MetPy Dev Call - 2021

MetPy Dev Call ([#1684](https://github.com/Unidata/MetPy/discussions/1684)) - 2021 === ## Dec 17, 2021 :::info - **Time:** 1:00 PM MST - **Participants:** - Eric Bruning - Russell Manser - Ryan May ::: ### :lightning_cloud: State of MetPy * Work towards 1.2 milestone. `moist_lapse` fixing is a pain. * BAMS-uscript in good shape, sent out to co-authors * Future breakages by dependencies (e.g. Matplotlib 3.5) will prompt immediate bugfix release rather than waiting for any feature release. ## Dec 3, 2021 :::info - **Time:** 3:00 PM MST - **Participants:** - Eric Bruning - Drew Camron - Kevin Goebbert - Russell Manser - Ryan May - Jon Thielen ::: ### :lightning_cloud: State of MetPy Milestone and BAMS-uscript progress. Highlights include `moist_lapse` improvements/fixes, updated implementation, support responses, and forward movement for spherical calculations. BAMS-uscript will go to coauthors soon. Goals for end-of-year -> AMS for training material deployment, with student conference workshop planning. ### :loudspeaker: Discussion, questions, and updates - Questions about Unidata access to or relationship with NWS repositories of information/calculations e.g. WBGT - [#2235](https://github.com/Unidata/MetPy/issues/2235) discussion on docs/implementation update on `kappa_star` terminology -> filtering and interpolation discussion on existing implementation and potential new multi-pass options - doc update will go in, with warning for users who may be exploring common kappa values and not know its non-standard value relative to literature - **updated timing**:: as of Dec 17 2021, transition to **every other Friday, 1 pm MST** ## Nov 19, 2021 :::info - **Time:** 3:00 PM MST - **Participants:** - Eric Bruning - Drew Camron - Kevin Goebbert - Russell Manser - Ryan May - Jon Thielen ::: ### :lightning_cloud: State of MetPy BAMS manuscript deadline requested for extension through the end of the year. MetPy and **Siphon** infrastructure clean-up with help from Daryl Herzmann! Lots of build effort around conda-forge and towards Python 3.10. We're testing on 3.10, and environments can nearly smoothly be assembled via conda. ### :loudspeaker: Discussion, questions, and updates - [1.2.0 milestone](https://github.com/Unidata/MetPy/milestone/29) discussion - Kevin shared a demonstration notebook [gist](https://gist.github.com/kgoebber/b3546977e55fef96d4b36f4ef573a788) for spherical calulcation discussion [#893](https://github.com/Unidata/MetPy/issues/893) - limitations and considerations on potential vector derivatives (not many use-cases?) - moderate potential for future xarray data model that may accomodate vector information/operations (Jon mentions [datatree](https://github.com/TomNicholas/datatree)) - path to creating vector gradient functionality with API pathway - setting `mx`/`my` to 1 here should reproduce cartesian functionality - API pathway decisions: - have kinematic calculations prefer this new pathway 1. check xarray inputs for CRS information **and** lat/lons: if yes, sally forth 1. if CRS **and no** lat/lons: calculate lat/lons, then use this path; **warn** about performance and background calculations 1. for non-xarray, require lat/lons input, perform dx/dy calculations if not provided - what if map factors included? future examination on demonstration of use-case and evaluation of e.g. WRF's map factors - limitations of `first_derivative`-based functionality on masked arrays; causing failures in Russell's testing infrastructure - solar angle and time discussion [#1440](https://github.com/Unidata/MetPy/issues/1440) - how much authority and precision do we stake here? - Kevin has a previous implementation of textbook equations here, will share ## Nov 5, 2021 :::info - **Time:** 3:00 PM MDT - **Participants:** - Eric Bruning - Drew Camron - Kevin Goebbert - Ryan May - Jon Thielen ::: ### :lightning_cloud: State of MetPy GitHub Actions refactor [[#2183](https://github.com/Unidata/MetPy/pull/2183)] going in, unclogging lincheck PR failures et al. New issue submission forms. Grant report will be upon us before we know! ### :loudspeaker: Discussion, questions, and updates - 2155/2156 names? - currently `field_smooth`, `field_zoom`; proposed `field_smooth`, `contour_smooth`? verb-first? - **DO:** verb-first: `smooth_field`, `smooth_countour` with helpful docstrings and a contrasting example - spherical calculations - between "solved" and implementation - will include multi-model vorticity example - needs support to move forward ## Oct 8, 2021 :::info - **Time:** 