GSoC 2021 Devansh Shukla
LaTex/submitted version of the proposal here.
Name: Devansh Shukla
Academic Program: Integrated Master of Science in Physics
Institution: Sardar Vallabhbhai National Institute of Technology, Surat, India (svnit.ac.in)
GitHub: @devanshshukla99
Matrix: @devanshshukla99:matrix.org
E-Mail: devanshshukla99@gmail.com
Contact No.: +91 9826887954
Location: Sagar, MP, India 470002
TimeZone: IST | UTC +5:30
Current Affiliation: Sardar Vallabhbhai National Institude of Technology, Surat, India
Academic Program: Integrated Master of Science in Physics.
CGPA: 9.71/10
Expected Year of Graduation: 2023
My Resume is available at Curriculum vitae
Languages: Python, C
Platforms: Linux, Windows
Source Control: git
Software & Tools: LaTeX, conda, WxMaxima, Mathematica, +.
- Developed a Python program for Radio Frequency Interference Scan at Sardar Vallabhbhai National Institute of Technology, Surat, India.
- github.com/devanshshukla99/SAS
- Developed a python program for straight-forward terminal-based socket communication with 128-bit AES encryption.
- github.com/devanshshukla99/Juno-ReDesign-Sockets-Communication/
- Package for getting colored text output in Python.
- github.com/devanshshukla99/Chromos
I have been very interested in the Astrophysics and Cosmology domain for a long time. SunPy specifically caught my eye due to my course, Intro. to Space Physics, which included a seminar on Sun-spots in which I utilized SunPy to generate AIA and HMI Image of the Sun, thereby showing the correlation of sunspots and magnetic field; I have been reading about SunPy since then, slowly learning about the incredible community. I chose this project since it brings out the best of both worlds, some Solar Physics and some Programming; for me, it looked like a natural choice in my career to learn and hopefully contribute positively to the community; more specifically, next subsection points to some reasons and benefits of such a project; of course, I'm open to finding more in the future.
- Super helpful when combining visualizations of multiple objects, for example combining HMI with AIA Image or adding a LASCO Map.
- Moreover, PFSSPy, a dependent of SunPy, which calculates the Potential-Field Source-surface model for computing the coronal magnetic field features, has an exciting attribute, which can overlap field over the image. One of the best use-cases of 3D plotting in SunPy.
- Suitable for intuitive understanding and visualization.
- Looks super cool.
Presently, PR #4591 by Dr. David Stansby(@dstansby) provides a foundation for the project. It's basically a proof-of-concept; it takes a sunpy.map.Map object, converts it to Heliocentric-Inertial Coordinate system, creates a mesh and finally plots in PyVista.
PR #5193: Under Review
- Adds support for file-handlers and file-type in Map and TimeSeries.
- https://github.com/sunpy/sunpy/pull/5193
Issue #3303: Currently Working
- Create a Test or a function to report status on all remote servers SunPy pings during a test run.
- Thinking about using a pytest plugin, pytest-vcr, to intercept the requests with a flag, but its likely to change.
- https://github.com/sunpy/sunpy/issues/3303
SunPy currently uses Matplotlib for plots and has limited support for 3D plots owing to PR #4591 . The project's goal is to pack PR #4591 into a new package, independent of SunPy, and develop a complete API initially interfacing with PyVista but general enough for other backends like plotly too. SunPy has excellent support for 2D plots with Matplotlib, but 3D support is missing; through 3D backing, one can easily combine visualizations of multiple objects such as overlapping HMI on a AIA image in 3D; it'll be a gem for intuitiveness. Moreover, PFSSPy, a dependent package of SunPy, which calculates the Potential-Field Source-surface model for computing the coronal magnetic field features, has an exciting attribute, which can overlap field over the image; this'll precisely be the best use case for 3D support in SunPy. I believe the project is evenly balanced between writing, testing and documenting code
- Convert code from PR #4591 into a new package under SunPy org, and understand it down it a fiber.
- Add infrastructure to plot astropy coordinate object.
- Add infrastructure to support other data types, too, like for SOHO, RHESSI, etc.
- Hoping for 90%, targeting 100% code-coverage. Benchmarks and extra optimizations, if possible.
- Documentation for the whole project.
