-
-
Notifications
You must be signed in to change notification settings - Fork 577
Home: Weekly Community Meetings Archive
This is a publicly and collaboratively editable document for the minutes and agenda items to be covered at the upcoming SunPy Community Meeting for this week. Community members should feel free to add agenda items for the upcoming meeting as they see fit. Following the meeting, the agend and minutes should be copied over to the wiki. The purpose of this document is only to provide a place to collaboratively and quickly record the minutes of the community meeting.
This document may change dramatically or be purged periodically.
- Welcome to Newcomers
- Governance
-
sunkit-image
follow-up - Coordination meeting planning
- Project board: https://github.com/orgs/sunpy/projects/3
- Meeting planning notes: https://demo.hedgedoc.org/BOOMHZEpQ4esaQrh_iwmCA#
- Jack has requested a tag-up with the VSO folks one afternoon since they are meeting the same time we are
- What would be most useful to discuss? Collect list of outstanding VSO/data provider issues?
- PR Review
- Project-wide gallery discussion
- spurred by eclipse example--where does it go, need a better place than the blog
- back to this idea of "learn sunpy"
- Executable book? Example gallery?
- We've talked about this for a long time--can we just do it?
-
sunkit-image
WOW- Laura to follow-up Monday, if someone else wants to do it in the meantime, feel free
- Governance
- Look over the SEP
- Announce in Solar News (Will) and UKSP (Stuart)
- Regular project-wide updates
- Make community more aware of what is happening on the project
- some don't see parts of the ecosystem very often
- discuss at coordination meeting
- Welcome to Newcomers
- Upcoming events
- Will giving a talk at DoE Python Exchange event https://meetup.doepy.org/
- Albert will be the lone SunPy representative at TESS (week of 8 April)
- Nabil and Will will be at the PyHC Summer School (week of 20 May)
-
sunkit-image
follow-up - PR Review
- Optional dependency discussion
- What should be included in
pip install sunpy[all]
- Dask and asdf
- special cases because you have to import them in your code in order for
sunpy
to use them
- special cases because you have to import them in your code in order for
- glymur (lxml), scikit-image, opencv
- add extras for each of these
- then decide whether these should be included in
all
- decisions
- Remove dask extra
- Keep asdf extra
- image gets turned to scikit-image
- opencv stays
- spice
- what goes in
all
?- Stuart and Will inclined to make
all
actually include - Making
all
the default
- Stuart and Will inclined to make
- See this PR for all final decision making: https://github.com/sunpy/sunpy/pull/7536
- Welcome to Newcomers
- PyHC Spring Meeting is next week
- sunpy talk is the first day
- PyHC Summer School--is anyone else going?
- Will would like help if possible
- Boulder,
-
sunkit-image
plans? - PR Review
-
sunkit-image
- Discussion of https://github.com/sunpy/sunkit-image/pull/188
- Consensus is that we want this and are fine with it just being a wrapper function
- We all seem to agree that there is value in wrapper functions--collect all functionality in one place, enforce something like a consistent interface
- Needs
- tests
- some changes to the function signature
- gallery example entry
- Needs the map/data array decorator
- Laura to follow up with Fred when it comes to next steps
- Scope out what
sunkit-image
is--we need to codify it and put it in the docs- We could devote a small amount of time at the coordination meeting to doing this
- This has the potential to be one of our most used affiliated packages
- This will help us vet future contributions
- Discussion of https://github.com/sunpy/sunkit-image/pull/188
- PyHC Summer School
- https://github.com/SolarOrbiterWorkshop/solo_workshop_gsfc/blob/main/solo_workshop_gsfc_tutorial_oct2023.ipynb
- Good inspiration for PYSPEDAS work?
- OSTFL Funding Call
- Will unsure about putting something in, not sure what it would be
- Stuart more convinced we should, it would be a shame to let it go by and not propose
- Astropy have gone the route of parceling out money to software people all over the world to do things
- Could we go this route? This would seem more beneficial as we just dont have, at any US institution, a large enough group of people that can be paid, have time to do the work, and have the expertise to do the work
- Spreading the money around would partially solve this problem, though it would mean this couldn't be 'science' money
- To do this we need two things:
- A list of things we want to do--this can be project wide; a good place to start would be identifying the gaps in the ecosystem currently
- A governance structure/system for evaluating proposals of work and handing out money; our current proposed restructuring of the governance has elements of this, but would probably need more
- We are static/shrinking in terms of active core members, but we are increasing in the number of people we are supporting in the community
- Welcome to Newcomers
- pfsspy successor discussion
- Coordination meeting planning
- PR Review
- Decision made: package called
sunkit-magex
🎉- Sam Bennet to run package template on
pfsspy
- Stuart to perform git foo to bring over git history to new repository
- Want to make sure that we also bring over issue history
- Looks like we need to do this manually
- Can filter by date and label
- Not many issues so this shouldn't be too onerous
- Sam Bennet to run package template on
- Coordination meeting discussion
- Advice: stay in Leiden, bus to ESTEC
- Planning session ahead of time (maybe middle of March?)
- Devote one afternoon to overview talk, stream publicly
- ESA People that we should be coordinating with?
- Welcome to Newcomers
- pfsspy successor discussion
- PR Review
- pfsspy discussion
- expected time commitment: 1-2 hours a week at the most
- roadmap for pfsspy development:
- decide on name
- review documentation
- review streamtracer documentation
- rust versus fortran
- per Sam, rust version is single-threaded and so slower (by a factor of 3 or 4)
- Fortran version has been completely removed as of v2
- seems to overall be an appetite for NOT maintaining the Fortran code
- Rust is easier, though not immediately clear how to do the parallelism
- licensing
- both pfsspy and streamtracer are GPL because they are based on code that was under a GPL license
- per Stuart, should not be an issue since these are separate packages
- reach out to David and ask for streamtracer to be relicensed
- Name?
- In general, we want to use the format "sunkit-{noun}"
- Some possibilities: sunkit-extrapolation, sunkit-extrapolate, sunkit-extrapolator, sunkit-mag, sunkit-fields
- The plan is to gather some input from others and then vote on a name at the community meeting next week
- TODOs
- Ask David about keys to streamtracer package (transferred to Stuart)
- Decide on a name
- Template repository
- Migrate existing pfsspy code to new repository
- Add Sam and Jan as maintainers
- Welcome to Newcomers
- Coordination Meeting Planning
- https://app.rallly.co/invite/nLRlV97pwYPq
- 22 April and 13 May are clear winners
- Week of 20 May is the PyHC summer school in Boulder which might complicate things...
- Presence at meetings this year (where is the link to the spreadsheet?)
- https://docs.google.com/spreadsheets/d/1muxbnL2h_pJgUVdEX_1ncrliKQeZzRiTV2BxQtJ_QtY/edit#gid=0
- TESS
- Solar Orbiter/DKIST Meeting
- SPHERE
- ESPD summer school
- ESPM (European Solar Physics Meeting) ?
- GSOC PROJECTS
- REVIEW THE DAMN PRS FOR OPENSCI OTHERWISE I WILL DELETE SUNPY
- PR Review
- Ask NF about booth at AGU
- Preparation for meetings
- Preparation for the eclipse
- Good opportunity for attracting new contributors (!)
- from outside of solar
- make use of this window in time where there is a large public focus on solar physics
- Revive the old eclipse processing repo: https://github.com/sunpy/solar-eclipse
- Update this to current sunpy
- Show example using the 2017 eclipse: https://github.com/sunpy/solar-eclipse/blob/master/eclipse/sample_photos/total_solar_eclipse2017.jpg
- 400 mm lens
- Advertise this ahead of the 2024 eclipse in Solar News
- After eclipse, follow up post with photograph from 2024
- Also, maybe other posts ahead of time, e.g. eclipse amount example for the eclipse amount
- Steve also suggested recreating the eclipse path map where a shadow is projected onto a map of the earth to show the path of totality.
- Action items:
- Blog post using the eclipse amount tool for the 2024 eclipse
- Revive old eclipse repo
- Blog post using the eclipse repo on the 2017 eclipse photo
- Advertise blog post (e.g. via NF, SolarNews)
- Blog post follow-up with a 2024 eclipse photo after the eclipse
- Good opportunity for attracting new contributors (!)
- GSoC Projects
- Eclipse processing (Steve)
- PR Review
-
https://github.com/sunpy/sunpy/pull/7271
- SOLARNET FITS validation
- Overall consensus is that this is a good idea, but need to work out details
-
https://github.com/sunpy/sunpy/pull/7271
- Welcome to Newcomers
- 5.1 has been released. Thanks David! 🎉
- SunPy Poster at AGU (to be presented week of 12/11/23)
- Will would like help assembling this poster--recommendations for where to do this?
- Abstract: The primary goal of the SunPy Project is to facilitate and promote the use and development of community-led, free and open-source solar data-analysis software based on the scientific Python environment. To accomplish this goal, the SunPy Project maintains the
sunpy
package for downloading, loading, and transforming solar data, in addition to a set of affiliated packages for instrument- or analysis-specific tasks. Together these form the SunPy Ecosystem, a set of interoperable Python packages for solar physics research. In this talk, we will discuss how the SunPy Project and Ecosystem promote and enable three of the pillars of Open Science: Open Software, Open Data and Open Results. In particular, we will discuss how open development practices have led to high-quality research software within the SunPy ecosystem and how such practices encourage community ownership of software. Additionally, we will discuss how the philosophy of Open Data in the solar physics community combined with the open development model of the SunPy Project allows users and developers to challenge the hidden assumptions of data and maximize its scientific potential. Finally, we will demonstrate how the adoption of software from the wider scientific Python community as well as the development of extensible tools by the SunPy Project has led to a more interoperable Python software ecosystem within astrophysics and heliophysics.
- Will to send out poll re: coordination meeting
- Leiden should work
- likely no cost associated with it, maybe for coffee
- This will be done before Christmas (hopefully!)
- Likely timeframe will be some time in the last few weeks of May
- Focus: NDCube/Map refactor
- PyOpenSci review is more or less done (Thanks David and Nabil!)
- All packages have seen maintenance releases since 5.1
- Except radiospectra (ongoing maintenance problem)
- radiospectra
- Radio data annoying because locations move around, formats change
- Would be great to use with astropy units, but support lacking
- Only precedent is the pint units plugin for xarray--could you model an astropy units plugin after this?
- Need a "crop by units coordinates"
- Welcome to Newcomers
- The
aiapy
repo has officially moved, all development is now on GH: https://github.com/LM-SAL/aiapy- Need to make announcements various places
- Migration still in progress
- SunPy Poster at AGU (to be presented week of 12/11/23)
- Will would like help assembling this poster
- Abstract: The primary goal of the SunPy Project is to facilitate and promote the use and development of community-led, free and open-source solar data-analysis software based on the scientific Python environment. To accomplish this goal, the SunPy Project maintains the
sunpy
package for downloading, loading, and transforming solar data, in addition to a set of affiliated packages for instrument- or analysis-specific tasks. Together these form the SunPy Ecosystem, a set of interoperable Python packages for solar physics research. In this talk, we will discuss how the SunPy Project and Ecosystem promote and enable three of the pillars of Open Science: Open Software, Open Data and Open Results. In particular, we will discuss how open development practices have led to high-quality research software within the SunPy ecosystem and how such practices encourage community ownership of software. Additionally, we will discuss how the philosophy of Open Data in the solar physics community combined with the open development model of the SunPy Project allows users and developers to challenge the hidden assumptions of data and maximize its scientific potential. Finally, we will demonstrate how the adoption of software from the wider scientific Python community as well as the development of extensible tools by the SunPy Project has led to a more interoperable Python software ecosystem within astrophysics and heliophysics.
- Coordination Meeting 2024
- Focused meeting to refactor Map on top of NDCube
- Location: ESTEC, Leiden (???)
- Early to mid May (???)
- Will needs to send out a poll to find out what dates work for everyone
- 5.1rc release is this Friday
- Review milestoned PRs: https://github.com/sunpy/sunpy/milestone/114
- Map/NDCube refactor planning
- We need to scope out what work we are going to do ahead of time
- What's the division of labor
- Discussion of this at coordination meetings past:
- submap
- still a role for this and crop
- API is tuned for 2 celestial dims
- will become an API helper/wrapper of crop
- could deprecate using pixel units in favor of slicing API
- when you slice an NDCube, you aren't guaranteed a FITS-serializable WCS back
- Current PR to ndcube could fix this: https://github.com/sunpy/ndcube/pull/649
- First cut: just recreate Map with NDCube
- Come up with a proposal for a new map api based on ndcube
- open discussion of each point of this new api
- break up different tasks of implementing this api
- Divide up effort based on implementing different parts of this API
- Different methods on the object, everyone takes a method
- e.g. some work on
.plot
, some work on.submap
- Changes needed in Map factory
- Visualization
- Map sequence
- The first thing we do is come up with a proposal for what this will look like
- Need to do this beforehand
- Write an SEP for this before the meeting
- Defines the new Map API as draft
- Start meeting with some discussion of this proposal
- Limit this to only one day at the most
- Edit SEP during this discussion and as we work throughout the week
- High-level breakdown (sorted by priority)
- Methods on map that alter the map
- I/O
- Visualization
- Factory
- Other non-mapbase things (map sequence, composite)
- Next steps after initially recreating map on top of ndcube
- Metadata (more esoteric...)