3:00 PM MDT - **Participants:** - Eric Bruning - Drew Camron - Kevin Goebbert - Russel Manswer - Ryan May - Jon Thielen ::: ### :lightning_cloud: State of MetPy Unidata UserComm status reports up to be discussed at next session. CSSI award features EarthCube funds and includes a request to attend EarthCube "new projects welcome". BAMS manuscript write-a-thon still on for Monday. ### :loudspeaker: Discussion, questions, and updates - [1.2.0](https://github.com/Unidata/MetPy/milestone/29) Milestone has "officially" passed; when to aim for? - "Officially" reschedule for a late November wrap-up, leave some - [PR1963](https://github.com/Unidata/MetPy/pull/1963) from Kevin still a priority to resolve regarding spherical corrections; to be "scrapped" in favor of new developments across a variety of grids - [IS2078](https://github.com/Unidata/MetPy/issues/2078) can implement a fix to restore previous behavior (e.g. Connor's proposed solution) for 1.2 - BUT, discussion: feasibility of preserving independent projection info across multiple fields and not defaulting to first projection in collection. 1.2? - Multiple issues around citations, authorship, infrastructure - Quick discussion on "contributor retention" - suggestion to explicitly invite PR contributors to dev call ## Sep 24, 2021 :::info - **Time:** 3:00 PM MDT - **Participants:** - Eric Bruning - Drew Camron - Kevin Goebbert - Russell Manser - Ryan May - Jon Thielen ::: ### :lightning_cloud: State of MetPy MetPy beginner virtual synchronous workshop was successful last week! Very active period for MetPy issues ### :loudspeaker: Discussion, questions, and updates - discussion on supersampling issue [[#2126](https://github.com/Unidata/MetPy/issues/2126)] - create example? - suggest upstream request to e.g. xrscipy - discussion on broadcasting issue [[#2123](https://github.com/Unidata/MetPy/issues/2123)]: 1-D calc output across a map (e.g. zonal mean jet) - create example? - Cartopy 0.20 released with PROJ8 support - matplotlib 3.5 on the horizon with breakage - flake8 meets gempak reader meets refactor - Eric's update on WBGT: has original C source code on NDFD WBGT implementation with supporting documents and information - planning for BAMS manuscript - content outline per section laid out by end of next week - Mon 11 Oct evening write-a-thon scheduled - Jon attended pydata "duck array discussion" [[repo](https://github.com/pydata/duck-array-discussion)] ## Sep 10, 2021 :::info - **Time:** 3:00 PM MDT - **Participants:** - Eric Bruning - Drew Camron - Russell Manser - Ryan May - Jon Thielen ::: ### :lightning_cloud: State of MetPy Solver has "debuted"! Cartopy NaturalEarth linking issues have trickled down into our CI. ### :loudspeaker: Discussion, questions, and updates - Relative Humidity issue vs ERA-5, discussion ([#2088](https://github.com/Unidata/MetPy/discussions/2088)) - RH-liquid vs -ice - [Eric's lecture notes](http://www.atmo.ttu.edu/bruning/5321/fall20/notes/CP1008.pdf) - [ECMWF operational satvap](https://www.ecmwf.int/sites/default/files/elibrary/2016/17117-part-iv-physical-processes.pdf#subsection.7.4.2) - Performance refactor ([#2064](https://github.com/Unidata/MetPy/pull/2064)) approach and go-ahead - [decorator-based gist](https://gist.github.com/jthielen/93ec9fcd8f3892a83215309fe5cda2c6) vs implementation in PR - Discussion on ballooning decorator approach for various function input/output metadata - **Decision on this being okay and digestible for now** - long term considerations for collapsing these approaches as appropriate - Implementation suggestions discussed including shift away from namedtuple and constant function input magic - Discussion on ([#2090](https://github.com/Unidata/MetPy/pull/2090)) brings back around discussions on future optimizations and "the great numba divide" ## July 8, 2021 :::info - **Time:** 3:30 PM MDT - **Participants:** - Eric Bruning - Lydia Bunting - Drew Camron - Connor Cozad - Steven Decker - Kevin Goebbert - Russell Manser - Ryan May - Jon Thielen ::: ### :lightning_cloud: State of MetPy Annual [MetPy Survey](https://forms.gle/jWq176wj9eMUBdk17) is available until July 31. ### :loudspeaker: Discussion, questions, and updates - Today's projection discussion ([#893](https://github.com/Unidata/MetPy/issues/893)): - [Eric's continuation](https://gist.github.com/deeplycloudy/7b8412094c7a249d19b771374f312750) of Kevin's initial exploration - Expand exploration to data on projected coordinates (NAM, HRRR, WRF) - Current proposed solution (more expensive, easier to implement) - implement