Release 6.5.0 #1845

Post release 6.4.3 reset

* - cache mpi headers and bodies * - final step in implementing caching: only cache message bodies (skip routing of headers and maps when possible) * - fix: initialize mpi request handles * - add MPI error check * - PETSc vec readonly access * - add more strategic MPI error checks

* chore(timer.f90): cleanup docstrings in Timer utility * Restoring uppercase of intrinsic

* docs(mf6io): correct description of CDIST in gwfgwf/gwtgwt * correct deprecation warnings to use proper version number

* integrate sim-tdis with idm * rebuild makefiles * fix comment * rebuild makefiles --------- Co-authored-by: mjreno <mjreno@IGSAAA071L01144.gs.doi.net>

* feat(GWE): Introduce Groundwater Energy Transport * fix typo in meson file * fix a formating issue that was popping up in an MST autotest * some legacy line leftover from a botched rebase, possibly * Need update initial autotest to conform to new autotest standards adopted with #1464 * Start looking for post-rebase breakages * Get gwfgwe and gwegwe exchanges up-to-date based on #1505 * Code that had been moved to set_active_status in FMI was still present in fmi_fc(). Removing. * Forgot to remove unused variables after making changes in 4d76729 * Adding another autotest specific to GWE * Adding another autotest after getting it #1464 compliant * Compliantizing another new autotest with PR #1464 * Fixes in response to #1493 (comment) * Fix in response to #1493 (comment) * Rerunning black in response to #1493 (comment) * Fix in response to #1493 (comment) * Made changes to dfn files in response to #1493 (comment) and #1493 (comment) and reran mf6ivar.py, which updated the tex files in this commit * Fix in response to #1493 (comment) * remove unnecessary line of script * Add missing lines to ConnectionBuilder.f90 related to GWE * Add a GWE vs GWT comparison autotest * GWE-GWE exchanges now working. Other clean-up for code uniformity * Rebrand constant temperature package acronym to CTP * Missed an import renaming update * Forgot to remove a now obsolete file due to renaming. * Rebrand energy storage and transfer package acronym to EST * Forgot to remove a now obsolete file due to renaming (again) * Rebrand gwe dispersion package acronym to CND (conduction) since that is the dominant process in heat transport * Remove gwe-related code from generalized transport code (tsp1.f90 & tsp1ssm1.f90) * add single-cell test for energy source loading (ESL) package * Adding energy source loading (ESL) package * remove unused variable * fprettify * Add another ESL autotest * Adding another autotest that compares gwe to three different analytical solutions from Carslaw & Jaeger (1947) * Bringing over Stallman autotest from previous GWE PR (#1237) * Adding streamflow energy transport (SFE) package * Add autotest for SFE * Adding lake energy transport (LKE) package. Includes new autotest * forgot meson update * Adding multi-aquifer well energy transport (MWE) package. Includes new autotest * Adding unsaturated-zone energy transport (UZE) package. Includes 2 new autotests * Attempting to reapply a failing autotest. Unable to discern why it is failing as downloaded contents from its failure are working locally. * removing troublesome autotest. Downloaded contents from failed run on Actions doesn't yield the same output that the logs are reporting. This particular autotest has an analytical solution (and a plot) that might be better shown on the modflow6-examples repo anyway * Removing a file that shouldn't have been added (snuck in among other staged files) * Update release notes

…1611) * suggest Chocolatey and winlibs distributions * mention conda distribution's incompatibility

* if CI tests fail it can be helpful to inspect their outputs * add workflow_dispatch trigger to `large.yml` * miscellaneous cleanup

add python script that can poll memory usage on HPC system

* v15 of exes distribution doesn't have double-precision program versions * temporary workaround until double-prec mf2005/mfnwt/etc restored to dist

* previously triggered on push/PR to master or release branch

* reset PR step was removed from release procedure, remove mention of it

Post release 6.4.4 reset

* correct inaccuracies from earlier versions of release automation * remove water mover performance entry from develop.tex * rename vx.x.x-template.tex -> vx.y.z-template.tex * describe release notes maintenance procedure * remove unused function from build_dist.py * reformat python scripts in distribution/