- Danny to start this SEP draft around the beginning of next year
- Welcome to Newcomers
- FF/Release discussion
- PR review
- SPICE/coordinates discussion
- https://github.com/sunpy/sunpy/pull/7237
- Debate about whether this should live in coordinates
- It could be useful to people outside of solar physics (e.g. planetary)
- Jack says there are kernels for JWST even
- Adding another affiliated package places a larger maintenance burden on us
- For now, consensus is to keep this in core
- If someone from outside solar comes along and wants it outside of sunpy, we can reassess
- What is going to make it into 5.1?
- scraper (https://github.com/sunpy/sunpy/pull/7227)
- Depends on whether Nabil has time
- Could bump it, not a huge priority
- Needs reviews!
- download sample data deprecation
- As it stands, would remove
parfive
as a core dependency for the sake of making sunpy installable with pyodide - David to pull out sample data deprecation into separate PR, address the dependency issue separately
- As it stands, would remove
- Would be great to get coordinates PRs into this release, will reassess after FF whether this is reasonable
- We could possibly bump the scraper PR
- scraper (https://github.com/sunpy/sunpy/pull/7227)
- Welcome to Newcomers
- GSoC
- PR review
- SPD
- overall, format of demo to larger audience + help session a few days later was effective
- informal feedback was overall positive
- we still did not get much feedback from the Slido that we ran during the demo...not clear why (maybe we need a longer-running survey similar to the 2020 one?)
-
sunraster
discusssion- fitting stuff could go in here
- question about where the SPICE reader could live long term
- Coordination meeting planning
- consensus between Danny, Will, Nabil was that we don't necessarily need a meeting in the style of the last coordination meeting so soon
- more focused meeting on rewrite of Map on top of NDCube
- target core group of people that are needed to do this rewrite
- this would be in place of a more "traditional" coordination meeting
- Potential timeframe: Late spring/early summer
- post elcipse and post FOXSI launch
- Will's proposal: Late April / early May
- Possible location: ESTEC
- Laura: is this possible? cost? planning?
- Goal: come away with a solid work plan for rebasing Map, a better metadata framework, bonus: plan for spectral data support
- Meeting next week: pitch this idea to wider group, see if people are onboard, get something on the calendar ASAP
- GSoC
- Reviews on the scraper PR
- Gilly's proposed ADAPT client would be a good test of this new API
- Welcome to Newcomers
- GSoC
- AGU abstract
- due today!!!!
- Will is going to submit a sunpy abstract, if you want to be on it, please add your name below by the end of this meeting, any/all feedback is welcome
- Session: https://agu.confex.com/agu/fm23/prelim.cgi/Session/191589
- Title: "Enabling Open Science with the SunPy Ecosystem"
- Authors:
- Will Barnes
- Albert Shih
- Nabil Freij
- Stuart Mumford
- Jack Ireland
- Laura Hayes
- ???
- SunPy at SPD
- SunPy Nuggets--https://github.com/sunpy/sunpy.org/issues/328
- Will, Stuart, Nabil planning demo
- We want live feedback during the session--Slido?
- Hacking/help session planning
- SciPy Postmortem
- Array API
- xarray+WCS
- PR review
- Will submitting AGU abstract today (by 18:00 UTC)
- please make edits by then
- add your name if you want to be a coauthor
- GSOC update
- Array API
- https://data-apis.org/array-api/2022.12/
- we need to think about how ndcube, map will be compliant with this
- xarray discussion
- nominal plan: work on putting together a demo of working with WCS/ndcube, Quantity, Time, SkyCoord to present to xarray devs
- depending on progress at SPD, schedule call with xarray devs some time in mid October
The primary goal of the SunPy Project is to facilitate and promote the use and development of community-led, free and open-source solar data-analysis software based on the scientific Python environment.
To accomplish this goal, the SunPy Project maintains the sunpy
package for downloading, loading, and transforming solar data, in addition to maintaining and supporting a set of affiliated packages for instrument- and analysis-specific tasks.
Together these form the SunPy Ecosystem, a set of interoperable Python packages for solar physics research.
Packages within the SunPy Ecosystem heavily utilize and/or extend capabilities provided by the astropy package.
Additionally, many packages within the SunPy Ecosystem are also part of the Python in Heliophysics Community (PyHC), an effort to coordinate Python software development across heliophysics.
In this presentation, we show how the SunPy Project and Ecosystem promote and enable three aspects of Open Science: Open Software, Open Data and Open Results.
In particular, we show how open development practices lead to high-quality research software within the SunPy ecosystem and how such practices encourage community ownership of software.
Additionally, we discuss how the philosophy of Open Data in the solar physics community combined with the open development model of the SunPy Project allows users and developers to challenge the hidden assumptions in data and maximize its scientific potential.
Finally, we demonstrate that the adoption of software from the wider scientific Python community as well as the development of extensible tools by the SunPy Project lead to a more interoperable Python software ecosystem within astrophysics and heliophysics.
- Welcome to Newcomers
- GSoC
- NF Summit
- Laura going for sure
- have we confirmed Nabil and Stuart?
- SciPy Postmortem
- Array API
- xarray+WCS
- SunPy at SPD
- SunPy Nuggets--https://github.com/sunpy/sunpy.org/issues/328
- Will, Stuart, Nabil planning demo
- We want live feedback during the session--Slido?
- Hacking/help session planning
- PR review
- SolarMonitor alpha release: https://solmon.dias.ie/
-
sunraster
JOSS paper (Danny)- sample data for slit spectrographs
- there is slit spectrograph
- downloading sample data
- should have a more readymade solution for affiliated packages
- right now each has to implement their own solution
- would be nice if this was part of some sort of package template, lived somewhere other than core
- gsoc update
- current behavior: interval is closed
- behavior should remain same as main
- https://github.com/exitflynn/sunpy/blob/main/sunpy/net/tests/test_scraper.py#L325-L326
- Welcome to Newcomers
- e-Calisto discussion
- GSoC
- NF Summit--need to decide who we will send
- SciPy Postmortem
- SunPy Demo at SPD
- PR review
- e-Calisto discussion
- https://github.com/sunpy/radiospectra/issues/98
- https://github.com/i4Ds/ecallisto_ng
- e-Calisto client would be a REST API that would stream data, rather than having to download files
- similar to discussions about HAPI data
- currently no metadata in these files--what should go in here?
- GSoC
- Nabil to finish up HEK changes, at least introduce units to HEK tables:
- Continuing progress on ScraperClient
- NF Summit
- Nabil, Laura, and (maybe?) Stuart will go
- need to make sure we actually follow up with NF ASAP about these plans
- SciPy discussion
- coc changes, want numfocus to just tell us what coc should be
- probably more information at the summit
- better to have more guidance on how to manage coc violations
- 3rd party mediator is best
- cookbooks: https://projectpythia.org/
- maintenance is a worry, online data and bugfixing when ci jobs start failing
- making ipynb files anyway: why not share them.
- consensus seems to be that we should go in the Jupyter book direction
- biggest challenge will be maintaining it--we don't want it to just rot!
- Alasdair, Stuart brought up challenges of using data remotely
- TODO for Will is to just start trying this infrastructure / scope out what could go in here
- deferred xarray/array-api discussion to next week
- coc changes, want numfocus to just tell us what coc should be
- Welcome to Newcomers
- What's going on with APE 21: https://github.com/astropy/astropy-APEs/pull/82
- Some details here https://groups.google.com/g/astropy-dev/c/QArGJfzPVeE
- Talks at SciPy and SPD?
- PR review
- AW to make pr on rsun, combining "why does rsun matter in reproject" and https://github.com/sunpy/sunpy/pull/6954#discussion_r1182488810 perhaps a "Why the solar radius is important in solar coordinates"
- Will to tackle timeseries tutorial refactor:
- Milestoned a few very minor docfixes but the decision still stands that only tutorial improvmements will hold up 5.0
- Coordinates section still needs work, but in decent shape and we can always backport improvements
- Re: scipy, Jack, Will, Nabil will work asynchronously to map out a plan in the next week and half and then meet the week of 15 May
- Welcome to Newcomers
- Scipy Presentation Prep
- SPD Demo Prep
- PR review
- Decided that it is worth holding up 5.0 until we are recently
- SciPy Presentation Prep
- we need to motivate the problem
- "why are we not just using xarrray?" "why NDCube?"
- decide whether we do it live
- we want to work in the ecosystem that we have and we want that to scale
- we also want to illustrate the challenges of the data products we are using
- stacking many coordinate systems into a single coordinate system
- we don't just start with a 3D cube
- what are we going to show?
- analysis with multiple AIA images
- on heliocloud
- scalable data analysis with ndcube
- should also mention current efforts in uniting WCS and xarray: https://github.com/sunpy/ndcube/issues/222
-
Welcome to Newcomers
-
Should we run CI on backport PRs before merging?
-
5.0 feature freeze! ❄️
- Failing Helio tests
- Increase parfive to 2.0 for 5.0
- Docs!!!
-
Presentation prep
- Scipy Presentation Prep
- SPD Demo Prep
- PyHC
-
PR review
- Running CI on backport PRs
- Decided to let backport PR CI run by default, but can merge without waiting in special cases (like lots of backports at the same time that are fixing different tests)
- Review GSOC application by 24 April to be submitted by 27 April
- Need to follow up and find out who is in charge of Helio/HEC because it's not clear whether the service is still maintained/whether we should be maintaining a client
- David has posted to Helionauts
- Next week we will actually talk presentations, specifically scipy demo
- If we do not end up having time, then Nabil/Will/Stuart/Jack will likely need a separate meeting
- Welcome to Newcomers
- Fix https://github.com/conda-forge/sunraster-feedstock/pull/9
- SPD Demo Prep
- PR review
- Issue with
sunraster
conda feedstock: https://github.com/sunpy/sunraster/issues/236 - Good review for
ndcube
ApJ paper, close to the finish line - SPD Demo
- Created repo https://github.com/wtbarnes/spd-2023-sunpy-demo
- Before next meeting, start writing Fido queries to get the data we want from AIA+SolO
- Will to create "How-to-Guide" section to start dumping things
- Next step is to comb through the gallery and start pruning examples that are better suited for how-to guides
- Fate of example gallery unclear...maybe try to convert these to tutorials? (this point not discussed, added by Will)
- Welcome to Newcomers
- Consensus on diátaxis reorganization
- SPD Prep
- Do we need money for anything else besides rooms, AV equipment? Food? Swag? SPD is offering.
- Work on demo, planning
- Solicit ideas from the community ahead of SPD? e.g. announcement in Solar News asking people what they want to know, what they are confused about
- PR review
- Wrap up open PRs from tutorial sprint
- GOES temperature functionality: https://github.com/sunpy/sunkit-instruments/pull/81
- diataxis discussion
- Clarified that the currently named "Guides" section is actually "explanation" (in the language of diataxis)
- Suggested renaming, either "Explanation" or "Topic Guides"
- So we have an "Explanation" section, but we do not have a "how-to guides" section
- Discussion around state of example gallery--so much useful content but an admixture of tutorial, how-to, explanation
- We need to assess the gallery in terms of this division
- Nashpy how-to was brought up as a good example of how-tos: https://nashpy.readthedocs.io/en/stable/how-to/index.html
- one issue with example gallery is that there should be an image with each entry--do not want our how-to guide beholden to that requirement and often not needed
-
Next step: create a how-to guide based on what is in the existing example gallery
- Maybe restructure and simplify some of the gallery examples to be how-to guides--moving explanatory text to "explanation" and removing graphics if they are not needed
- Could also add a "how do I..." table a la
xarray
--useful for one-liners (e.g.save
,plot
)
- Focus of the 5.0 release should be docs revamp--we do not want this to be halfway done when we release
- SPD plans
- Still need to do demo prep (next week)
- Money could be spent on food to attract people, especially during the lunch sessions
- Tentatively create a brief survey for people to fill out prior to SPD to find out what people are interested in hearing about, what kind of content do they want to see
- Alasdair mentioned that SolarMonitor user survey was really successful
- Laura pointed out that if we rerun survey, should be targeted in our questioning
- Too open-ended does not often elicit useful responses
- Brought up point about re-running the survey at some point--maybe pared down
- Welcome to Newcomers
- Any follow-ups from docs sprint last week
- Reorganization of the coordinates docs: https://github.com/sunpy/sunpy/pull/6881
- Last minute thoughts on DASH proposal: https://hackmd.io/ZTx626O3RJ6wAG64gjXDcg
- PR review
- GOES temperature functionality: https://github.com/sunpy/sunkit-instruments/pull/81
- Welcome to Newcomers
- DASH Astropy Proposal
- PR Review
- Docs sprint next week (3/22) 4 PM--7 PM (UK time) (12-3 PM EST US)
- make progress in rewriting the tutorial in the "continuous-sit-down-and-read-it" style
- Welcome to Newcomers
- Advertise Frontiers paper
- DASH Session Proposal
- PR Review
- SPD Plans--wrap-up discussion from last week
- Goals:
- Advertise the wider SunPy ecosystem
- Gather feedback from users that we can act on
- Encourage contributions from the wider community
- Tentative Plan:
- Ecosystem update/Demonstration (working title: "New Era of Solar Instruments with the SunPy Ecosystem")
- Schedule: 60 minute session to be requested during the "Town Hall" block either Monday or Tuesday
- Brief update on existing ecosystem and affiliated packages
- Advertise paper where people can find more information
- Demo of multiple affiliated packages working together
- Concurrent observation of same event/structure with Solar Orbiter and DKIST
- Use both in situ and remote sensing data from SolO--show how the SunPy ecosystem enables "multi-messanger" solar physics in a way that a more fractured/disparate software ecosystem does not
- Show an image with EUI, fit a spectra with SPICE, look at time series data from some in situ instrument, show alternate viewing angle with AIA, look at polarization with DKIST
- Allow people to provide real-time feedback during the session (e.g. via Slido)
- Prompt people to give feedback (e.g. during the presentation, make flyers with Slido QR code on them and example questions, offer stickers for giving feedback)
- Solicit input on what topics people want to hear about ahead of the meeting, e.g. announcement in solar news a month or two ahead of time with survey link
- Use this session as an advertisement for the office hours/hacking session later in the week
- Should run this on Binder or some equivalent platform to allow for people to follow along if they want
- Office hours/hacking
- Schedule: 120+ minute session to be requested during the "Splinter" session before free afternoon on Wednesday
- Hands-on session meant to follow on from the town hall/demonstration
- Appeal to different experience levels: office hours session for less experienced users wanting help with installation, getting started, or maybe even doing their science
- At the same time, have a hacking session for those who are wanting to add features, fix bugs, or even "hack" on their own science question (which may lead to bugs, feature requests)
- Have different stations for different experience levels
- Have some way of systematically gathering feedback during this session as well--follow-up survey? make sure that users are creating issues (or we are creating them for them) during the session?