lat-lon solution for corrections - update signature (to assume radius?), accept crs from pyproj and pull out radius - update signature to add latitude - calculate 2-d lat-lon grid of projected grid cells - consideration: earth- vs grid-relative u, v ## Jun 24, 2021 :::info - **Time:** 3:30 PM MDT - **Participants:** - Eric Bruning - Drew Camron - Kevin Goebbert - Russell Manser - Ryan May ::: ### :lightning_cloud: State of MetPy Progress on PR reviews and community engagement towards 1.1 release. GEMPAK IO going in soon, but not likely as a full xarray backend yet. ### :loudspeaker: Discussion, questions, and updates - per comment below, Eric has produced the beginning of a grid-based map factor example [here](https://github.com/deeplycloudy/demos/blob/main/3D-grid-spacing.ipynb) - prepare upcoming survey for status report, to try and get out Monday - Trim questions about other python packages and community support - Expand on question about missing functionality/capability/speed - Introduce questions to help gauge equity of our tools and resources ## Jun 10, 2021 :::info - **Time:** 3:30 PM MDT - **Participants:** - Eric Bruning - Drew Camron - Kevin Goebbert - Russell Manswer - Ryan May - Jon Thielen ::: ### :lightning_cloud: State of MetPy BAMS article proposal accepted, draft expected within 6 months. Catching up on community PRs, intern contributions, and milestone tasks; 1.1 release delayed to July. ### :loudspeaker: Discussion, questions, and updates - Discussion on enabling grid-aware calculations of derivatives, vorticity - Eric will put together resources on calculating nominal map factors based on grid information - Kevin will put together resources on digging into pyproj for `Proj`-based map factor retrieval ## May 13, 2021 :::info - **Time:** 4:00 PM MDT - **Participants:** - Eric Bruning - Drew Camron - Kevin Goebbert - Russell Manswer - Ryan May - Jon Thielen ::: ### :loudspeaker: Discussion, questions, and updates - Tackling backlog of community PRs, namely Showalter (#1673) and GEMPAK IO (#1656) - Begin discussions on division of labor and plans for grant deliverables and project future - BAMS article proposal submitted ## Apr 29, 2021 :::info - **Time:** 4:00 PM MDT - **Participants:** - Drew Camron - Kevin Goebbert - Russell Manser - Ryan May - Jon Thielen ::: ### :lightning_cloud: State of MetPy NSF CSSI Proposal awarded! [1.0.1](https://github.com/Unidata/MetPy/discussions/1838) out! BAMS abstract nearly re-wrote. ### :loudspeaker: Discussion, questions, and updates - Discussed that grant award should motivate [updated roadmap](https://github.com/Unidata/MetPy/issues/1655) ## Apr 15, 2021 :::info - **Time:** 4:00 PM MDT - **Participants:** - Eric Bruning - Drew Camron - Kevin Goebbert - Russell Manser - Ryan May - Jon Thielen ::: ### :lightning_cloud: State of MetPy Codecov security breach leading to ci security cautionary updates, no malicious behavior noted. 1.0.1 nearly out-the-door including lcl/wetbulb nan-handling, Kevin's contour fixes, and _removal of a packaged viridis colortable_. ### :loudspeaker: Discussion, questions, and updates - Meetings will move to 4:00 pm MDT for the remaining meetings of the semester, calendar invite to be updated ## Apr 01, 2021 :::info - **Time:** 3:30 PM MDT - **Participants:** - Eric Bruning - Drew Camron - Kevin Goebbert - Russell Manser - Ryan May ::: ### :loudspeaker: Discussion, questions, and updates - Kevin's plans for post-semester contributions; seeking suggestions for examples - Discussion on setting "bounties" for targeteted community involvement from experts ## Mar 18, 2021 :::info - **Time:** 3:30 PM MDT - **Participants:** - Eric Bruning - Drew Camron - Kevin Goebbert - Russell Manser - Ryan May - Jon Thielen ::: ### :lightning_cloud: State of MetPy We have generally "caught up" on the internal/external support workload. Our portion of the AMS short course was taught successfully from a new notebook designed with support from Unidata's instructional designer. [Completed notebook available here](https://gist.github.com/dcamron/0c8ebdd607dd4b36bbf8f1d4e76af2cc) and will be used as a basis for repeat half-day virtual classes or expanded into a full-day class in the next few months. A few usability issues were captured as a result of that exercise. Ryan made an intentional effort to respond to every issue, PR, discussion to have appeared for MetPy for some time. Ryan has made progress on the BAMS abstract. Drew will be prioritizing the python-training materials and site for the next weeks. ### :loudspeaker: