search

AlexsLemonade / refinebio-examples

Example workflows for refine.bio data
https://www.refine.bio
Other
11 stars 5 forks source link

Try different strategy for ortholog file download #411

Closed cansavvy closed 3 years ago

cansavvy commented 3 years ago

Purpose

I read #299 and gathered the main problems with the ortholog file download were all the Safari caveats:

To summarize, the general aspects to think about here are how do we handle getting the data while considering: Safari users who cannot access the ftp link Safari users whose downloaded file will be decompressed instead of compressed (.gz) Other users for machines that we have not yet tested (like Windows)

Issue addressed

299

Strategy

1) I made this download by download.file() which should work no matter the browser someone may be using. 2) I found a thing on Apple that says that if you want to view ftp using Safari, you can change it to an http. So for looking at the files available, I steer the user to that file page but a different link (with http) if they are using Safari. 3) I trimmed out the GUI instructions since I think this should work no matter what. (I wasn't able to get the GUI to work for me for some reason anyway -- the files won't download).

I tried think link on Safari and it works.

Additionally the data file got changed to the wrong dataset that was mouse? This got messed up when I did #298 apparently. It still was running like that but not really, everything was returning 0 rows and no errors. I changed it back to the correct zebrafish so it was actually working. (This is a tad scary and when we are taking a look at polishing the other dataset examples, we should make sure nothing else like this slipped through)

Remaining concerns and questions

Do we feel like this is less confusing to follow and more flexible for people who might be using different web browsers?

cansavvy commented 3 years ago

LGTM! Wondering if we should apply the same method to the ORA file download? That might be less confusing than replacing a URL with a file name here:

Well, for sure I'm going to file a follow up PR for the RNA-seq ortholog example.

I can see your point about the downloading the file being less confusing there too though. I'll file a separate issue about that. Seems reasonable. (GSEA examples have the same kind of URL read in step so this would probably apply to them too).