- Ecosystem update/Demonstration (working title: "New Era of Solar Instruments with the SunPy Ecosystem")
- Questions:
- Is this too ambitious? Not ambitious enough?
- Lessons learned from tutorials/presentations at past meetings?
- What are some things we can do to maximize feedback from users?
- Are there goals that I'm missing?
- Will there be a sufficient number of us at the meeting in order to adequately execute something like this?
- Will DKIST be an affiliated package by the time of this tutorial?
- Goals:
- Newcomers--Welcome Eve!
- Re: the paper, David suggested advertising to the mailing lists listed in the release notes
- should also go to twitter
- DASH
- Brett Morris has reached out to organizers
- We should draft something ASAP, deadline by 3/31
- "Uses of Astropy in the Heliophysics Ecosystem"
- Full draft by next week
- Lots of enthusiasm for SPD tutorial given the plan above
- ambitious, should start as soon as possible
- Nabil, Will, and maybe Stuart will be at SPD in person (the more the better)
- Welcome to Newcomers
- Governance discussion
- Review this in the next week: https://github.com/sunpy/sunpy-SEP/pull/62
- Discourse announcement: https://community.openastronomy.org/t/updating-the-sunpy-projects-governance/584
- The more in attendance the better
- SPD Plans
- Decide what we will present at SPD
- We at least have an hour long "splinter" session
- Do we want to ask for something longer?
- If we have an hour, what do we want to show? Two hours?
- Retrospective on what went right/wrong at TESS...
- DASH Session Proposal
- PR Review
- Governance
- "Benevolent dictator for as long as the board wants"
- New gov structure in https://github.com/sunpy/sunpy-SEP/pull/62
- Involve dev community more in how project is run and make things a bit more democratic
- spread around responsibilities of the lead developer a bit more
- Jack's question: why reorganize?
- Stuart: current governance not really doing much...
- current board not doing much currently and not much interaction between developer community and the board
- financial decision now sits with the lead developer--should make this more transparent
- Maybe most importantly: need a governing body that a funder can give money to if the Project is ever to win a grant
- Maybe pressure from NumFocus as well...
- Within the steering committee, should we just have two lead developers?
- rather than a lead developer and a deputy lead developer
- Alasdair: shouldn't members of the steering committee be project members?
- Project members
- formalize the process of who gets to decide how the project is run
- ultimately this is where the power lies
- project members delegate responsibilities to the steering committee
- as many people as meet the criteria can become project members
- Alasdair: should add coherence to code of conduct to project members criteria
- Don't necessarily have to be members of the solar physics community; Stuart, Albert pointed out that we've had significant contributions from those outside of the solar community
- Debate about whether we should just explicitly list the initial project members in the SEP
- Extended debate about initial member list
- Albert: we should see the list in order to determine whether our criteria our adequate
- Stuart: starting with a criteria, trying not to be exclusionary and to not have the appearance of "an old boys club"
- Albert: presenting it this way is nearly disingenuous because we are really just using it to pick the people we expect to be there anyway and in this way is less transparent and just obfuscates what we are trying to achieve
- Stuart: to some extent, it does not matter because we can always add people after the initial membership
- Danny: argument does not matter because the initial members will add people according to the membership criteria
- Conclusion: not really much of one but debate should continue on PR
- Project member election
- Albert: What stops people spamming nomination?
- Stuart: Astropy batches nominations and votes on some particular cadence (e.g. every 6 months)
- At this point, favoring flexibility as opposed to being overly bureaucratic about it
- Nabil: Is this some "time out" period for unsuccessful nominations?
- Albert: don't box ourselves in to a particular election cycle, may be really beneficial for some to have more frequent elections
- SPD Plans
- Discussion
- Danny: different stations showing off different stations so that people can move around and learn about the things
- Different sessions for developers and for users depending on level of interest and granularity of the question
- DKIST session that could be combined with more developer/power-user session
- "New Era of Solar Instruments with the SunPy Ecosystem"
- Several suggestions that we target users at multiple levels, give opportunities for newcomers as well as more advanced users
- Rough consensus after discussion today:
- tutorial showing off multi-instrument capabilities, perhaps combined with a DKIST demo, would be best; DKIST + SolO would be really great, plus multi-messanger stuff
- also a more hands-on office hours/hackathon
- plan for a 60 minute splinter/townhall
- also ask for space to do office hours/hackathon
- At a minimum we will have 1 60 minute session at SPD
- If we want to do something more ambitious, we need to come to them with a fairly well-scoped proposal of what we want to do
- Q: How would we fill a larger block of time?
- Q: What are are our goals in having a presence at SPD?
- Schedule type (ordered by decreasing probability)
- A: 1 60 min "splinter" session (scheduled during lunch, evening, before "free afternoon" i.e. probably less attendance)
- B: 1 60 min "splinter" + open-ended time during free afternoon
- C: 90 min session (part of normal program, parallel with others)
- D: 2 90 min sessions with a coffee break in between
- Formats
-
Tutorial
- what we've done in the past
- does not seem to fit well within allotted time
- easiest
- could we do a more focused tutorial that shows off newer parts of the ecosystem? different from what we've done in the past? i.e. show multiple affiliated (or adjacent) packages working together to do science
-
Town hall
- similar to what we did at TESS last year
- what would be different?
- main purpose would be to update community on ecosystem
- also gather feedback
- less interactive, more "talking at"
-
Office hours / hacking
- More free form and direction based on user needs
- Offer help to newer users
- Market as hacking session for more experienced users, i.e. fix a bug or implement a feature
- Potential for lowest turnout
- Most informal
-
Contributor workshop
- most ambitious
- "Anatomy of a SunPy contribution"
- Goal would be to bring in new contributors from the solar physics community
- Live demo of how to contribute to sunpy
- Emphasis would be on contributing rather than using, but not limited to just code either (e.g. creating issues, fixing docs)
- Not clear whether this would scale and if we could actually do this in any of our a
-
Tutorial
- Proposed scenarios
- Tutorial, schedule A
- Town hall, schedule A
- Town hall + office hours/hacking, schedule B
- Tutorial, schedule C
- Contributor workshop, schedule D
- Discussion
- Welcome to Newcomers
- Governance discussion next week (3/1/23)
- Review this in the next week: https://github.com/sunpy/sunpy-SEP/pull/62
- Discourse announcement: https://community.openastronomy.org/t/updating-the-sunpy-projects-governance/584
- The more in attendance the better
- TimeSeries
- SciPy Meeting Abstract (due today!!!)
- Submitted
- If you see this before 17:12 UTC, we still have time to change it
- PR Review
- Governance review is in the wild, see above for details. Will discuss next week.
Work funded by a small dev grant, to scope out the options for a new data container for the time series class.
Collated a list of requirements from various mailing lists etc. Came up with 5 options:
- pd
- astropy.timeseries
- xarray
- numpy
- ndcube
compared to what other python packages are using, whole selection other than astropy represented.
Recommendations:
numpy isn't really appropriate, too many missing features. xarray is pretty much a straight upgrade on pandas, only real selling point to pandas is things stay the same. NDCube, stores data with WCS. Timeseries data has different requirements to imaging / spectral data, and NDCube is more focused on the latter data. Also NDCube exists because of WCS, without WCS we would probably be using xarray.
The main comparison is between xarray and astropy timeseries:
Xarray doesn't support different time scales (as indices) but there may be a path to extending it. Xarray also doesn't support physical units. There is some ongoing work, but it's a big problem for the whole scipy ecosystem.
Astropy timeseries has limited out of memory support (columns can be dask arrays). Astropy timeseries has no multi-index / multi-dimensional data. Less mature data manipulation support in timeseries.
Discussion on relative priorities of requirements:
Leap seconds is a known requirement on our currently supported timeseries data. Multi-dimensional data support in astropy timeseries, is a bit of an unknown, you can store the data in the table, but you can't access it based on index. Observer coordinates: Supported as SkyCoord columns in Astropy, but SkyCoord probably not supported in xarray, although xarray's non-dimensional coordinates support the workflow. Swarm spacecraft, PSP, SolO, etc all need to know the spacecraft location, so probably an important future feature.
Can xarray ever be brought close enough to sunpy to feel sunpy or will it always feel different to our other classes. Stuart: xarray brings massive interoperability with sci python, docs, api etc are good, but doesn't have astronomy domain specific things (skycoord).
- Welcome to Newcomers
- Planning for upcoming meetings--who is going to what and do we have a vision for what we would want to do at each of these meetings?
- B.20 Heliophysics Tools and Methods
- Coordination meeting?
- NDCube things (Danny)
- PR Review
- Python meetings
- Nabil is going to PyCon (Salt Lake City, UT)
- scipy
- Nabil
- maybe Will? Jack?
- should be we submit an abstract?
- Birds of a Feather session?
- Abstract deadline
- Working title: "The SunPy Ecosystem in the Cloud"
- where does our ecosystem break down at scale?
- discussions with Dask, Jupyter folks could be fruitful
- Will and Jack to draft an abstract
- Talk about what the limitations are in terms of using scalable compute with our ecosystem
- Have others have encountered these issues before?
- NDCube paper
- Section on scalability of NDCube--does not have to be a numpy array! (Will to add a paragraph about this)
- No in person coordination meeting this year
- Maybe an in-person meetup at Dash?
- Maybe an online meeting?
- Some confusion about how to cite sunpy
- Danny could not find it in the docs--worrying? Are other users not finding it as well?
- Maybe it is too hidden in thedocs
- Add a button to the landing page of the sunpy.org
- SDAC looking to hire a Python developer to support the sunpy ecosystem
- Welcome to Newcomers
- Planning for upcoming meetings--who is going to what and do we have a vision for what we would want to do at each of these meetings?
- SPD (Minneapolis) 8/{13-18}/23
- SPHERE (DC) 6/{20-23}/23
- Hinode/IRIS (Japan) Fall 2023
- others?
- Spreadsheet: https://docs.google.com/spreadsheets/d/1muxbnL2h_pJgUVdEX_1ncrliKQeZzRiTV2BxQtJ_QtY/edit?usp=sharing
- Session proposals for DASH?
- 10/{9-11}/23 (DC)
- NumFocus Small Development Grant
- sunpy v5.0
- Coordination meeting?
- Do we need one? In person or remote?
- When?
- PR Review
- Welcome to Newcomers
sunkit-instruments
-
https://github.com/sunpy/sunpy/pull/6743 and the point of
meta_keywords()
- PR Review
- Quick discussion on Solar Orboter SPICE + other fitting tools available
-
GOES XRS NRT discussion
- where should the code to read these kinds of files go?
- two files needed
- consensus seems to be that it isn't worth a Fido client
- Separate TimeSeries source? Combine with existing XRS source?
- Example for now
-
sunkit-instruments
- Will asking if it makes sense to put code in sunkit-instruments that is too specific for core but can be used across instrument packages
- context: instrument response functions for aiapy and xrtpy
- abstract/base classes for defining an API for computing effective areas and temperature response
- bits of xrtpy are a good starting point for this (e.g. objects for channels, filters, mirrors, etc.)
- No strong objections to starting this effort in
sunkit-instruments
- Does not preclude a move to core in the future (if that makes sense)
- Laura also suggested things like PSF deconovlution--does it makese sense to move this to
sunkit-instruments
? other functionality?