* skip, don't crash/fail if downloaded programs or rebuilt regression mf6 programs not found * only require dev build to gracefully run autotests

* move logic into external model repo tests where used * simplifies framework state management

* compare_output and run_sim_or_model

* double precision mf2005, mfnwt, mflgr and mfusg are back in exes dist v16

* test_mf6_tmp_simulations.py was unused/skipped, remove it * test_examples.py still had a reference to TestFramework.setup(), neglected in 886e394

* Introduce SWF Model

* refactor(swf-disl): remove unused routines * fprettify * some minor reformatting

* add macos-14 (m1 runner) build to release matrix * native support for apple silicon, removes need for rosetta * use devtools os tags to distinguish intel from arm mac artifacts * remove unused inputs from release_dispatch.yml: commit_version, reset

* refactor: remove filename numbers from ims and tsp * remove filename numbers from gwf * remove filename numbers from gwt * sim, model and exchange file names too * fix msvs proj file * gwe updates * swf updates * update makefile * update mf6memvar * run mf6ivar * restore gwfswf -> swfgwf naming * rebuild makefiles with newer pymake

* refactor: cleanup swf * minor tweaks

* another step toward build-time model selection * introduce dfns.txt specifying paired IDM <-> DFN files * paired IDM files now named <dfn>idm.f90, no need for explicit mapping * introduce options --dfn, --outdir, --verbose for dfn2f90.py * add pyaml to environment.yml (dfns.yml also supported) * colocate generated IDM files in src/Idm/ subdirectory * update utils/idmloader/IDM.md per above

* update DEVELOPER.md so that meson setup command can be copied separately for Linux/MacOS and Windows

* add description of the bug fixed by PR #1633 * extend test_gwf_sfr_inactive autotest to reactivate sfr reaches after inactivating 3 of 6 reaches * update test_gwf_sfr_inactive autotest test description

* add check of cond and auxmult values in DRN, RIV, and GHB packages * add test that DRN, RIV, and GHB packages issue error message if negative cond and auxmult values found * update release notes * update mf6io guide for DRN, RIV, and GHB with information on minimum COND and AUXMULT values and that the program will terminate with an error if negative values are specified. closes #1639

The main purpose of compilers.yml is to check which compiler toolchain/version work on which OS/version. The motivation for the scheduled runs was that github runner images might change underfoot, but this does not impact local development and is only relevant if we want to change how releases are built in CI in future (e.g. adding a gfortran build, using a newer ifort or ifx). So we can just run compiler checks manually whenever needed

* second try at ARM mac dist, replacing 95aec27 * add statically linked distribution built with gfortran * link static libgfortran and libquadmath by hiding dylibs * -static* flags don't seem to work on macOS * set LDFLAGS for macos >= 13

* fix 'ZDRN + DDN' -> 'ZDRN + DDRN' * same fix in DRN-Discharge-Scaling.ipynb

At this time PRT supports particle tracking on all DIS and some but not all DISV grids. DISV cells containing adjacent parallel faces are not properly handled and may cause crashes, this includes quad-refined grids as well as DISV grids generated by FloPy's voronoi module. A GWF-PRT exchange is included, which can be used to couple GWF and PRT models in the same simulation. PRT can also consume flows from a previously run GWF model via FMI. Support is planned but not yet implemented for: * configurable particle mass * local mass conservation — conserved between cells but not yet guaranteed at the subcell level * exchanging particles between models * DISU grids (longer term) * solving models in parallel (longer term) Docs/dfns and tests make up about half the changeset, the rest is fortran source. Input/output format is likely to change without notice until PRT reaches a release-ready state. PRT's primary author is Alden Provost. Some code is adapted from MODPATH and originally due to David Pollock. Co-authored-by: Alden Provost <aprovost@usgs.gov>

* in cases with multiple outlets in a lake the outlet observation used the incorrect index when evaluating the package ibound * added test that compares outlet observations to known values

* refactor(misc): improve doc * improve auto flow reduce description in dfn to indicate it is only for unconfined cells * mark swf as a dev_feature * minor name change to tsp list file header * fprettify

