GSOC'20 project OpenDevoCell Integration'Final submission for a OpenWorm
Mentors - Dr. Bradley Alicea, Vinay Varma
Working Link for Project parts deployed Online
- DevoWormAI center - https://devoworm.github.io/DevoWormAi/index.html
- DevoZoo - https://devoworm.github.io/devozoo/index.html
- C.elegans Segmentation - https://celegans-devolearn.herokuapp.com/
Repository which is for sole purpose of my project/s.
- Devolearn - https://github.com/DevoLearn/C_elegans_DevoLearn
- Devoworm - https://github.com/devoworm/devoworm.github.io/tree/master/DevoWormAi
- DevoZoo - https://github.com/devoworm/devoworm.github.io/tree/master/devozoo
Repository in which I contributed.
- Devolearn - https://github.com/DevoLearn/devolearn
- Devolearn - https://github.com/DevoLearn/Education
Weekly Meetings and Updates.
Our group use to have a weekly update formal meeting on every Monday which is uploaded to the Youtube under Devoworm channel. It not only consists of our update, also the paper review, projects that we need to do in future etc.
One can find final presentation slide on this link - https://drive.google.com/file/d/1KZiUsqpJV0Sm-wWLGgYBUnPvxLnWv5xk/view?usp=sharing
This project is about integrating the resources available at the OpenWorm and DevoWorm group. This project will focus on improving the data science and machine learning infrastructure of the DevoWorm group. In this project, we need to integrate the different previous year GSoC projects under one roof. The priority this year is to improve the web interface to make it more and more user-friendly. We want to provide a general analysis and statistical tools so that researchers around the world can use these tools for their analysis. This includes the ability to incorporate new forms of analysis as well as algorithms for new types of data. The ability to extract quantitative data from the movie images is key to conducting comparative and time-series analysis
Work that has been done so far.
- During the GSoC period, I have succesfully created the pipeline to host all the current existing finished project on web online. I have given links to all currently live project above. A brief detail about them are given below.
- It is a front most page (face) of DevoWorm Group. There we have listed our all the projects that we are doing right now, and we aim to make them available as soon as we can. It is kind of a splash page, which contains links to web app, colab notebooks etc.
- The main task this year is to improve DevoZoo web portal, to more it more attractive and easily to access. I have improved the core of DevoZoo this year, give it a complete fresh look. Now, user can have a idea from the very front page itself like what kind of data is available in which format.
- This is the improvisation of GSoC 2019 project OpenDevoCell. It has now a very user friendly and good looking front end, and has the description of our newly formed package DevoLearn. How to use and install it.
- It is a package that is being developed by Mayukh Deb and Ujjwal Singh under our mentors supervision. I have helped Mayukh in improving the algorithms involved and setting up PyPI package. Mayukh Deb is major contributor in this project as it is very deeply linked with his GSoC project.
Things that are developed but not hosted yet.
Job Scheduler: We hae assumed a scenerio where people are extensively using our portals for there research work and they might not get there results instantaneously. For this, we have created a Job Scheduler script which is there in GSoC repo, but it is not integrated with the platform as of now, as there is not much need of it as of now.
Devoditor: This is a iamge labelling online tool, which is 80 percent completed and might be hosted before GSoC or in first week of September. This tool will have many functions like crop, annotate, rotate, adding text, drawing manually etc.
Things to do:
Host bacillaria model: We need to host bacillaria model also on our portal, but due to some memory limitations and other factors, we are not able to upload it as of now.
Creating educational resources for researchers under DevoWorm Academy.
What I learn from GSoC
It is a very great experience for me, I learn how to work with the organisation, how to write production ready code. I have mentors and most the members, who lives in different continents, still we manage to match our timing and get things done. My mentors have taught me what are best practices when you are doing open source development. I have learnt a lot from them, as biology is never my jam, I only studied bio till high school, After that I am all computer science guy. They(mentors and other members) make things so interesting, I am totally unaware that how Computer Vision and other ML tehniques can make impact in traditional scinece fields like Biology, Zoology.
Overall it is a very enriching experience and I would like to thank my mentors and all the community to make this GSoC one of the best experiences.
Also, a big thanks to INCF, OpenWorm, because their effort, the whole process become very easy and fluid.