Discussion, questions, and updates - As part of surface discussion for new issues - similar (same?) `reduce_point_density` kdtree issue with large collection of sfc data - fixable formatter functionality - upper air barb offset - Kevin gives class updates - Proposal and longer-term general discussion ## Feb 18, 2021 :::info - **Time:** 3:30 pm MST - **Participants:** - Eric Bruning - Drew Camron - Kevin Goebbert - Russell Manser - Ryan May - Jon Thielen ::: ### :lightning_cloud: State of MetPy Discussion on future integrated access to [Synoptic Data PBC](https://synopticdata.com/) API for near-real-time data, an offer from their team - Boundaries and limitations? - MetPy core can't guarantee support, won't advertise explicitly, minimal example documentation - MetPy core won't guarantee maintenance unless this lives to demonstrably serve the community - Precedent this sets? - Provided data must be freely re-usable - MetPy core can always modify and clarify stance on this in the future ### :loudspeaker: Discussion, questions, and updates - [intake-thredds](https://github.com/NCAR/intake-thredds) release yesterday! Future of siphon as it relates to this? - Declarative discussion - [Future functionality?](https://github.com/Unidata/MetPy/discussions/1717) - [`copy` method](https://github.com/Unidata/MetPy/issues/1719) - [Complex layout support](https://github.com/Unidata/MetPy/issues/1707) ## Feb 04, 2021 :::info - **Time:** 3:30 pm MST - **Participants:** - Eric Bruning - Drew Camron - Kevin Goebbert - Russell Manser - Ryan May - Jon Thielen ::: ### :loudspeaker: Discussion, questions, and updates - [MetPy and Siphon (#1678)](https://github.com/Unidata/MetPy/discussions/1678) - Jon discusses [intake](https://github.com/intake/intake)! - a standardized API connecting your data container (numpy, pandas, xarray, etc.) intake's drivers to desired remote data via yaml/"catalog" interface - NCAR [intake-thredds](https://github.com/NCAR/intake-thredds) does exist using siphon to map to TDSCatalog -> opendap - our adaptation might create a new, more robust intake driver interfacing with TDS, NCSS - **MetPy to move to absorbing non-thredds, met-specific data access/read capabilities from siphon** - **siphon thredds capabilities to look to towards intake** ## Jan 21, 2021 :::info - **Time:** 3:30 pm MST - **Participants:** - Eric Bruning - Drew Camron - Kevin Goebbert - Ryan May - Jon Thielen ::: ### :loudspeaker: Discussion, questions, and updates - Short discussions on AMS workshop, potential workshop opportunities in the future - Takeaway to plan to cancel "AMS intermediate" short course - Will be participating in "AMS intro" short course with others - Kevin has a potential opportunity in discussion - Unidata will begin constructing its own standalone workshop - Discussions on 1.0.1 milestone, things to wrap up - Posted milestone issues/PRs to be cleaned up and reviewed - Declarative issue with united mathematics losing units, unquantified -> quantified - Larger issue regarding cyclic boundaries and plotting (came up with declarative contour eSupport issue) - direction to take regarding interpretation of boundaries and fix - add duplicate entry to data a la `cartopy.util.add_cyclic_point` or geocat utils - Organizing `declarative.py` interface differently? Potentially breaking up into smaller files while maintaining module/import structure - Release cycle discussion? Timing/frequency - Faculty discussion on student computing resources, local hardware for jupyter hubs, cloud resources ## Jan 07, 2021 :::info - **Time:** 3:30 pm MST - **Participants:** - Eric Bruning - Drew Camron - Kevin Goebbert - Ryan May - Jon Thielen ::: ### :loudspeaker: Discussion, questions, and updates - Discussion on [#1635](https://github.com/Unidata/MetPy/issues/1635)/[#1639](https://github.com/Unidata/MetPy/pull/1639) regarding pressure weighting behavior with Temperature -> 1.0.1 - Falls back to unit conversion/reassignment behavior down to `get_layer()` - Will triage with Kevin's update+fix for now - Discussion on [#1636](https://github.com/Unidata/MetPy/issues/1636) regarding expected behavior with CAPE/CIN shading, particularly decision not to shade below LCL - Handle `.parse_cf()` fallback assumption to latlon when given no crs, given dimensional coord x, y, and given non-dimensional 2-d coord lat, lon ([stackoverflow](https://stackoverflow.com/questions/65567078/question-about-the-coordinates-of-metpy-cross-section)) -> [#1648](https://github.com/Unidata/MetPy/issues/1648) -> 1.0.1 - Addition to what could go wrong?