-
Meta keywords function
- Laura wrote this originally for documentation
- We should just make it a table in our docs because this useful information
- Deprecate the function
- change it to make a copy
-
case-insensitive file systems
- avoid filenames that clash with object names (e.g.
database.py
andDatabase
)
- avoid filenames that clash with object names (e.g.
-
SPICE Team fitting tools
- informal chat with people developing spectral fitting for SPICE
- see what they're doing, where they are having problems
- demo of how ndcube and sunraster
- Welcome to Newcomers
-
xrtpy
discussion - Discuss Change all uses of GenericMap.world_to_pixel/pixel_to_world to GenericMap.wcs.world_to_pixel/pixel_to_world aka https://github.com/sunpy/sunpy/issues/6435
- PR Review
- XRTpy functionality being adapted from XRT functionality in SSWIDL
- xrtpy docs
- XRTpy has functionality for representing channels, calculating the effective area & temperature response, and (soon) using filter ratios to diagnose the electron temperature and emission measure.
- Current refactoring about how times are handled
- Anything that is handled by parse_time can be put in to the effective area and temperature response functions
- Channel object holds all the things you need to compute the effective area
- Users likely will
- Roadmap?
- xrt_dem_iterative Python version is most requested
- PSF deconvolution
- Composite images
- Take a long & short exposure, and replace saturated pixels in the long exposure with unsaturated ones from the short exposure
- Need to think about what the FITS header will look like
- Will not recreate
xrt_prep
in xrtpy, will work with level 1 data
- Question about Data availability of Level 1 data
- Archive being transferred over to SDAC, done fairly soon
- Have a common data model for working with temperature responses, etc.
- Where would the data model go? SunPy? Maybe start off in AIApy?
- Harder to do this since it wasn't done in SSWIDL
- XRT & AIA are not the only instruments that need this; also for other X-Ray and EUV imagers too
- Abstract base class?
- Word of caution about trying to program against every possible imager
- Will, Joy, Nick (and other interested parties) will follow up with some rough design of what a more general framework could look like for doing this calculation; plan such a meeting
- From Mark Weber: Quick note on units. For XRT, there are e-, DN, and ergs. You cannot get to photon count, but you can get to total energy.
- NDCube rebin discussion
- This is an API we may be stuck with if we rebase Map on top of NDCube
- Are we happy with it?
- Discussion of meta keywords dict on #6722
- Some debate about whether making a copy of the dictionary (without returning from a function) adequately protects a user from overwriting it.
- Also question of why we need this thing at all?
- Discussion of XRT unit detection #6725
- Discovered that there was an inconsistency between how the dummy test maps exposed the
HISTORY
key versus how actual maps read from FITS file expose it - Now that that is merged, should simplify the unit checking
- Discovered that there was an inconsistency between how the dummy test maps exposed the
- Welcome to Newcomers
- GSoC 2023 Project Discussions
- PR Review
- Are there things that we can do to make it more clear that people should cite sunpy when using it and also how they should cite it
- GSoC
- https://github.com/OpenAstronomy/openastronomy.github.io/pull/330 proposed project ideas
- Database
- Shane: we should scope out what database could be useful for
- Scope currently is very muddy and not clear what the utility is and why it couldn't be accomplished with a custom scraper
- Should not be a general solution for all scales for deploying a database
- The original scope of database is not particularly important
- Welcome to Newcomers
- GitHub Releases release notes (what is their purpose, and can we just link to release notes in the docs?)
- PR Review
- GitHub Releases release notes (what is their purpose, and can we just link to release notes in the docs?)
- Issue at https://github.com/sunpy/sunpy/issues/6134 for any future discussion
- AGU updates
- NASA open source meeting
- Welcome to Newcomers
- Frontiers Paper Referee Response
- Revisions technically due today
- Working on response to referee
- Decide what we actually want to change
- Timeseries Refactor (David S.)
- PR Review
- Timeseries refactor discussion
- David has NF small dev grant for refactoring Timeseries
- decide on a new base data container
- currently using pandas but this has issues (e.g. leap seconds)
- https://hackmd.io/-0y4uN9lSYyAHa_3_LjQ-A
- want to accommodate multi-dimensional data as well (e.g. in-situ data from SolO)
- Welcome to Newcomers
- AGU talk
- Letter to Frontiers Editor
- Send by today
- Extended deadline to respond to reviewer 1 to 14 December
- PR Review
- Question about progress on WCS and xarray
- Welcome to Newcomers
- PR Review
- AGU talk
- Will is giving a sunpy talk at AGU
- Input welcome
- Talk has to be recorded before next Thursday (8 December)
- Title: "The SunPy Project: Past, Present, and Future"
- Abstract: "For over a decade, the SunPy Project has facilitated and promoted the use and development of community-led, free, and open-source Python software for solar physics data analysis. Consisting of both the core library and an affiliated package ecosystem, the SunPy Project provides critical science enabling functionality such as data search and download, data structures for remote sensing and in-situ observations, solar-specific coordinate systems, and a variety of image processing techniques. As such, a wide variety of science analysis workflows and instrument data pipelines depend on the SunPy ecosystem. In this talk, I will provide a short retrospective on the development of the SunPy Project and give an update on the current state of the project, including the latest release of the sunpy core library and the status of the affiliated package ecosystem. Furthermore, I will discuss the broad goals of the SunPy Project moving forward, particularly those enabled by the NASA Open Source Tools and Frameworks proposal. These include community outreach efforts, the relation of the SunPy Project to other Python-based heliophysics packages, support for current and future solar missions, and support for scalable computing in heliophysics."
- VSO health reports discussion
- https://vso1.nascom.nasa.gov/vso/health/daily/daily_vso_health_summary_python.html
- Nice UI for seeing which tests are passing over the last month
- Notes on talk
- ablog
- Pull material from Frontiers paper
- Roadmap
- ndcube
- slide on 4.1 release
- "sunpy needs you" recruitment
- wider ecosystem
- advertise that we have an annual planning session, re: the coordination meeting
- summarize coordination meeting decisions
- people know the core package but they don't necessarily know the other stuff in the ecosystem
- don't necessarily know the difference between ecosystem/project
- Welcome to Newcomers
- PR Review
- Frontiers paper
- got not so positive response from reviewer 1
- there appears to be a mismatch in understanding of what the scope of the paper should be
- some suggestions are helpful
- first action item (for Will): contact the editor to clarify what the purpose of the article should be
- docs sprint summary
- need a place for a
- more sprints
- quarterly?
- another docs one but maybe more focused
- Potentially: Staurt and David to tackle user guide
- docs structure
- avoid more sphinx galleries for now
- appetite for putting most things in RST files
- Welcome to Newcomers
- Nick and Joy from xrtpy 🎉
- Docs sprint scheduled for 18 November: https://demo.hedgedoc.org/8z21gCI8QpeH4wvpzDo-1g?view
- 4.1 released 🎉
- PR Review
- xrtpy discussion from Nick and Joy
- Joy showing the xrtpy poster and docs page
- motivated by what capabilities are in SolarSoft
- main capabilities are
- calculation of the effective area as function of wavelength
- calculation of temperature response as a function of temperature
- future capabilities
- temperature diagnostics
- better integration with CHIANTI
- what does better interoperability between instrument packages mean?
- affiliated package guide is a good place to start
- Laura asked about prep routines
- Nick says no tentative plans for this
- Maybe more definitive discussion with Kathy Reeves
- Some general confusion about what the XRT prep routine does
- some discussion about a general interface for computing effective areas, accessing instrument information
- would be good to define a common API for this
- aiapy and xrtpy both provide interfaces to similar information but the APIs are fairly different
- more and more packages will want to do the same
- better to define some sort of interface now
- Will to follow up with issue on xrtpy trying to scope out an interface
- Welcome to Newcomers
- Docs sprint scheduled for 18 November (unless there are strong objections)
- Feedback from 4.1 RC
- status of testing on all sponsored packages?
- any feedback from other affiliated packages?
- PR Review
- What should we do ahead of the docs sprint?
- Will: map rotation example separated out using the diátaxis framework
- Identify what content we have that falls into the certain areas
- What does our content look like if we split it out
- What has been done thus far?
- Goals of the sprint:
- Plan for making incremental progress towards this goal
- Welcome to Newcomers
- DST and meeting time
- Status of 4.1 release
- Docs sprint doodle: https://doodle.com/meeting/participate/id/b2xxQnAe
- SunPy LINCC talk next week: https://www.lsstcorporation.org/lincc/tech-talks
- Reviews on
eispac
PRs - PR Review
- radiospectra discussion
- pydantic???
-
header_helper
discusssion- should we provide an additional namespace now that we have multiple helpers
- proliferation of namespaces but also of things that live under the
.map
namespace - decided that since there are likely to be more header helpers, makes sense for them to live under their own namespace
- Welcome to Newcomers
- Database deprecation announcement
- Docs Sprint Planning
- Diátaxis refactor is large
- We should figure out a day where we are all working on components of this refactor
- Scope out a plan before hand
- Preferably before the end of year
- Possible dates--need a poll here
- 4 November
- 16 November
- 18 November
- 5 December
- 8 December
- Governance Updates
- PR Review
- Priority list for 4.1: https://github.com/sunpy/sunpy/milestone/90
- JSOC cutout example failing; disable it for now
- Should nail down what exactly is failing--download or the actual query service?
- It looks like its actually the initial VSO query that is failing
- Remove sunpy-dev
- Pending deprecation warning
- Nabil posted to mailing list, no responses
- Use sample data for HMI magnetogram upside down example
- Should have a "How-to" for downloading HMI magnetogram data
- This uses
physobs.los_magnetic_field
which may not be obvious
- This uses
- Stuart thinks most gallery examples should be how-tos, not tutorials
- We need a larger-scale plan about how to accomplish this
- There still seems to be some confusion about terminology in the diátaxis
- Need both common agreement and common understanding before moving forward
- We should come up with some sort of minimal example of this applied to our docs
- e.g. splitting out gallery examples into how-tos and tutorials
- taking some of the coordinates pages and splitting them into the four quadrants
- Welcome to Newcomers
- Paper wrap-up
- Almost done
- Submit by tomorrow (if at all possible)
- PR Review
- Feature freeze is Friday (!)
- What PRs need reviews?
- Anything that we really want in 4.1 that has not been opened yet?
- Review deprecations to make sure we've removed everything
- Frontiers paper
- Submit by Friday at lunch time
- Jack reading over this afternoon
- Will wrapping up the roadmap/governance and community section
- Final read throughs today and tomorrow
- Feature freeze is Friday (10/21)
- PRs that make it into 4.1 should be opened by Friday and reasonably close to complete
- Extended debate about
__author__
/__email__
tags
- Welcome to Newcomers
- Paper progress
- Lots of content is there
- Still needs lots of tidying up though
- Deadline for final draft EOB Friday (14th Oct)
- Will be submitted a week today (19th Oct)
- Devstats are live: https://devstats.scientific-python.org/_generated/sunpy.html
- Anything else we would want to see on this dashboard?
- e.g. plots from our previous papers?
- Sprint Planning
- Bi-monthly or quarterly
- Lasting 1 day (starting in Europe, ending on West Coast)
- Intention is two-fold
- Get core devs working semi-synchronously for a confined period of time on a regular basis
- Onboard new contributors (advertise on Twitter, AAS News, etc.)
- Scope out sprints ahead of time
- Structured, but not too structured...
- Temper expectations appropriately
- SunPy Office Hours at GSFC
- DEMs in Python
- PR Review
- What is the PyHC story?
- Steve asks if there has been any progress agreement on standards for units, time, coordinates
- Thus far, no progress...
- There is a movement to have connectors between them, but no standards thus far
- xrtpy
- ask them to join the sunpy channel
- what can we do to better support them
- also follow up on involving them in the instrument working group
- There is also the "Instrument Working Group" channel
- Reach out to Joy and Nick (maybe ask them if they would want to present at a community meeting?)
- DEMs
- David will be working on DEMs for two weeks after first of the year
- Documenting and wrapping current routines
- Maybe implement MCMC version (time permitting)
- Setup a call for others to weigh in towards the end of the year
- Packaging in Python
- poetry
- Stuart has issue with non-setuptools setups not supporting "-e" option
- interesting, but too new?
- We can use modern setuptools stuff and get most of the way there
- Don't need a
setup.py
file to use setuptools
- Roadmap
- Solo Workshop
- ndcube releases
- PR Review
- Welcome to newcomers
- Standardized Metadata SEP: https://github.com/sunpy/sunpy-SEP/pull/61
- Frontiers paper status
- ASV
- Status Albert's PR: https://github.com/sunpy/sunpy/pull/6392
- What does this look like on PRs?
- What is the role of ASV in our workflow?
- How do we use ASV to evaluate PRs? visually? quantitatively?
- PR Review
- Make sure data examples achieve a single task #6384: https://github.com/sunpy/sunpy/pull/6384
- Deprecate timeseries peek() positional arguments #6310: https://github.com/sunpy/sunpy/pull/6310
-
Metadata SEP Discussion
- Question of whether this should be done in SunPy
- This is an ontology defined in Python objects
- Trade study on existing ontologies and understand which ones are useful before completely inventing a new one
- We want complexity Python objects--do we also want the components blown out as separate attributes to be accessed?
