Flora/documentation/overview network

partly addressing documentation sprint issue #83 regarding overview

@j.guez you can check the section on prediction inside the overview, it might be too long here and details could go to prediction.rst

@embray I'm having a some trouble to decide what should go to overview, is it just a rough summary of the dedicated page. Not sure whether I was too long/too repetitive (compared to the dedicated page)

@fjay I forgot to mention --output OUTPUT in the prediction part

I think this is fine in terms of level of details. My intent with this page was that it serve as a quick reference.

In the case of prediction it might be that it's detailed enough that we can do away entirely with having a separate page on the topic, unless @j.guez is already far along on that...

added 1 commit

• 86ed257b - [documentation] minor updates, mostly for spelling/formatting

Compare with previous version

Indeed, it seemed pretty detailed to me, I don't see what I could add. It might be a good idea to use this prediction text as the main prediction.rst page and summarize it for the quickstart tutorial ? @fjay @embray I can do this eventually.

ok, yes you can do this, add whatever you think is missing and summarize to the most important bits in the overview section (you can do all this within this branch), thks

added 1 commit

• 963cfe85 - Update prediction.rst

Compare with previous version

added 1 commit

• 3b468554 - Update prediction.rst : minor changes

Compare with previous version

added 1 commit

• 732e72d2 - Update overview.rst - summarize the prediction part

Compare with previous version

@fjay I did the summary and added the full text to prediction.rst in this branch. I also changed small things in the full text.

I'm a little confused by the intended meaning here. Currently (although we could and should) we do not define a format for storing an entire dataset in a single file. A single file can contain one SNP matrix + position array, so isn't it misleading to call this a "dataset"?

yes you're right, for now the user has to pre-cut the genome into multiple pieces (corresponding to replicates), each saved into an .npz file I guess the definition of real dataset is changing depending on the context (for selection it could be indeed just a piece of a genome). Or people might in the fure built a network that scan the whole data store into one file. (in particular for species with tiny genomes). However I see that it can confuse people. Let's change this.

@j.guez are you working more on it?

I agree, everything you write makes sense and I believe should be an option. It may also be useful for improving I/O performance on some distributed filesystems. But for now each file contains a single "datum" as it were.

@fjay No, I don't have anything to add, at least for the moment.

changed this line in version 8 of the diff

I changed this a bit in b5bd9735

OK thanks ! So I think this can be merged

added 1 commit

• 0f721dd9 - [documentation] misc minor nitpicks

Compare with previous version

added 70 commits

• 0f721dd9...c0a45224 - 62 commits from branch master
• 14774d9d - start working on overview
• c9c5dfbb - mention of filenam_format propertie
• 502a4c39 - first version of completed overview
• 15a59790 - [documentation] minor updates, mostly for spelling/formatting
• ba8ac4a2 - Update prediction.rst
• f953e9b9 - Update prediction.rst : minor changes
• 7466e46e - Update overview.rst - summarize the prediction part
• 3d95d116 - [documentation] misc minor nitpicks

Compare with previous version

added 1 commit

• b5bd9735 - [documentation] change potentially confusing verbiage about "datasets"

Compare with previous version

resolved all threads

approved this merge request

marked this merge request as ready

enabled an automatic merge when the pipeline for b5bd9735 succeeds

merged

mentioned in commit 0475ac08