These are chat archives for biojs/biojs

15th
Feb 2018
rowlandm
@rowlandm
Feb 15 2018 10:04
Hi @/all - BioJS community meeting on next Thursday at 10.30am GMT
biojs/organisation#25
rowlandm
@rowlandm
Feb 15 2018 21:16
@alkesh47 - I saw you contributed to AstroPy - very cool!
Alkesh Srivastava
@alkesh47
Feb 15 2018 21:21
Yea, just testing the waters of open source contribution. I hope that's okay @rowlandm
😅
rowlandm
@rowlandm
Feb 15 2018 21:54
Of course it is!
I know some people in Astronomy
I'm trying to understand the data more
Alkesh Srivastava
@alkesh47
Feb 15 2018 21:55
I was just getting familiar with the code base, and I agree the data can be a bit overwhelming at times
rowlandm
@rowlandm
Feb 15 2018 21:56
I'm at the level where I want to know the types of data
let alone look at the data!
Alkesh Srivastava
@alkesh47
Feb 15 2018 21:59
On similar note, the data that we are fetching from the workman, is it streamed response, or a single big JSON object?
I was trying to work a solution for displaying the data using VueJS and streamed respsonses could really speed things up
rowlandm
@rowlandm
Feb 15 2018 22:03
ideally we would pick the faster one
but I also think we could do some preprocessing on the workman
and provide an ability to categorise or force users to do a search
no user is just going to want to browse.... I don't think
@/all - of course there could be other schools of thought
at Stemformatics, we never do a full browse
on 300+ datasets
Alkesh Srivastava
@alkesh47
Feb 15 2018 22:06
segregating the data into categories would surely improve the search time mainfold
rowlandm
@rowlandm
Feb 15 2018 22:07
exactly
Alkesh Srivastava
@alkesh47
Feb 15 2018 22:07
*manifold
rowlandm
@rowlandm
Feb 15 2018 22:08
did you see this?
Its to make the front page load time fast by providing only one or two json calls
we could make it faster still by running Http2
Alkesh Srivastava
@alkesh47
Feb 15 2018 22:11
Http2 would reduce the latency issues that would prove beneficial to the end user
even when datasets become large
rowlandm
@rowlandm
Feb 15 2018 22:16
yup
And also not providing all the data for the categories will make a difference too
I'm assuming everything is being sent
which is why it's so slow
Alkesh Srivastava
@alkesh47
Feb 15 2018 22:19
Separating the data upfront into categories would prove beneficial for the speed as well as provide a more simpler uncluttered UI with more relevant items
rowlandm
@rowlandm
Feb 15 2018 22:20
yes
Alkesh Srivastava
@alkesh47
Feb 15 2018 22:20
Quick question, what does the left hand section on the landing page https://app.moqups.com/rowlandm/qTdAndgOF4/view/page/aa9df7b72 does? What is the purpose of all those dropdowns and buttons?
rowlandm
@rowlandm
Feb 15 2018 22:20
does it just read in the whole thing at the moment?
Oh that's just an example of a BioJS component
So I wanted to show something straight away that would demonstrate BioJS
Alkesh Srivastava
@alkesh47
Feb 15 2018 22:22
Currently it does return a very large JSON object and it performs a search everytime something new is typed
But Dennis said a few days earlier that he was working on a more improved backend that returns responses from hapi.js and uses Mongo for database.
rowlandm
@rowlandm
Feb 15 2018 22:24
Right
I'd think we could work something pretty simple in the long run
Alkesh Srivastava
@alkesh47
Feb 15 2018 22:26
We can implement a browser caching mechanism to further hasten the process
rowlandm
@rowlandm
Feb 15 2018 22:30
I'm pretty confident that with some simple engineering on the backend for data transfer, this will be more than enough
I think the key here is readability and simplicity of code
You can see we don't have many volunteers so it's improtant that we can easily fix problems as they comeup
Make it too complex and only the person who did it can fix it - that's a bad idea for BioJS I think
Alkesh Srivastava
@alkesh47
Feb 15 2018 22:33
I agree, code maintainability should be of prime importance.