- e.g. distance from the Sun to observer as its own property
- Living document that can be updated as we find more metadata that should be part of the standard
- Common metadata that people want
- Should it be an SEP?
- Is it an ontology if it is constantly being added to?
- What is the criteria for adding to this metadata "standard"?
- Not all metadata objects have to use all of these
- but if they do use that, they should use this name
- Provides a vocabulary with expected return types and also meanings
- We should leave WCS out of this (a harder problem to solve)
- Map kind of already does this, we use the Map properties as an informal/implicit metadata standard
- This is just formalizing that existing vocabulary
- Can we at least agree on the names?
- Then maybe take a stab at implementing it in Map and see how it goes
- What existing solar/heliophysics vocabularies can we pull from?
- VSO - https://umbra.nascom.nasa.gov/vso/docs/dm_1.8.html
- EGSO - https://hesperia.gsfc.nasa.gov/ssw/vobs/egso/doc/EGSO.Data.Model.v1.4.pdf
- SPASE - https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2018SW002038
- HAPI - https://github.com/hapi-server/data-specification/blob/master/hapi-3.0.0/HAPI-data-access-spec-3.0.0.md
- More general astrophysics vocabularies
-
Kevin is working with three graduate students taking a course on scientific software engineering who are looking for a semester-long software project
- Where is the best place to contribute?
- Concern about amount of work on the side of the maintainers
- Open PRs early, come talk to us early
- December deadline for the course
- Engage with the process from the beginning
- David just opened a bunch of issues on
sunkit-instruments
that are scoped relatively small--good intro to opening a PR, making a contribution - Dan has mentioned that
ndcube
has some less esoteric issues open at the moment as well - See the "Good First Issue" tag on multiple repos
- Also see Newcomer's Guide: https://docs.sunpy.org/en/latest/dev_guide/contents/newcomers.html
- Potential projects
-
sunkit-image
- GOES EM/T functions in
sunkit-instruments
(Dan/Laura) - Refactor of spectral fitting code in
sunkit-spex
(Dan/Shane) - Adding resampling and/or arithmetic operations to NDCube (Dan)
-
-
Paper discussion
- Continue with use of "affiliated"
- Aim to have a full draft by end of September; will give us two weeks to iterate, make corrections before submitting ~15 October
- Welcome to newcomers
- Frontiers paper status
- We've been given an extension until mid October (no more extending!!)
- TODO: divvy up sections and start writing
- TODO: make a "science-y" figure
- TODO: make a figure laying out project structure
- Small development grant application: https://docs.google.com/document/d/1bS6mFd3XJ7cbHvgpPfdVmFLFErLOEYqeeN_ndeLQNY4/edit
- PR Review
- Make sure data examples achieve a single task #6384: https://github.com/sunpy/sunpy/pull/6384
- Deprecate timeseries peek() positional arguments #6310: https://github.com/sunpy/sunpy/pull/6310
- Mask values when reading CDF files #5956: https://github.com/sunpy/sunpy/pull/5956
- Discussion about https://github.com/sunpy/ndcube/pull/541
- Suffers from issues with dask bringing arrays into memory
- Frontiers paper
- Deadline is mid-October
- Discussed plan of action for dividing up work
- Divided up work
- Discussion on affiliated/sponsored nomenclature
- present: Will, David, Nabil, Alasdair (Jack, Laura and Danny present for some but dropped off before the discussion concluded)
- use of the terms "affiliated" and "sponsored" is (potentially) confusing
- everything is an "affiliated" package
- those that are maintained by the project are sponsored packages, i.e. the project is responsible for their maintenance and direction
- affiliated packages are maintained by the community (not necessarily the project) and must meet criteria for being an affiliated package as established by our affiliated package criteria: https://sunpy.org/project/affiliated
- It was proposed that we make the following changes to the nomenclature around our package structure:
- Do away with the terms "sponsored" and "affiliated"
- Define "The SunPy Ecosystem" to include all packages which we previously considered "affiliated"
- The requirements to be part of the Ecosystem would be the same as those laid out by the current affiliated package requirements
- On our website, we would then list each package in the Ecosystem and list who maintains it.
- Those packages previously referred to as "sponsored" would just be listed as maintained by "The SunPy Project" (e.g. sunpy, ndcube)
- All others would list the appropriate person/organization as their maintainer (e.g. David Stansby for pfsspy)
- The core package (sunpy) is part of the ecosystem
- If we want to make this change, it would require a new SEP or changes to the existing affiliated package SEP
- Regarding Figure 1 for the Frontiers paper
- it was suggested that we have an umbrella labeled "The SunPy Ecosystem" and then have all packages underneath, including the core package
- these could be grouped by functionality but perhaps not explicitly so
- All packages live under the umbrella
- We seemed to decide that we don't want to indicate any sort of hierarchy in this figure
- Nabil suggested we should start thinking about logos for the other packages in the ecosystem
- It was suggested that we may want to have a separate logo for the package and the ecosystem
- Keep the current logo for the package, come up with alternate logo for ecosystem?
- After action report from TESS
- Coordination meeting logistics
- Coffee??? We at least need a plan for where it will be picked up from
- Remote participation setup--what needs to be done on the ground before we get there?
- Identify a presenter for each session
- Master running notes doc: https://demo.hedgedoc.org/fqLNg0xpStac_M9v92v99Q?edit
- Everyone can contribute notes here
- Post this on the meeting webpage
- Another Twitter announcement--when?
- Others?
- PR Review
- AGU Abstract (due today!!)
- TESS Session Planning
- Another Tweet on Monday (i.e. the first day of the meeting)
- Coordination meeting logistics
- Coffee
- Remote participation setup--what needs to be done on the ground before we get there?
- Instrument working group
- Finalize agenda by 15 August
- Identify a discussion lead for each session by 15 August
- Another Twitter announcement--when?
- Others?
- Contributor/Reviewer Checklist
- Build contributor checklist from this: https://github.com/sunpy/sunpy/pull/6346
- Reviewer checklist as well? Maybe auto-generated once the PR is moved from "draft" to "ready for review"?
- Backport bot
- Infrastructure priorities
- Coordination meeting
- Stuart to follow up with Shane re: setup for streaming
- Coffee can be expensed to OpenCollective, but we need to figure out where to buy it...
- Agenda still needs some shuffling around
- Recent bugfix release (3.1.8)
- Thanks Nabil and David!
- User feedback pointed this bug out to us and it was fixed in ~hours, released in days
- A success story!
- TESS Session Planning
- Tweet (Laura/Stuart)
- Solar News/UKSP (Will)
- Discussion questions for second half of session
- Coordination meeting logistics
- Coffee
- Remote participation setup
- Instrument working group
- Others?
- PR Review
- TESS Session Planning
- Will: Put together slides for guiding discussion
- Discussion questions for second half of session
- What has been your experience using SunPy? The good? The bad?
- What is missing from SunPy that you would like to see?
- What would you like to be able to do with SunPy that you can't currently do?
- What do you think about SunPy's online documentation and examples?
- If you're using Python packages in place of SunPy for your solar physics data analysis, why is that?
- Do you know how to contact SunPy if you have a question?
- Do you know what PyHC is?
- Do you use SunPy with other parts of the PyHC ecosystem?
- Coordination Meeting planning
- Need to add an item for roadmap discussion
- Who to invite for scalable computing discussion
- Raphael Attie
- Brian Thomas (and ask anyone from APL...Sandy Antunes, John Vandegriff)
- Mark Cheung / Paul Wright
- Frontiers paper deadline is now 31 August
- Coordination meeting logistics
- Coffee
- Remote participation setup
- Others?
- TESS Session Planning
- 6-7:30 PM on 10 August (Wednesday)
- Rough outline of program
- Advertise session (Twitter, mailing lists, etc.)
- Notes from SciPy?
- PR Review
- Frontiers paper delayed
- Error messages re: failed downloads
- What's the best way to propagate information to the user about failed downloads/unavailable servers?
- Need to close loop between users and upstream data providers
- Better error messages spit out by Fido?
- TESS Session--finalized for Wednesday evening (8/10)
- SunPy Tutorial at SPHERE Workshop
- Finalize Coordination Meeting Schedule
- We should post a first draft of this on the webpage by the end of the week
- https://demo.hedgedoc.org/C9AvEImbTyav1NIy36PuZg?edit
- Thoughts on correspondence with instrument teams
- PR Review
- Frontiers Paper Progress
- Coordination Meeting agenda
- PR Review
- Frontiers paper progress
- Will will add text from proposal
- Will and Nabil to reorganize that text
- Need to think about a coherent way to present affiliated package organization and justify/flesh out our philosophy
- Coordination meeting agenda
- See the agenda doc here
- Instrument Team Day notes:
- Instrument presentations (state of analysis tools, show us what you're doing)
- HERMES (Steven Christe) -- instrument package template
- Feedback forum
- First, focus on functionality
- Then, foucus on standards
- It is on us to demonstrate why an instrument package would want to foster interoperability
- We should demonstrate why playing in the ecosystem benefits an instrument team
- "Why would you want to use sunpy"
- This is mostly focused on user-tools, but what about processing pipelines as well?
- HERMES, PUNCH, MUSE conversation
- We want to build an ecosystem
- How close to the ecosystem do you want to be?
- being in the ecosystem doesn't necessarily mean being an affiliated package
- Where are the gaps
- What could sunpy be doing better?
- Are our affiliated package requirements too onerous?
- Feedback on the affiliated package requirements/process?
- Instrument presentations (state of analysis tools, show us what you're doing)
- TESS Meeting session
- Aiming for a ~2 hr session Monday evening
- will cost $650 but should be able to pay with SPD money
- unknown if remote participation will be possible (it is not a hybrid meeting...)
- Thoughts?
Title: SunPy Showcase and Community Forum Description: We invite members of the solar physics community to showcase the interesting ways that they've used SunPy in their research and to contribute feedback to guide the direction of the package and the future of the SunPy ecosystem. The first half of this session will feature a lightning talk round in which users are invited to present a single slide for 3 minutes showcasing how they've made use of SunPy in their research. In the second half of this session, we will hold a community forum to discuss how the solar physics community at large is making use of the SunPy ecosystem, with the discussion informed by the preceding presentations on uses of SunPy. Several core SunPy maintainers will be present to give a brief introduction to the ecosystem and guide discussion. In particular, we would like to know where researchers are making use of SunPy, where they are not using SunPy, and where they feel the strengths and weaknesses of the ecosystem are. We invite all members of the solar physics community to contribute feedback on the current state of the SunPy ecosystem, whether they are actively using Python and SunPy or not. Additionally, we encourage participation and contributions from both senior and early-career members of the community.
- Coordination meeting planning
- Check in with Shane re: LOC things
- What do we need to do "on the ground" prior to the meeting?
- Agenda: ideally we would firm this up by 1 July
- Accommodation?
- Oh god
draw_
methods make it stop: https://github.com/sunpy/sunpy/pull/6251#pullrequestreview-1006105303 - PR Review
- coordination meeting
- user feedback session from those at DIAS/Trinity
- coffee catered, lunch we will go out (adhere to schedule for sake of remote participants)
- Shane will chase down guest Wi-Fi codes
- IT needs--Stuart will handle machine for streaming to Jitsi, Shane to contact IT about audio in the room
- need to set agenda soon-ish--Will and Stuart to draft agenda, people can add to it as they see fit; main motivation is that participants coming in for only a session or two know when they need to drop in
- Draw methods
- proliferation of GenericMap API--too much code in mapbase (Stuart)
- Can we just move all
draw_*
methods to visualization?
- PyHC Summer School Debrief (Will and Stuart)
-
SunPy Frontiers paper
- Due 29 July--we should begin a draft now
- Will submitted the abstract with Stuart, Nabil, and Laura as coauthors (Authorship on the paper itself can be changed at time of submission)
- This is a high-level overview of the package and affiliated packages (little to no implementation detail)
- This is going to be expensive--is it worth it?
- Presence at TESS
- Response from SPD: we can have a shorter afternoon session or a longer evening session
- Latter may compete with other evening events
- Need to decide ASAP
- SunPy nuggets
- Put call out for people who have done interesting things with SunPy
- Publish a (short!) blog post summarizing what they've done as notebooks
- Stuart suggests having a few in the pipe before we start putting them up
- Could give them on a DOI via Zenodo such that they're citable (people can get credit!)
- Coordination meeting planning
- Check in with Shane re: LOC things
- Agenda: ideally we would firm this up by 1 July
- PR Review
- SunPy Frontiers paper
- Can use ROSES money
- there is confusion about core versus affiliated package--good to have something to point people towards
- can use the text from the ROSES proposal as a starting point
- Laura will create Overleaf template--move to GH later
- Should think about using showyourwork: https://show-your.work/en/latest/
- SO/PyHC wrap up
- Lots of interest in contributing to SunPy from those at SO Workshop
- support for SO in-situ data is an area we need to improve in--lots of demand!
- Established relationship with SOAR--hopefully close loop when users report problems to us
- Laura said some people at SO Workshop still using HelioPy for in-situ data--should identify what functionality in HelioPy has not been covered by other packages
- TESS
- Albert suggested asking participants for 1 slide showing a cool things (i.e. research) they've done with SunPy
- Maybe use this as a jumping off point for a community forum about what functionality users want to see in SunPy
- We want to keep this relatively high-level--should not turn into a tech support session (but gladly will do that other times throughout the week!)