* move function for effective horizontal hydraulic conductivity out of gwf-npf.f90 * introduce HGeoUtilModule for standalone hydrogeologic functions * eliminate dependency on gwf-npf from gwt-dsp and gwe-cnd * todo: more generic routine and argument naming

#1637) * move discretization modules to src/Model/Discretization/ * rename gwf-dis.f90 -> Dis.f90, likewise for disv and disu * define "found type" in discretization modules, remove generated IDM file dependency * move toward clean separation between generic framework and model-scoped code * future work may introduce model-specific dis types extending these * for now this allows moving forward with feature toggling * rerun mem_allocate.py and mf6ivar.py

* add rules: no trailing return statements, clean up comments (no double dashes, no empty comments, internal comment spacing) * run black/isort on scripts in .github/common/

* restrict yanked pytest version

* - coupled gwf with buy requires interface model * - update makefile * - update makefile

* workaround for netcdf flopy branch not yet having PRT utils (needs rebase) * unblock mf6 netcdf work for mjreno

* does not account yet for idomain

…lel testing of GWE (#1659) * test(gwe): New autotest for GWE that will facilitate additional parallel testing of GWE * get more info on why autotests are hanging & hanging * Forgot to turn off the plotting

* base workspace was selected where sim/model-specific workspace should have been * only collect expected output files if comparison is enabled * a lot of these complications will go away with snapshot testing

…#1655) * remove DEV_MODFLOWUSG_UPSTREAM_WEIGHTED_SATURATION option * exclude test models that use mfnwt or mfusg conductance options * add staggered_thkfrac function to containerize staggered logic * move conductance functions to GwfConductanceUtilsModule * use staggered_thkfrac function in gwe-cnd and gwt-dsp * add vscode task for rebuilding makefiles * remove saturation recalculation in HCOND for newton * modify test_gwe_drycell_cnd0.py to skip evaluation of first timestep * exclude test001d_Tnewton until v 6.5.0 is released * add ENUM for horizontal conductance options * add ENUM for 3-D cell connection types

* exclude models using options removed in f8885ec * running out of disk, remove some software preinstalled on the runner image to free some space * remove test_examples.py, example models can be tested nightly in mf6 examples repo, with regression comparisons accomplished via snapshot testing in that repo — mf6 test framework makes assumptions about workspace dirs that new GWE models don't (and need not) follow

* ignore well-known python-dateutil and pandas warnings * use .item() to extract single items from numpy arrays * mark PRT voronoi tests slow (todo: refactor, speed up) * use raw docstrings if they contain backslashes

* fix test name in test_prt_drape.py * reuse functions from test_prt_voronoi1.py in ..._voronoi2.py * hardcode cell IDs in voronoi tests instead of re-intersecting every run * remove slow markers, tests now run in a few seconds

* updates in preparation for keystring input types * rebuild makefiles * fix rebase conflict * some refactor and cleanup --------- Co-authored-by: mjreno <mjreno@IGSAAA071L01144.gs.doi.net>

* test(par/gwe): second autotest for parallel gwe * clean-up of a clean-up * fix my misunderstanding * case name was too long * adding missing scenarios to cases variable * refactor test_gwe_split_analyt.py fitting better with parallel testing * Remove commented-out plotting commands

* Implement conveyance approach for diffusive wave flow calculation * update tests to work with conveyance formulation * fprettify * test-drive test for cxs utils * fprettify * add an n-point cross section test-drive autotest * fprettify * try to figure out why failing on windows * Model misbehaving on Windows

* build petsc with ifort/cl from cygwin * factor out scripts for cygwin build steps * factor out actions for mf6 parallel testing * meson test serially to avoid errors on windows

* refactor(swf): generalize dis usage to support dis/disv * preparation for moving width out of dfw package * removal of width from dfw; starting to implement dis/disv * DIS appears to be working with SWF; new test is passing * fprettify * update latex files * remove backtracking

* libpetsc.dll wasn't found if PETSc cache miss, maybe interference from cygwin setup? setup PETSc environment after building PETSc to workaround * simplify ONEAPI_ROOT usages