I plan to fully commit and immerse into the project, with a minimum investment of 20hr/week and with an average of 25hr/week. Honestly, I am not a big fan of assessing based on time investment; my priority would be to make and follow weekly task schedules, with some buffer time for unforeseeable circumstances and additional tests.
I have attached a thorough timeline of the proposed project. Regarding my commitments, I'll have my course classes from August 10, so I've shifted some of the workload to July. I plan to to publish a weekly blog for the project with all relevant details to help future developers and users; I believe this would be critical in the long run.
-
May 17, 2021 - May 24, 2021
- Familiarize with the code and the community.
- Since I am already familiar with the dev env and test systems used in SunPy so I'll quickly go through it again just in-case, and will focus more on the code and how it all is integrated together.
- Create a separate package with PR #4591 as foundation and set-up all necessary workflows.
-
May 24, 2021 - June 1, 2021
- During this time, I'll discuss with my mentors and come with more specific details about the project, like how the public face of the API should behave, etc.
- I'll learn and try to solve the challenges I'm already aware-of in the project, see subsection V.IV
- Learn more about dependencies like AstroPy, plotly, MPL, PyVista and see how their API is designed.
-
May 24, 2021 - June 1, 2021
- Plan for the API.
- Thoroughly understand and test the existing code.
- Continue learning about the dependencies.
-
June 7, 2021 - June 14, 2021
- Define tests and documentation for the existing code.
- Discuss & Work on the API.
-
June 14, 2021 - June 21, 2021
- Add code for plotting sunpy.map.Map
- Work on the API.
- Crush the bugs along the way.
-
June 21, 2021 - June 28, 2021
- Add tests and documentation, hoping to achieve 75%+ code coverage.
- Check the Integration with SunPy. – Waypoint - 1 Check
-
June 21, 2021 - June 28, 2021
- Crush the bugs.
- Add tests and documentation, hoping to achieve 90%+ targeting 100% code coverage.
- Finalizing the API.
-
June 28, 2021 - July 5, 2021
- Gather Feedback.
- Finalizing the API for the evaluation.
- Crush the Bugs.
-
July 5, 2021 - July 12, 2021
- Finalizing the API for the evaluation.
- Plan for plotting astropy coordinates in 3D infrastructure.
- Buffer time.
-
July 12, 2021 - July 16, 2021
- Buffer time, for incorporating required changes as per feedback.
- More debugging.
-
June 16, 2021 - July 23, 2021
- Work on infrastructure for plotting astropy coordinates in 3D.
- Discuss on adding infrastructure for other data type.
-
July 23, 2021 - July 30, 2021
- Test infrastructure for plotting astropy coordinates in 3D.
- Plan infrastructure for other data type, SOHO, EUV, RHEISSI, tbd.
- Buffer time to crush bugs.
-
July 30, 2021 - August 6, 2021
- Work on adding infrastructure for other data type, SOHO, EUV etc, tbd.
- Add tests and documentations, hoping to achieve 75%+.
-
August 6, 2021 - August 13, 2021
- Work on support of pyvista for jupyter-notebook.
- Crush the bugs.
- Finalize the API.
-
August 13, 2021 - August 20, 2021
- Add tests and documentations, hoping to achieve 90%+ targeting 100% code coverage.
- Finalize the API, prepare for the pre-release.
- Crush the bugs.
- Check Documentation for mistakes/typos/bugs.
-
August 13, 2021 - August 16, 2021
- Finalize the API, incorporate the feedback from additional review.
- Crush the bugs.
- Gather feedback.
-
August 16, 2021 - August 23, 2021
- Pre-Release.
- Crush the Bugs.
-
August 23, 2021 -
- Maintain support and work on improving the performance and stuff.
- Keep crushing the bugs.
- The API will initially depend upon pyvista, which works on VTK, unfortunately Read the Docs doesn't support a headless-display, but this need to be investigated after reply from @cadair suggesting a possible solution.
- As of now, whenever I run pyvista from Jupyter Notebook it bugs out crashing the notebook, I'm not sure if this is a wide-spread or localised issue
This is the first time I'm applying to a project.
This is the only proposal I'm submitting.
- Before - 20hrs/week.
- During - 25hrs/week.
- After - 15hrs/week.
Yes, I am eligible.
