These are chat archives for biojs/biojs

21st
Jun 2018
Yo Yehudi
@yochannah
Jun 21 2018 08:42
I think there might be a slight bug with the star icons :laughing:
image.png
Sarthak Sehgal
@sarthak-sehgal
Jun 21 2018 08:44
Yes, @yochannah. I am currently using some icons library which messed up while compiling. I'll find a better way out and correct it. Opening an issue on the frontend repository so that I don't forget this. Thanks for reminding! :smiley:
Yo Yehudi
@yochannah
Jun 21 2018 08:44
good idea
I'm fond of inline svg because they are embedded in the doc and can't end up being the wrong character or displaying strangely if there's a CORS loading error. icomoon has good demos of how to use them: https://icomoon.io/app/#/select
Dennis Schwartz
@DennisSchwartz
Jun 21 2018 09:47
@Megh-Thakkar I’ll check. I think I just added it to the package.json but actually didn’t publish a new version :see_no_evil:
I published a new version now. Give it a few minutes and try again :)
Yo Yehudi
@yochannah
Jun 21 2018 11:21
@Megh-Thakkar @sarthak-sehgal do either of you use Twitter? Let me know your handles if you do and I'll mention you in future tweets https://twitter.com/BiojsLibrary/status/1009717814314389506
Megh Thakkar
@Megh-Thakkar
Jun 21 2018 13:20
Sorry @yochannah made an account years back, haven't used it :laughing:
Thanks @DennisSchwartz !
Yo Yehudi
@yochannah
Jun 21 2018 13:53
haha ok
Sarthak Sehgal
@sarthak-sehgal
Jun 21 2018 14:02
@all, I was implementing the visualizations earlier today. I could get most of the things to work right except a few so I had these doubts:
  1. Every component listed in BioJS which has a visualization, had a "sniper" object in its package.json file. The object lists dependencies like the scripts and stylesheets. Wzrd.in is used to get the bundled build version of the script. Now, there are some packages which do not use Browserify to bundle their code. The "sniper" object then has a property "noBrowserify" which is set to be true. I have taken these into account but then I suspect that there might be other exceptions too. So, is there any documentation or a schema kind of a thing which the components having visualizations followed to integrate with sniper?
  2. In Cytoscape (might be the case with a few more components), the performance-tuning.js file for the performance tuning visualization fetches a file from another relative path to its github repository. Do I scrape through all the JS files to check if they require any other file?
  3. We are currently using rawgit.com to get all the scripts from the GitHub repository as raw code but I don't think we can use it for the production version. Any alternatives? @Megh-Thakkar?
The core problem is that Sniper was integrated during the inception of BioJS or maybe when there were less components. All the components actually follow a set of rules to integrate with sniper (as the "sniper" object exists in the package.json). Can we get those rules somehow? It would make things a lot clear as I won't come across any "exception" then and the code will probably be much structured.
Yo Yehudi
@yochannah
Jun 21 2018 14:12
I'm hunting through old docs now @sarthak-sehgal
BUT looking at notes from early 2016:
"Make it easy to deploy the registry locally (“dockerise” the BioJS registry?) (easier sniper?) Revise Documentation? Main Devs have moved on."
and i've another doc saying "we need to document the sniper, can someone who knows something about it please do so?" 😆
Yo Yehudi
@yochannah
Jun 21 2018 14:22
I assume you've looked through https://github.com/biojs/sniper
apart from that I don't know of anything else - if those aren't any help we could try to contact sebastian and see what he remembers
let me know :)
Sarthak Sehgal
@sarthak-sehgal
Jun 21 2018 15:10
@yochannah, I did go through the README of the sniper repository. I skimmed through the BioJS Edu docs and they contain more or less the same thing.
The current workman, as per my understanding, clones the whole repository (or at least the "build" folder) and then creates the static page having the paths that of the server itself.
I think so because:
Have a look at this file: view-source:http://workmen.biojs.net/demo/cytoscape/performance-tuning
It is the source of the "Performance Tuning" example of Cytoscape
Search for "fetchJSON("
The path mentioned there is an absolute path to a JSON file being served on the server.
The file is there in the /build/data folder of the cytoscape github repo btw.
The visualizations might be a bit slow for this very reason!
Megh Thakkar
@Megh-Thakkar
Jun 21 2018 15:29
@sarthak-sehgal it's in the snippets/data folder, the snippets folder where js files are
Also, there are certain discrepancies which are confusing.
For example, in a Component msa the package.json states the path differently for one of the dependencies.
Yo Yehudi
@yochannah
Jun 21 2018 15:32
it might be worth making a ticket on the sniper repo and mentioning @wilzbach if you are up to that?
unless possibly that mention summoned him :D
Megh Thakkar
@Megh-Thakkar
Jun 21 2018 15:32
Shall I open an issue in the msa repo?
Yo Yehudi
@yochannah
Jun 21 2018 15:33
whichever you think is appropriate!
Megh Thakkar
@Megh-Thakkar
Jun 21 2018 15:37
Opened a new wilzbach/msa#239. Will update as soon as he replies :smile:
@sarthak-sehgal and I are able to fetch the dependencies mentioned in the package.json and I have also written the code to update the database accordingly. But these confusions are setting us back so we thought it would be best to discuss it with everyone!
rowlandm
@rowlandm
Jun 21 2018 22:15
Remember @sarthak-sehgal and @Megh-Thakkar that the only up to date documentation is the code itself
It might be worthwhile running the workman to see it in action and to debug how it works - but I'm unsure how much detail you are missing.
When in doubt, you will have to re-engineer it from first principles via the code...
Megh Thakkar
@Megh-Thakkar
Jun 21 2018 22:55
@sarthak-sehgal and I will start working around that!