…1679)

* probably want to maintain compatibility until version EOL * catch syntax usages only supported on newer versions * update print format syntax in test_swf_dfw_bowl.py (no nested quotes)

* conditions to determine whether particle should continue advancing were not independent * set cell when initializing subcell methods, not currently needed but might be in future * minor cleanup in test_prt_budget.py

Add a parallel Windows distribution to the release workflow, making it available both on the nightly build and in official releases. The archive is named win64par.zip and contains all non-system dependencies. This avoids the need for software installations or other actions requiring elevation for the end user. Bundled dependencies include: - MSVC runtime * ucrtbase.dll * msvcrt.dll * msvcp_win.dll * MSVCP140.dll * VCRUNTIME140.dll * VCRUNTIME140_1.dll - oneAPI (Fortran compiler, MKL, MPI) * libifcoremd.dll * libmmd.dll * mkl_mc3.2.dll * mkl_core.2.dll * mkl_avx512.2.dll * mkl_avx2.2.dll * mkl_def.2.dll * mkl_sequential.2.dll * libfabric.dll * impi.dll * mpiexec.exe * hydra_bstrap_proxy.exe * hydra_pmi_proxy.exe * hydra_service.exe - PETSc * libpetsc.dll Miscellaneous: - factor out action to build mf6 parallel on windows - add developmode toggle to release_dispatch.yml

* previously hardcoded which broke the nightly build

* test(vcatch): add v-catchment overland flow test for SWF * adjust range for quadratic smoothing function * test fails on CI but not locally: adjusting solver tolerance

#1687) * installs can take 30+ min and not handled by setup-fortran, so need to cache manually * factor out an action for mf6 parallel oneapi install

Co-authored-by: mjreno <mjreno@IGSAAA071L01144.gs.doi.net>

Performance issue discovered while working on MP7/PRT examples. Previously cell definition arrays (face flows, face neighbors, etc) were reallocated for each subcell regardless of tracking method. This introduced overhead for each subcell solve, scaling with the total particle count. Reallocation can be skipped for structured grids since subcells are guaranteed to be 4-sided, and can happen lazily for unstructured grids. Removing the unnecessary allocations speeds up the relevant example model from a minute or more to subsecond runtime on my machine

* feat(disv1d): simplify and replace DISL with DISV1D * No more support for the TOREACH option in DISL * DISV1D is a generalized version of DISL * fprettify

…1557) * add echo file option for layered input params * some naming adjustments * use per package export option to control writing variable layer array files * restore mf6ivar.py * silently create export files * post rebase definition update * add release note for export files --------- Co-authored-by: mjreno <mreno@IGSAAA071L00066.gs.doi.net> Co-authored-by: mjreno <mjreno@IGSAAA071L01144.gs.doi.net>

* only need to cache intel-classic 2021.7 and intel 2024.0

* add export array testing to autotest/test_gwf_npf01_75x75.py * update tex files There is still an unresolved issue in that export_array only works for "layered" variables, but if we add "layered true" to disv1d variables, then the io guide says "layered" for these variables. Also, for the swf-dis2d grid, if we include support for export_array, then we get nrow files created, each with ncol values in it. So there is some work left to do in supporting disv1d, dis2d, and other variables, such as gwf-dis-top, that might be considered "layered".

* add pixi.toml environment file, equivalent to conda environment.yml * pixi.toml also defines tasks: aliases for development commands * checkin pixi lock file, add ci workflow to update it monthly * use pixi tasks in ci actions, scripts, workflows * update developer docs with pixi info

* add version to pixi.toml * include doc/Doxyfile and pixi.toml in update_version.py * consolidate tasks * refactor custom install as pypi dependencies * remove branch filters from push trigger in commit.yml * update dev docs: keep conda language, tidying * ignore pixi beta feature warning in ci * restore vscode config to avoid breaking contributor workflow — when vscode python * supports pixi I think we can update the tasks to use "command": "${command:python.interpreterPath}"

) * add intel 2024.0 and 2024.1 to compiler test matrix * remove incompatible intel versions 2021.1 - 2022.1 * remove gcc 7 - 10

