Orange Rocks! Tips on building a GUI

[snewhouse]

Hi All
Orange is a great package,the GUI is great, the functionality is great and has been a fantastic resource to help some of our students overcome their fear to Machine Learning and computer/data science!

I was wondering if you could point me to some "How To's" or tutorials/resources for building a similar kind of GUI.

We are working with Docker and looking to build some easy pipelines for common bioinformatic workflows - focusing on NGS analysis first, and I think the structure off the Ornage GUI would be a great backbone to build upon.

I'm new the GUI building world, and only just dipping my toe in the Python world (dont hate me)...

Cheers

Steve & Team

[janezd]

Would you like to use the Orange's code for the GUI or would you like to implement your own environment from scratch, but you'd like it to be similar to Orange?

[snewhouse]

Hi

Thanks for the quick reply!

To be honest, a bit of both - we would need to test both options to see what would work best or could be adapted easily.

I really like the layout and pipelining/workflow widgets of orange, and think it will be great for bioinformatic pipelines in general.

To start off with, I'm wondering if using your Code for the GUI might be the best starting point (If you are all good to share that is?), but happy for any suggestions/tips you may have.

If adapting the code isnt an option or is too much of a headache to port over for standard bioinformatic pipelines, then we will build from scratch - again something similar to Orange, as we have had computer shy folks, find it fairly intuitive and quick to learn...

Cheers

Steve

[janezd]

Hi,

I think you should try writing a few widgets on your own. That is, you
reuse the GUI, but add your custom elements for your pipelines. (What do
you mean by "bioinformatics pipelines"? There already exists a strong
bioinformatics add-on for Orange. Have you checked it?)

We used to have a tutorial for writing widgets, but it was for an older
version of Orange and we haven't ported it to the new one yet. But you can
just take a few existing widgets as a template. And, of course, ask if you
have any questions. ;)

Best,
Janez

On Thu, May 14, 2015 at 11:29 AM, Stephen J Newhouse <
notifications@github.com> wrote:

Hi

Thanks for the quick reply!

To be honest, a bit of both - we would need to test both options to see
what would work best or could be adapted easily.

I really like the layout and pipelining/workflow widgets of orange, and
think it will be great for bioinformatic pipelines in general.

To start off with, I'm wondering if using your Code for the GUI might be
the best starting point (If you are all good to share that is?), but happy
for any suggestions/tips you may have.

If adapting the code isnt an option or is too much of a headache to port
over for standard bioinformatic pipelines, then we will build from scratch

• again something similar to Orange, as we have had computer shy folks,
  find it fairly intuitive and quick to learn...

Cheers

Steve

—
Reply to this email directly or view it on GitHub
#23 (comment).

[snewhouse]

Hi
Would love to try writing a few widgets - please point me to a basic tutorial or some of the exact code and I'll give it a whirl.

I have just added to bioinformatics add-on for Orange and will have a play..

By, "pipelines" (sorry I was just talking to some medic-PI's who glaze over when I mention the details and I was stuck in "vague" mode) I meant NGS analysis pipelines going from raw fastq to BAM to VCF to annotations using tools like our NGSeasy or Brad Chapman's bcbio-nextgen (https://github.com/chapmanb/bcbio-nextgen). If this works, could be a nice addition to Orange also - a NGS add-on?

I'd really like to see if I could (with some help) use something like Orange to wrap up these pipelines and provide some easy GUI tool to manage and rum Dockerised applications. When I am bored and have more time, I may just look into Dockerising Orange Installation etc

Also (I haven't looked yet) but is there a way to add custom or additional databases to the bioinfomatics add-on ie i'd like to update with all of the Illumina array chip Pobe id mappings.

On another note - do you or the group run any tutorials/training?. If we can arrange something, for later this year or early next year, it would be great to have you over to give a talk and present Orange and its capabilities to more of the staff and students at KCL. I will bring this up with my boss next week also. Let me know if this is something you & Orange would be interested in

Cheers

Steve

[snewhouse]

am starting from here http://docs.orange.biolab.si/extend-widgets/rst/basics.html...lets see how I get on