 And then we can close that issue.

jonorthwash commented 5 years ago

@ftyers and @maryszmary, I'm adding you as participants in this issue.

maryszmary commented 5 years ago

Just after our online-meetup I discovered (@yaskevich told me), that corpora, when uploaded, are always loaded to browser memory, independent of it's size and whether anotatrix is running on server. As far as I remember, this wasn't the case at the end of 2017, which means it's regression.

If this is true, it means that no big corpus can be loaded to annotatrix and I think that fixing it has more priority than any other task we discussed today. Do you agree with this, @jonorthwash, @ftyers?

In any way, @yaskevich, please create an issue for this when you can.

jonorthwash commented 5 years ago

I discovered (@yaskevich told me), that corpora, when uploaded, are always loaded to browser memory, independent of it's size and whether anotatrix is running on server.

I don't totally understand the issue. Could you clarify what the expected behaviour is in relation to what it does now?

maryszmary commented 5 years ago

The expected behavior would be to upload the corpus directly to the database without storing it in localstorage or, god forbid, global var, when the server is running. If I got it right, now it writes any corpus uploaded to annotatrix in every version to browser's memory.

jonorthwash commented 5 years ago

Ah, yes, good. There was some question last summer about what "upload" means. For entirely client-side operations, it means "load to browser memory". And for server-side operations, I suppose it should mean "upload to server". I wonder if we should allow users to do one or the other regardless, though: "import file and play with it in client-side memory". Perhaps we could distinguish between "upload" (send to server) and "import/load" (load to browser memory)? @ftyers, thoughts?

jonorthwash commented 5 years ago

I haven't seen any updates here in the last week. @yaskevich?

ftyers commented 5 years ago

I don't like the idea of distinguishing, I think if it's client side it should say "load" and if it's server side it should say "upload". "upload" for me has he semantics of sending it to another server over the network.

jonorthwash commented 5 years ago

"upload" for me has he semantics of sending it to another server over the network.

Agreed.

I think if it's client side it should say "load" and if it's server side it should say "upload".

This is a distinction. I agree with this one.

ftyers commented 5 years ago

@jonorthwash I meant more like "there should be only one button" and the semantics should change if it's standalone or connected to a server. But either way, we agree.

jonorthwash commented 5 years ago

@ftyers, ah, got it. That means that if run in server mode, there's no browser-only functionality. I can live with that, but I could see it making sense either way. I guess it's more confusing if it allows both, though...

jonorthwash commented 5 years ago

@yaskevich, any updates? What progress have you made over the last several weeks?

yaskevich commented 5 years ago

@jonorthwash In the beggining I investigated issue with big files and analyzed the code of keyboard shortcuts, but most of the work was devoted to priority issue – Github integration. All the progress is thoroughly noted in corresponding issue comment, and now it is accessible via comment history.

I thought you as a repo owner receive comments to issues of this repo.

Since you've just merged #356, I am able to push all the changes in my repo for preparing new PR.

jonorthwash commented 5 years ago

Since you've just merged #356, I am able to push all the changes in my repo for preparing new PR.

Couldn't you have done this before via a branch?

jonorthwash commented 5 years ago

What's the current status? How has progress been going?

yaskevich commented 5 years ago

Github integration progress is logged there: https://github.com/jonorthwash/ud-annotatrix/issues/376#issuecomment-509345460 Also some smaller task (server-side logging): https://github.com/jonorthwash/ud-annotatrix/issues/378 I commit to the branch dev19: https://github.com/yaskevich/ud-annotatrix/commits/dev19

jonorthwash commented 5 years ago

Why are you working in a separate branch?

yaskevich commented 5 years ago

You mentioned couple of times recently (in this conversation as well) that branch strategy would be good for previous work, so I tried to use it, just in case.

If it is unhandy, it is possible to merge current dev19 to master, AFAIK it must have been quite easy in git.