* until refactored and better tested

* maximum tracking time was set incorrectly, particles were solved beyond the end of the time step for multi-step simulations — max time within the tracking loop should be end of time step unless it's the last time step in the simulation, in which case it should be particle stop time * if, on recording a particle datum, the tracking time fell exactly on the boundary between time steps, the datum was assigned to the subsequent time step and stress period, while MODPATH 7 assigns it the previous — factor out a routine in the base MethodType to save particle data, with some extra logic for an inclusive upper bound on the time step like MP7

* ci: set pixi-setup frozen to true * update pixi.lock file * make sure smoke test is a smoke test * exclude test001h_rch_array4 and test001h_evt_array4 * fix openmpi version for Linux --------- Co-authored-by: Joseph Hughes <jdhughes@usgs.gov>

The last attempt apparently didn't work

…ps (#1714) * fix(ats-budget): cumulative budget incorrect with failed/retried ATS timesteps * add and require a finalize_step method for the budget object * cumulative terms are calculated only when the finalize_step method is called * required adding calls to budget%finalize_step() before any budget_ot() calls were made * update release notes

* refactor(swf): working on coordinate invariant overland flow * using dhds in denominator of conductance * working to get swr conductance in as a dev option * added capability to save velocity * updates to gradient implementation and to swr conductance dev option * fprettify * rebuild makefiles * tighter tolerance; fix to dev_swr_conductance * trying to get test to pass

* pin pixi version to 0.19.1 * bump setup-pixi to 0.5.2 * avoid breaking changes while pixi matures — pinning the action alone leaves us exposed to regressions in pixi itself

* test works in mf6 CI but not in flopy CI * allocate budget array instead

…. infiltration (#1719) * fix(tsp-apt): pair the user-specified concentration for finf with rej. infiltration * missing an escape character in the latex * apply same change to gwe-uze.f90 as was applied to gwt-uzt.f90 * change new function name to get_mvr_depvar for improved consistency

…i.lock (#1724) * use no_parallel marker in distribution tests * upgrade modflow-devtools and flopy for pixi

* add test that evaluates groundwater discharge with reach inflow equal to zero

* feat(disv2d): introduce limited 2D vertex grid (for overland flow) * needs testing * add io guide docs for disv2d * msvs proj and makefile updates * remove unnecessary files

…1727)

…1731) Showing the end of the list file in case of simulation failure makes CI logs extremely verbose, making it difficult to scroll/read. We upload failed test outputs as artifacts, so the list file can be inspected that way if needed.

@Hofer-Julian

…1728) Restore the custom task approach to installing flopy, modflowapi, pymake, and modflow-devtools originally introduced by @Hofer-Julian in #1696. This unlocks them (removes them from pixi.lock) and allows rolling updates in CI, as we had before #1703. Going forward, contributors will need to run `pixi run install` instead of `pixi install` to get a full development environment. Also remove hatchling from pixi.toml as it is not needed.

* fix a bug * update test to load the binary grid file * use consistent naming in binary grid file

* fix(uzf): rename uzf package variables to avoid name clashes * 9 uzf variables had counterparts in uzf%uzfobj with the same memory path * renamed the uzf package variables by adding _pvar to them * close #1741 This does not fix underlying memory problems with UZF. Instead it is a simple way to avoid multiple uzf variables with the same memory path. A proper fix will require additional work as the variable contents in uzf and uzf%uzfobj are not the same. * fprettifying * clean up a few variables that can be replaced with dis%top, dis%bot, and dis%area * gwet was named to gwet_pvar * fprettify

* GNC not allowed when parallel * 'gently' mpi abort instead of stop * fix interface mismatch

Co-authored-by: mjreno <mjreno@IGSAAA071L01144.gs.doi.net>

* perf(UZF): add deprecation warning for SIMULATE_GWSEEP * add variable to use only list * fprettify * Missing escape character in latex

* Add evaluation of DHNOFLO value in output tables in the listing file * Add evaluation of DHDRY value in output tables in the listing file * update BudgetObject to correctly handle SFR reaches without connections