- Suggestion was that Monday evening would probably be best--people still energized and then gives them rest of week to think about any other problems they want to discuss
- Could even hold "SunPy office hours" during coffee breaks (very informally)
- ASV Demo for Albert?
- "Getting Started with ASV"
- Presence at TESS
- Draft of abstract for Frontiers article
- Should we support integer data?
- See Will's issue about rotate and
missing=NaN
- See Will's issue about rotate and
- Meta (Steve)
- PR Review
- Presence at TESS
- Email organizers and ask for room (Dale Gary is a good first contact)
- Charge? Use SPD money for things like this..
- Hybrid meeting? Should we figure out how to broadcast this regardless?
- Is this a good idea?
- SC: still continegent of people that know nothing about SunPy
- The distribution of advanced/beginners is bimodal
- Formal/informal discussion: discussion of what you like/dislike about SunPy, anatomy of a PR
- Live poll for questions
- Some way to capture the feedback we get from the community
- Frontiers abstract (https://www.frontiersin.org/research-topics/33555/snakes-on-a-spaceship-an-overview-of-python-in-space-physics)
- Outline the difference between the package and project
- Say what is in core, what are the primary capabilities (coordinates, data download, data structures)
- Summarize the current affiliated package landscape (new ones! aiapy, pfsspy) as well as possibly new affiliated packages? (e.g. sunxspex, fiasco, ....)
- Probably worth summarizing our affiliated package review process as this is new post our 1.0 paper...
- what is our vision for the future (e.g. standardizing on NDCube)
- Better support for reprojection
- Up and coming high level features: rebasing on NDCube, metadata model, scalable performance with Dask
- Adopting sunpy-soar: Maintenance promises and precedents.
- Will got a response back from the help desk re: SOAR query errors--open up line of communication this way? or better for those closer to SolO (i.e. Laura/Shane) to make this contact
- Presence at TESS
- SunPy after-hours tutorial per usual?
- Do we have money/space to do this? Would SPD/AAS allow it?
- Who do we contact about this?
- Too late in the game?
- "New contributor" mini workshop?
- Anatomy of a sunpy contribution
- Do a live PR submission and review cycle
- Rather than demoing how to use sunpy, how to contribute to sunpy
- Emphasis on contributions that are not just code too!
- Can we reasonably do this in a few hours?
- PR Review
- Nabil: move all external clients besides JSOC and VSO to an external package?
- Enables us to make changes to these clients not on a usual sunpy release cycle
- No new clients in core?
- Will and Stuart to meet in person with SOAR team at ESAC in late May
- Will needs to follow up with TESS organizers re: getting space for workshop
- Seems to be universally agreed upon that this is a good idea
- PR Review
- Please register for coordination meeting if you've not already
- If we don't have a true critical mass of in-person people, is it worth considering going virtual-only?
- Frontiers Snakes on a Spaceship submission
- consensus is that we should submit something
- abstract deadline impending and submission deadline as well (29 July)
- Laura will contact Sophie re: scope of the papers
- Think about what we want to emphasize: what science does sunpy enable?
- Also start thinking about another ApJ paper for v5.0...
- Docs gallery refactor using things from Jupyterbook; see https://github.com/orgs/sunpy/projects/1
- Sign up for the coordination meeting! https://sunpy.org/project/meetings
- Upcoming release of v4.0
- Need for RC2?
- More milestones?
- Anything that really needs to make it in?
- PR Review
- Coordination meeting
- Things to add to the website:
- hotels
- google map
- travel funding information
- Shane will look into costs for coffee/snacks
- Maybe can use numfocus money
- May need to collect small amount of cash from everyone
- Who do we want to invite from instrument teams?
- SolO--particularly in-situ stuff (EPD, RPW)
- PSP
- PUNCH
- EUVST
- Radio people--LOFAR, EOVSA, solar with ALMA
- would be nice to get as many in person as possible!
- Agenda
- have a relatively
- Breakdown of full session versus break out sessions?
- Main lecture hall
- 3 side rooms
- Maybe 1 office
- This will depend on number of participants and who comes
- Things to add to the website:
- SciPy BoF Session ideas? https://www.scipy2022.scipy.org/bof-sessions
- Coordination Meeting Announcement (https://demo.hedgedoc.org/C9AvEImbTyav1NIy36PuZg?edit)
- Duration: 4 days? 5 days?
- Hard cap on in-person attendance? I've put 40.
- Mailing lists: sunpy, sunpy-dev, UKSP, Solar News, Helionauts, OA Discourse
- GSOC
- PR Review
- Discussion of what "draft" means
- See #6085
- Should the context of what "draft" is be in the PR template?
- Need a consensus about what we mean by "draft"
- Does GH have a way to adjust permissions on who can mark as draft?
- SSL Certificate Verification Failure on things that access NOAA data?
- https://github.com/sunpy/sunpy/runs/6083862087?check_suite_focus=true#step:8:23179
- The NGDC NOAA domain is currently undergoing maintenance until the 21st https://www.ngdc.noaa.gov/index.html, this is probably the issue
- PR Review
- Skip the NOAA tests? Or don't check the SSL cert?
- Tell someone they need to fix their certificates
- Will and Nabil giving short PyHC presentation on affiliated package ecosystem
- Draft status discussion
- shouldn't mark as draft without explanation
- Will: should "mark as draft" be used as a review tool?
- Stuart: particularly in context of new contributors, should encourage them to mark as draft if we think the PR is not ready for review (e.g. needs significant work before doing review)
- Firming up coordination meeting plans
- Timeseries refactor PR: #5834
- PR Review
- We need to book a room/firm up dates for the meeting
- Will will get in touch with Shane to book room at DIAS for 22 August barring any major objections in the next few days.
- Significant discussion on Albert's new affine transform PR
- Agreement that default for missing in rotate should be NaN
- Timeseries PR review still needs to be addressed--Stuart and Will will provide reviews
- Coordination meeting--quick planning meeting on Friday?
- WE WILL BE ADDING MAXWELLS AND STUART CAN NOT STOP ME
- Status of ASV?
- Deprecations
-
sunpy.io.fits
-- going private -
sunpy.image.coalignment
->sunkit_image.coalignment
-
sunpy.physics.solar_rotation
->sunkit_image.coalignment
(in progress) -- we need to refactor coalignment capabilities insunkit_image
-
GenericMap.shift
->.shift_reference_coord
(in progress)
-
- What features to try and get into 4.0?
- Feature freeze is in ~1 month
- Current list at https://docs.sunpy.org/en/latest/whatsnew/4.0.html
- PR Review
- Ideally would like to lock down a date by Monday
- Quick planning meeting Friday?
- Will will put a message in the chat
- ASV is up and running
- https://github.com/sunpy/sunpy-benchmarks
- Only a minimal set of benchmarks exist currently
- We should be strategic about what benchmarks we add and when
- Adding new benchmarks and running them over the history of the project takes a long time (~6 hours)
- For PRs, comparison is just between the PR and main
- Weird regression in
parse_time
?
- Milestoning more things for 4.0
- Good progress on changelog bot
- Fill out the coordination meeting planning survey if you have not already (https://forms.gle/dtaYfaw2csDqUoaW6)
- SunPy Tutorial at PyHC Summer School
- GitHub Actions
-
physics.solar_rotation
(https://github.com/sunpy/sunpy/issues/5965) - IRIS SJI Maps (https://github.com/sunpy/sunpy/issues/5962)
- PR Review
- Should there be a models subpackage in core? (Dan)
- Models would have to be physics-based and solar-specific, e.g. thermal X-ray bremsstrahlung using solar abundances.
- Models would be widely used/integral to at least a sub-field of solar physics.
- Models are stable over years, i.e. are rarely going to change.
- Subpackage would not contain complex tools like fitting infrastructures.
- What models should go in? What should the bar for inclusion be?
- Should this even be in core?
- This seems like a topic for the coordination meeting...
- (Jack) I have to leave, but one issue for the coordination meeting an issue I would like discussed is the use of SunPy in cloud environments. Are there changes to SunPy that could enable it to use a cloud environment more efficiently? We should have more experience to report on this topic at the end of the summer.
- (Will) It depends on what you mean by "enable it to use a cloud environment more efficiently", but I think this is somewhat covered by our OSTFL proposal--one of our proposed improvements is improving support for large datasets, e.g. making Map operations Dask-friendly.
- GH Actions migration is going well/mostly working
- Mapsequence derotate
- Agreed it at least needs to be renamed
- Deprecate and put into
sunkit-image
?
- IRIS SJIMap needs to be fixed, but also will be deprecated soon (in favor of irispy)
- Coordination meeting planning survey: https://forms.gle/pj74HicDgjBAWcj36
- PR Review
- Please fill out coordination survey
- PyHC have asked about presentations for some affiliated packages
- More general presentation on affiliated package ecosystem?
- PR "cheat sheet" (see https://github.com/sunpy/sunpy/issues/5944)
- should this go in the template?
- is the current template useful? reevaluate?
- 2022 Coordination Meeting
- Hybrid format
- Proposed locations: Washington, D.C. (American University) or Dublin (DIAS)
- Proposed dates: mid to late summer--will depend on location
- Duration? 4 days? 5 days?
- Will needs to put together a larger poll to get a sense of best date/location combination
- Policy for supporting multiple versions of test deps (https://github.com/sunpy/sunpy/pull/5926)
-
conda-forge
errors: https://github.com/conda-forge/sunpy-feedstock/pull/114 - CI Status?
- Kicking off instrumentation working group again
- PR Review
-
conda-forge
errors- asdf problems
- Stuart cannot look at them at the moment
- Skip the tests?
- Something wrong with
asdf
on conda? - Trying pinning to 2.9.2?
-
asdf
onconda-forge
is broken
- Test deps
- We now support versions of test dependencies released in the last 12 months
- CI
- Work is happening upstream in openastronomy repositories to replicate Azure functionality in Github Actions
- Conor suggested adding Apple Silicon M1 builds
- Note that these would be untested because we don't have CI there
- Coordination meeting
- DIAS is looking like a more attractive option
- Poll needed, but summer is ideal
- Is there a meeting we should be looking to coincide with? (maybe the SO workshop...)
- As far as American meetings go, there is SPD/TESS in Seattle
- TimeSeries summary
- Continued saga of Azure builds
-
conda-forge
feedstock for affiliated/sponsored packages - GSOC?
- Comments restructuring of examples: https://github.com/sunpy/sunpy-project/issues/11#issuecomment-1043342482
- Kicking off instrumentation working group again
- PR Review
- Timeseries quicklook
- Matt gave us a demo of the new quicklook functionality being developed for timeseries
- Plotly and non-plotly versions--very neat!
- There is significant resistance to adding plotly as even an optional dependency
- Healthy debate on purpose of plot versus quicklook methods
- It would be cool to support plotly (other plotting backends), but we are largely bound to matplotlib because of WCSAxes
- David points out that matplotlib does do interactive plotting in the notebook
- Stuart: people can pull their data into any kind of visualization framework
-
bokeh
is another option too (David: the docs are really great!) - Laura: used both
bokeh
andplotly
, found the former to be more user-friendly. - Potential Topic for coordination meeting: how much work do we do to support other kinds of plotting?
- Azure
- Stuart has call with them soon
- Is it time to just move on? to GH actions
- David: basic testing and doc are working in the
sunpy/sunpy
repo - Stuart: need to recreate the OA template repo
- David: Will open issue to collect TODOs for the template: https://github.com/sunpy/sunpy/issues/5913
- Feedstocks
sunkit-image
- GSoC
- All of our projects are listed (Will/Stuart still need to scope out DB project a bit more)
- Not as much OA participation as in previous years
- Examples
- Laura/Shane did straw poll of students and many said searching for examples was an issue
- Even core contributors have issues finding which examples show what
- docs.sunpy.org could become docs for the ecosystem, not just for the core library.
- Stuart: have to be careful about breaking URLs
- We should be motivated by a plan to restructure our docs/guide, not just a particular technology choice
- Release blockers for 2.1
- GSoC Projects
- Anything else?
- NDCube 2.0
- QueryResponses
- 2.1 Schedule
Issues to be tackled, but not much time between Stuart and Danny at the moment. All contributions welcome.
This is the last community meeting of 2020
- NDCube 2.0 API and how it relates to sunpy Map
Dan Ryan presented a notebook showing a comparison between Map
and NDCube
2.0: https://github.com/DanRyanIrish/notebooks/blob/master/ndcube_as_map/ndcube20_as_map.ipynb
- Side-by-side comparison of Map and NDCube in their current states
- NDCube strictly follows NumPy array ordering in pixel space (row, column)
- Can you create a
WCSAxes
instance ahead of time and then pass it intoNDCube.plot()
?- Will will try this and follow up
- This is easy in
Map.plot
, should be easy inNDCube
-
submap
->crop
-- generalized N-dimensional cropping- Spatial bounds are now
SkyCoord
- Shouldn't have to specify bounds in dimensions that you don't want to slice
- Spatial bounds are now
- NDCube does not assume alignment between pixel and world space
- In HPC, can never have true independence between lon/lat because of projection of 3D image onto 2D screen
- Can animate any arbitrary number of axes
- Global coordinates
- How do we go about reimplementing
Map
on top ofNDCube
?- First pass: just try to reimplement
Map
exactly - This will reveal rough edges, pain points
- Opportunity to rethink API for 3.x?
- First pass: just try to reimplement
-
Stuart: Maybe we should advertise our weekly community meetings better (Some outreach work in order?)
-
PUNCH (Guests: Matt West and Dan Seaton)
- To be open-sourced and based in Python
- To be afiliated with SunPy (in the works)
- Have a frozen version to be remained untouched (kind of like SunPy's LTS?)
- GitHub to host the new addition from PUNCH to SunPy
- PyPi for release with
pip
(Maybe conda as well?) - Laura suggested talk about package template for affiliated package applications (i.e., packaging for "structure" stuff)
- Matt expressed concerns about accuracy or relevance of latest version of official SunPy docs
- Matt: Actual transform code
- Stuart: Elements.io as the de facto SunPy IRC channel, bridged with Slack via Matrix
- Matt: Concern about conversing about code dev over video telecons / call
- Matt: Have specific questions to ask
- PUNCH
transforms
package- Ingest data products as FITS files
- For specific transformation between different coordinate systems
- Using an azimuth for coordinate transformations
- Optical astronomy/ system - distortions in the imaging (due to various factors such as atmospheric aberrations?)
- Polynomial distortion algo needs to be worked into (interfaced with) SunPy's
map
module - Can use Astropy
wcs.WCS
to handle the projection - (Thanks so Solar-Probe)
- SunPy map objects - represents meta translator / standardizer (for the conversion between coordinate systems)
- Aiming for APE 14 compliance
- I/O issues as observed by Matt (exporting/importing after transformation), cannot be regenerated with SunPy currently without a patch
- Stuart: A valid FITS header conforming to APE 14 standard should allow such I/O issues to be resolved
- Matt: Issues with WCS (only recommondations for nonlinear projections), will need to adopt a non-linear approach upstream by Astropy (Action item: PR's to follow?)
- Recommendations: "Perfect" FITS (with proper header specs) format for input data products should work with existing Astropy/SunPy ecosystem
- Matt: Feel there is a need for a community docs for best practices for FITS handling for solar physics? (Answer: A recent publication as suggested by Stuart)
- PUNCH is currently working feverishly to develop code products for software infrastructure
- FITS header in solar physics observed to be in dire need of standardizaiton
- Matt's bigger question: Having SunPy involved in PUNCH would be mutually beneficial
- Application as a transform package (suggested IRC channels? workflow for integration between PUNCH and SunPy)
- Dan R.: Incorporation of use of NDCube in PUNCH work?
- Dan R.: Respect PUNCH is mission-based
- PUNCH: Open to relying on SunPy chatroom for the time-being, with specific questions to be redirected to PUNCH specific channel later (as suggested by Dan R.)
- Stuart: Affiliated package application is currently being updated to remain relevant, all are welcome to open an issue there
- Matt: What happens to deprecated affiliated packages (Stuart: As long as it works we would keep it, despite it being outdated, unless there is serious issues with the existing package)
- Matt: Concerned for time to process the application (Laura clarified), worried might get stuck
- Stuart: SunPy is very open to feedback from PUNCH
- Matt: See SunPy as a potential "wrapper"
- Stuart: Expected something that are missing from the transforms package originally written in Perl by Dan S.
- Matt: Package architecture developed about 10 years ago, so may need reworking
- Matt: Some mem tests have been conducted on the existing codebase as benchmarking
- Matt: Jacobian in there probably will not be rewritten as it is written (def: a matrix for storing 2nd order derivatives for the transformation?)
- Craig may be attending future SunPy meetings
- Specific Python questions... (to be continued)
- Superclass with a whole bunch of methods
- Subclass to reverse the name of two of the method the superclass for some inverse transform but would like to call forward transform (a layered superclass)
- Super-superclass
- Stuart: We will need to see the code
- Albert: Should be doable, but... (with caveats)
- Not private functions, but protected ones to be manipulated
- Will: We would like to see more interaction with instrument teams
- Laura: PUNCH contact emails are kind of hidden, but found
-
NDCube 2.0 (post Beta)
- Dan needs feedback
- Could be useful if more feedback can be gotten regarding the API
- Previously received feedback is too technical and not general enough
- Dan would really want new user to try it out and give honest feedback regarding functionality
- Before next week would be great (hoping possibly before Christmas)
- A focused discussion during next week's / next next week's community meeting would be great
- Ideally NDCube 2.0 should come out before SunPy 2.1, at least that's the expectations
-
2.1 PR's, etc.
- Metadata is holding up branching, needs another review
- CI is largely passing
- Stuart will give it a review one more time
- Doc PR needs week before feature freeze
- Will need to update "What's New" as well
- 2 PR's open for a bug open on the googlemail mailing
- Can backport the fix for only one of the two
- The CTYPE's been missing
- Should default to a particular projection instead of leaving it "open"
- Pixel-to-world bug needs attention
- World-to-pixel returns an array of floats, wouldn't be able to used in submaps directly, will need to be multiplied by pixels first...
- Basically boils down to: What do submaps do (by design as well as construction)?
- Stuart: WCS attributes coming from an NDCube perspective is fundamental
- Stuart: Why are we special-casing particular attributes
- Not to milestone the self tests (according to Nabil, someone says...)
- Will: Too hasty in approving a PR before some real CI/CD check failures
- Stuart: We do have the option of adding some "checks and balances" to the existing CI/CD mechanism as a secondary safeguard (for foolproof of genuine test failures)
- Stuart working on Azure configs for CI/CD checks (in "Stages")
- Stuart: API breaking stuff should be one thing we should be avoid doing
- (Maybe a warning to the console should be in order?)
- Dan: Deprecation could be a solution
- Stuart: pending deprecation warnings to the console (so as not to yell at people at their face figuratively)
- Laura: many people complains about SunPy morphing / evoling too fast
- Stuart: warning fatigue observed
- Albert: Should encourage adoption and usage of WCS for it to gain traction
- Stuart: wary of making significant changes to the
map
API - Dan: Would like to understand the metadata problem mentioned more (Stuart: why the
map
approach for the metadata transformation) - Stuart: Metadata issue much like the polynomial distortion correction / fitting algo raised by PUNCH
- Dan: generic data object to a map object (for which Stuart agrees)
- Metadata is holding up branching, needs another review
-
NASA ROSES grant application (1st call)
- Albert is in the process of writing the proposal for SunPy to be due in mid-January 2021
-
Outreachy
- Skipped for this week as agreed with mentors
- 2.1
- 2.1 pull Requests (https://github.com/sunpy/sunpy/pulls?q=is%3Aopen+is%3Apr+milestone%3A2.1)
- Core dependencies https://github.com/sunpy/sunpy/issues/4655
- 2.1
- 2.1 pull Requests (https://github.com/sunpy/sunpy/pulls?q=is%3Aopen+is%3Apr+milestone%3A2.1)
- Test behaviour discussion
- Should the tests fail if all the deps aren't installed? (https://github.com/sunpy/sunpy/issues/3701)
- Should the tests fail if servers are offline?
- Should the tests fail if pytest plugins are not installed? (https://github.com/sunpy/sunpy/issues/4612)
- 2.1 release triage
- PR Discussion
- Issue discussion
- Emphasize that these duties are not constrained to the core code base
- This role is largely focused on the project as a whole--make sure to emphasize this in the proposal
- Daily issue triage
- Back catalogue
- As they come in
- Mailing list/Discourse maintenance
- Dev-ops automation infrastructure
- Package template
- CI monitoring
- Make this less human in the loop!
- Development work
- Adopting stale PRs
- Less development in the first year, more in later years
- Big feature requests from the community!
- Offload science features to community and review
- Non-science features that are important to the package and useful, but less likely to be implemented by the community
- Fill open roles
- Organize coordination meeting
- Affiliated package
- Review / Re-review procedures
- Webpage
- Recruitment -- interfacing with instrument teams (instrument teams should front some cash)
- Updates for breaking changes in sunpy/other dependencies (e.g. Fido in radiospectra)
- Interfacing with upstream changes
- Need to react quickly to these changes! These are the most time-sensitive
- How do we monitor these changes?
- Data providers
- e.g. GOES XRS
- Packages
- CI
- Mentoring
- Filling open roles
- Educating community
- Onboarding contributors
-
Large future project design Scope out and describe on issues project plans for large features i.e. metadata rewrite. To facilitate contributions or funding from other places etc. Also have discussions about these things before the dev work happens.
-
Day-to-day Duties
- Issue triage
- PR review
- CI monitoring
- Mailing List / Chat support
- Helping with releases / doing bugfixes
-
Technical infrastructure
- Development of Project Automation
- Package template & auto PR bot
- CI dashboard
- Release reminder bot
- Completely autonomus bugfixes
- Document the shit out of all of this for benefit of wider community
- CI maintenance
- Sponsored package template / updates
- Development of Project Automation
-
Project infrastructure
- Milestone policies
- Drafting SEPs
- Organize weekly meetings
- Organize coordination meetings
-
Interfacing with and reacting to upstream changes
- Need to react quickly to these changes! These are the most time-sensitive
- How do we monitor these changes?
- Data providers
- e.g. GOES XRS
- Packages
- CI
-
Community Interaction
- What features do we need in the project?
- Fill open roles
- Affiliated package
- Review / Re-review procedures
- Webpage
- Recruitment -- interfacing with instrument teams (instrument teams should front some cash)
- Updates for breaking changes in sunpy/other dependencies (e.g. Fido in radiospectra)
- Milestones
- 2.1 release triage
- PR Discussion
- Remove self_test() references from docs https://github.com/sunpy/sunpy/pull/4586
- Issue discussion
- Policy on PR milestones
- Milestones should only go on PRs that block a release
- Backport PRs should be milestoned to the bugfix release
- Manual backport PRs should include only one backport per PR and should be milestoned in the same way as automatic backports and should reference the original PR
- Where do we record this new policy?
- PR Discussion
- Issue discussion
- Small dev grant
- Discussion of LASCO headers--fix is fine given the inconsistency between the metadata and image
- NDCube 2.0
- Done when it's done
- We should wait on 2.0 before starting discussion about how to use it as the base data structure for
Map
- Coordination meeting
- Need to discuss revamping of
Map
- Probably best to wait until January/February 2021
- Need to discuss revamping of
- The sponsored package dashboard is cool
- We have a lot of sponsored packages, maintaining them is hard
- e.g.
radiospectra
anddrms
need retemplating - Every time the package template gets updated, it would be nice to be able to propagate these changes downstream
- e.g.
- Laura interested in doing more work on
radiospectra
, possibly with help of students; needs reviving!
- PR Discussion
- Issue discussion
- map.rotate PR addressed https://github.com/sunpy/sunpy/pull/4502/files
- We need to properly test how different package affine transform and their effects on the data.
- We should test the 'ground-truth' of a map.rotate with reproject (flux conservation) with other affine transformations on a SDO image. This is worth writing a paper upon with the results - as a 'standard'/best practices that can be referenced.
- Mike has lots of work done on different methods on image processing - hasn't published (yet) but definitely worth talking about more.
- See this PR on aiapy regarding the agreemement between SSW and aiapy
- Project repo (https://github.com/sunpy/sunpy-project/issues)
- sunkit-instruments
- outreachy
- Coordination meeting
- PR Discussion
- Issue discussion
- outreachy
- Project ideas: https://demo.codimd.org/n5vdYiSnSAC33jNXmUtvdQ#
- ndcube
- Experiment with extra coords on real world data
- Path towards using NDCube to replace Map?
- Needs discussion at coorindation meetin
- PR Discussion
- https://github.com/sunpy/sunpy/pull/4432 - Ignore out-of-range years for an SRS query
- https://github.com/sunpy/sunpy/pull/4476 - Force metadata to be a MetaDict
- Issue discussion
- https://github.com/sunpy/sunpy/issues/4478 - OpenCV as an alternate affine transform library
- Transferring
instr
issues fromsunpy
core tosunkit-instruments
-
CompositeMap
woes (issue #2745 et al.) -
aiapy
JOSS paper--feedback welcome! - NDCube 2.0 NDCube strikes back (https://demo.codimd.org/54dm5gPMQmmJux6fV-3AOQ)
- PR Discussion
-
https://github.com/sunpy/sunpy/pull/4432 - Ignore out-of-range years for an SRS query
- Merged, might backport
-
https://github.com/sunpy/sunpy/pull/4476 - Force metadata to be a MetaDict
- Decided that it's probably not possible to trigger the fallback to basic wcs?
-
https://github.com/sunpy/sunpy/pull/4394 - Functionality to client to download new GOES 16/17 data and reprocessed 13/14/15
- Discussion of which data to default to (new or old data)
- Discussion of whether to force the user to specify a version
- Decided to add provider option to switch between old/new data, and default to new data
-
https://github.com/sunpy/sunpy/pull/4432 - Ignore out-of-range years for an SRS query
- Issue discussion
-
https://github.com/sunpy/sunpy/issues/4478 - OpenCV as an alternate affine transform library
- Lots of different affine transform implementations will be out there
- First bit of work is defining an
AffineTransform
base class that provides an interface for custom transform implementations
-
https://github.com/sunpy/sunpy/issues/4478 - OpenCV as an alternate affine transform library
- Discussion of affine transform/rotate
- Result of rotate() depends on whether you have scikit-image installed or not
- This can result in different results between the two implementations
- General reluctance to change/add stuff without first understanding the current behaviour
- Decided that the best way forward is to allow users to had their own functions into affine_transform.
- Transferring
instr
issues fromsunpy
core tosunkit-instruments
- Way forward with this is to
- Do a first release of sunkit-instruments
- Deprecate code in sunpy core
- Transfer issues form core to sunkit-
- Way forward with this is to
-
CompositeMap
woes (issue #2745 et al.)- Need to build
reproject
intoCompositeMap
- People often assume that this will be done automatically
- Need to build
-
aiapy
JOSS paper--feedback welcome!- Being written right now!
- PR Discussion
- https://github.com/sunpy/sunpy/pull/4451 - Use bunit in sunpy.map
- https://github.com/sunpy/sunpy/pull/4455 - Improve title of differential rotation example
- Affiliated package re-reviews https://github.com/sunpy/sunpy.org/issues/242
- Affiliated package for various solar models (Will)
- Performance versus elegance, i.e. how do we deal with the question "why is sunpy so slow???" (Will)
- PR Discussion
-
https://github.com/sunpy/sunpy/pull/4451 - Use bunit in sunpy.map
- Comments added
-
https://github.com/sunpy/sunpy/pull/4455 - Improve title of differential rotation example
- Merged
-
https://github.com/sunpy/sunpy/pull/4451 - Use bunit in sunpy.map
- Affiliated package re-review
- Concerns about how much extra effort this will place on reviewers, editors for affiliated packages
- We should probably have a re-evaluation PR/issue template for affiliated package maintainers to request a rereview
- We should explicitly note which version (e.g. via the commit hash or version) which will make it easier to diff the current versus review versions. See this issue
- Affiliate package for solar models
- Similar goal to the datasets repo
- Would bring together "canonical" models in solar physics, e.g. interior temperature, density as a function of radius, VAL/VAL-C atmospheres
- Will's use cases: analytical loop models like RTV, Martens, Serio, isothermal
- Others? Scope for now is any kind model that can be reference from literature
- Ideally, this should all be analytical/empirical and not complex, computationally intensive codes (i.e. should not include field extrapolation codes, MHD, even 1D loop models)
- Action item: Will to create a gist outlining a possible vision for this package
- How do we deal with "sunpy is slow"?
- There is often feedback about sunpy being "slow," particular in the context of
sunpy.map.Map
- sunpy is very general. While we do not ignore performance considerations, we do not optimize for them.
- There is a fear that as Python use in solar physics increases, users may drop sunpy or not use it to it's full potential because of actual or perceived performance bottlenecks
- We want to encourage users to report (or provide code contributions!) if they find that performance can be increased at different places or if some part of the package hampers their research workflows.
- See this issue on the aiapy repo where Stuart exhaustingly benchmarked an aiaprep workflow
- Opened https://github.com/sunpy/sunpy/issues/4459 for future discussion.
- There is often feedback about sunpy being "slow," particular in the context of
- PR Discussion
- Ready for review:
- https://github.com/sunpy/sunpy/pull/4433, Move colormap data to .csv files
- https://github.com/sunpy/sunpy/pull/4432, Ignore out-of-range years for an SRS query
- https://github.com/sunpy/sunpy/pull/4418, Add some missing bits to the API docs
- Ready for review:
- 2.1 Release
- Other Topics
- PR Discussion
- 4433 is ready to go with a rebase if tests pass
- 4432 should wait for the GenericClient refactor PR
- 4418 waiting on a doc build to pass
- 3738 is not generic enough, and will need so much refactoring that it will be worth starting again, so decided to close.
- 2.1 feature freeze is the 23rd October
- GSOC
- Fido
- Glue Solar
- Sunspotter
- PR Discussion
- Other Topics
- SPD Chat
- GSOC
- Glue Solar
- Docs build
- tox issues
- Dev guide
- Sunspotter
- Fido
- Gallery example combining HELIO, HEK, and JSOC
- Old IRIS JSOC issue
- Integration of scraper/dataretriever with redesign of GenericClient #4321
- Glue Solar
- PR Discussion
- GOES PR
- waiting on GenericClient updates
- we still aren't quite sure how to support the old FITS data for 13-15
- maybe an extra
attr
? - But how best to handle data from multiple servers...
- JSOC url-tar PR
- Can this be used with
Fido
? - Should this be an
attr
rather than a keyword argument? - Discussion about using
JSOCClient
directly rather than throughFido
- What do
request_data
andget_request
do thatsearch
andfetch
don't?
- Can this be used with
- GOES PR
- GSOC
- PR Discussion
- GSOC
- Fido
- Ideally we shouldn't have separate
Time
attrs for HEK and non-HEK queries. But then how to discriminate between HEK and non-HEK queries? - Documenting the return type of all the HEK columns; is this a worthwhile exercise?
- Ideally we shouldn't have separate
- Sunspotter
- Glue
- Fido
- GOES XRS Discussion
- We seem to ...
- GSOC
- PR Discussion
- GSOC
- Fido
- Combined metadata and data product searches now working, e.g. HEC and XRS queries
- Helio queries require table name. If not specified user is prompted for it.
- Sunspotter
- Glue
- WCS autolinking PR
- Preferred colormap PR
- 1D Spectrum with WCS PR
- Fido
- PR Discussion
- Differential rotation test and HMI synoptic map PRs pretty much ready to be merged
- Put off ERFA discussion until next week (or the next time that interested parties are present)
- Nabil's PR for removing all deprecated 1.1 code is pretty much ready to be merged as well.
- We need a 2.0.2 release soon, mostly for the NOAA fix
- GSOC
- PR Discussion
-
Updates on GSoC projects:
- Fido
- Sunspotter (Rahuul is unwell, so he gave an update on the channel) 1. Further progress on the Search Events Object. We figured out where to get the NOAA number matched with SunSpotter data. There was a problem with the units that was mostly resolved. The SEO as was discussed should be done mostly by the end of the week. 2. I will be preparing a presentation for DavidPS on the different architectures of Neural Nets. Since we're interested in time series analysis, I'll include RNNs, LSTMs, etc as well. 3. The basic pipeline is done for the CNN in pytorch, so it'll be plug an play mostly from now on, for experiments.
- Glue
- Updates to 1D profile viewer--now working with time slider 🎉
-
Change in NOAA sunspot indices files (PR #4340)
- Old text files on FTP server no longer exist
- Now in JSON format so we support both
- A lot of changes for a bugfix release--how to best document this?
- Big changelog? In the docs?
-
PRs need reviewing--we are up to 40+
-
Stuart will not be present next week; Will or Nabil will run the meeting
-
SciPy is this week
- GSOC
- CI changes
- Figure tests
- RTD PR builds
- Post 2.0 changes.
- Pull Requests
-
Updates on GSoC projects:
- Fido
- Generalized the way we get URLs from DR Clients from query attrs
- Concatenation of query results
- Question regarding duplicate GOES files--just different formatting or actually different results
- Sunspotter
- Glue
- Completed a working version of the upstreaming of the WCSAxes enabled
glue
1D Profile Viewer (where PR needs debugging and tidying up) - Introduced a new icon for the pixel extraction tool originally in
glue-solar
(to be upstreamed) - Added a non-collapsing
slice
function option for the 1D Profile viewer (in additional to the standard statistics operations)
- Completed a working version of the upstreaming of the WCSAxes enabled
- Fido
-
RTD now builds on pull requests!
- Do we want to continue to have CircleCI and RTD builds for each PR?
- Move docs build off of CircleCI?
- Figure tests should stay on Circle
- We don't want to build the docs in 3 places on every PR
- Parallelize the doc build?
- Running docs build on dev
- Proposal
- Build docs + gallery on RTD
- "No-gallery" doc build on Azure (no dev dependencies)
-
Figure tests
- Previously, running figure tests under conda
- Matplotlib wheels include version of freetype
- No freetype dependency so we can run figure tests in tox!
- Can easily run figure tests locally now 🎉
-
Mysterious macOS conda-forge failures still mysterious. See PR #4293
- GSOC
- What is blocking the 2.0 release?
- Updates on GSoC projects: Fido, Sunspotter, Glue
- 2.0 Release
- Finish the "What's New"
- Failures on the "devdeps" test point to problem with bleeding edge astropy and behavior of how
astropy.time.Time
behaves under numpy operations; possibly related to astropy/astropy#10337
- Pull Requests
- GSOC
- 2.0 blockers
- #4055 ready for review
- #4238 table matcher, in progress with euclidean distance and cosine similarity.
- Requested to have it also as example in the gallery.
- Some people not convinced this should live on core
- click and see a spectrum profile of that pixel
- Discussion of potential Fido clients: https://docs.google.com/document/d/1KYpqebFisjMRYJEoYKBJxqwIMW_jO2gaUsNVBCu_B7Y/edit#heading=h.ybxw3flop3dy
- sunpy 2.0 outstanding PRs: https://github.com/sunpy/sunpy/milestone/41
- 2.0 Release
- How should we deal with the
instr
subpackage in terms of the newsunkit-instruments
repo?
The collaboratively editable version of the minutes and agenda for the upcoming meeting can be found here. These will be archived here on the wiki following the meeting.
No community meeting was held due to the Python in Heliophysics Spring Meeting.
We need to review all PRs https://github.com/sunpy/sunpy/milestone/41 to get ready for an RC this week.
See https://docs.google.com/spreadsheets/d/1gsEQ3_xxHw4_gPnkfEna2IPqN9e7Cdom8KM_o6TVaUE/edit#gid=0 for the official list
- Magnetogram sources/clients/examples (led by dstansby)
- Meeting streamed onto YouTube for posterity
- 2.0 PRs
- Rectangle - leave for Albert to review
- Attr registration - waiting on changes in response to review
- User config - needs another review, should change function name to
copy_default_config
- JSOC error - re-milestone to 2.1
- VSO fetch fix - re-milestone to 2.1
- Moving constants - restarted failing build, waiting for Albert to review
- Map singeldispatch - merged
- What's new - add ideas!
- General chat about differences bewtween old and new differential rotation code
- Added an HTML summary for Map and MapSequence -https://github.com/sunpy/sunpy/pull/3951
-
Add contour map method - https://github.com/sunpy/sunpy/pull/3909
- Suggestions for meaningful tests?
-
Map now raises the file that failed to load if it errors on reading - https://github.com/sunpy/sunpy/pull/3727
- David S. suggests opening a new PR/branch, as devs are now just force pushing to the original contributor's branch...
- Nabil suggests its ok.
- Merged sunpy.sun.sun Pr
- #3951 needs review
- Discussion on methods on map vs functions in maputils relating to #3909, decided to put
contour()
inmaputils
.
- Follow-up from SunPy Coordination Meeting last week
- Issue triage
- PR review
-
add .cmap and .norm property to plot_settings
- https://github.com/sunpy/sunpy/pull/3624 - we said we would make a decision on this at the dev meeting but I don't think we did
-
Implemented Groundclients in continuation with PR 2619
- https://github.com/sunpy/sunpy/pull/3763 - need to check the clients and VSO return same data products, and whether VSO can provide real time data now.
- modified search metadata to return Astropy.Table
- Add a other issue template
- Added an HTML summary for Map and MapSequence
- Fix deprecated decorator
-
add .cmap and .norm property to plot_settings
- Co-working hour
- Time(s)? Day(s)?
Present: Jack, Albert, Shane, Nabil, Stuart, Will, David S.
- Follow up from coordination meeting
- How often do we want these to happen?
- Nabil suggested having more frequent, shorter, and focused meetings, e.g. around 3.0 -- weekly?
- Should we tie times of meetings to the release schedule? (this happened, somewhat by coincidence, this time around)
- At the very least, annual coordination meetings
- Issue triage
- Boring, but needs to be done (mostly a TODO for Will)
- Be more aggressive about closing issues, they can always be reopened
-
instr
issues need to be moved over to thesunkit-instruments
repo - PRs
- HTML
__repr__
PR #3951- *nearly done
- other features should go into other PRs
- would be good to generalize to NDCube, but a lot of work
- Deprecated decorator PR #3982
- Merge soon
- Create issue/PR upstream to astropy
- break version calculation out into its own function
-
cmap
andnorm
property PR #3624- We underestimated how integrated
plot_settings
is inMap
- Replace with a
plot_settings
subclass ofdict
that pulls out the plot-specific parts of the API - Deprecate
__getitem__
on this new object and then providecmap
,norm
, etc. as properties on this object - Squash existing commits to preserve the contribution, close this PR, and open a new PR on top of this branch (TODO for Nabil)
- We underestimated how integrated
- HTML
- Coworking hours
- 9:00 UTC and 17:00 UTC
- Tuesday (but reassess to see if this works for everyone)