…solver (#1688) * add ims convergence logic to PETSc solver * build petsc in release model for linux and mac * wire IMS preconditioner into PETSc solver * add 1-based ia,ja,amat to petsc matrix for local block * removed unused delta residual from precon * removed unused get_result fct * improved diagnostics * removed petsc serial code, unused * cleaned up newton test case * petsc convergence when exact solution found within machine precision * add level fill to IMS ILU pc in PETSc * use unpreconditioned residual * preallocate residual vector * cleanup convergence routine * add context constructor

* look up dfns in doc/mf6io/mf6ivar/dfn/ instead of hardcoded list * preserve top-level component ordering: sim, exg, sln, models, utl * order packages alphabetically within each component section * add argparse CLI allowing model selection, default to all

* select model types to include in docs/dfns/examples at release time * don't rebuild mf6io.pdf in build_documentation() if it already exists * don't build examples in dist build job (not needed) * last step toward feature toggling

* - move 'parallel' logic to mpi control * - incremental reorg. * - fix recursive model search * - allow asymmetry in send/recv - test still failing * - fix test: no element map when halo model has no actual contribution * - fix format

* fix(lake status): setting lake to inactive * implement fix for lake status set to inactive * update test * update develop.tex to reflect changes

* add BEDK sfrsetting in period block * add test of bedk with and without timeseries * update release notes * fix sfr rain observations for inactive reaches

* fix nightly build broken by #1746 * fix model selection, can't assume models provided * fix example filtering in build_dist.py, use rmtree not unlink * unrelated: remove old workaround from autotest/get_exes.py

* docs: spell-checking LaTeX and other files * docs: add codespell check of source and documents * add lint step to docs.yml

* refactor(ts): skip search for repeated boundary time series file The subroutine read_value_or_time_series checks to see if a time series file (TS) has been previously included in the time series manager (tsManager) before adding it. The check compares against the package name and the input parser row and column (ii, jj) location in the input file. If the package name and location match, then the TS is not added to the tsManager. This check is never true for boundary (BND) type TS, so a conditional is added to only perform the search for for non-BND TS types. This results in a substantial performance increase when there are a large number of time series files. * Remove link check for auxiliary variable * fprettify * cleanup --------- Co-authored-by: Langevin, Christian D <langevin@usgs.gov>

* function to adjust runoff and evaporation based on inflow

* words removed - blak, thck, numer, strng, htmp * added thckstrt

Optional support for local z coordinates. This allows positioning particles based on the water table. The motivations here are 1) modpath 7 parity and 2) to provide the same capabilities whether prt is running alongside a gwf model or as a post-processor. Previously dynamic positioning was possible only if prt runs as a post-processor.

…eters (#1723) * add enumeration type for discretizations * add get_dis_enum interface to dis types * add export_array_ascii option to additional packages * extend and refactor input export capability * add export_array_ascii tests for disv1d and dis2d * some cleanup * update export description for mf6io * assign exports as layered or not layered * remove export support for array-based input for now --------- Co-authored-by: mjreno <mjreno@IGSAAA071L01144.gs.doi.net>

* set boundary face flows for triangle disv grid test * also applies to test_prt_ternary_methods.py * unrelated fix: set found in prp module

* - streamline convergence csv report between IMS and PETSc solver - fix typo in csv header (inner report now contains model data for PETSc solver, similar to IMS) * - remove timer (fixes fp excep) * - model id printed instead of (local) index - a bit of cleanup/comments

Co-authored-by: scottrp <45947939+scottrp@users.noreply.github.com>

…mall change to assert setup (#1758)

Co-authored-by: GitHub <noreply@github.com>

* allow filein tags in mfsim.nam options block * update developer doc to describe updating fortran definitions * update fortran input definitions * additional detail added to developer doc --------- Co-authored-by: mjreno <mjreno@IGSAAA071L01144.gs.doi.net>

* continue excluding test014_NWTP3Low_dev until convergence failure resolved

) * feat(swf): implement proper perturbation for jacobian calculation * calculate jacobian based on C.T. Kelley method * refactor flow area calculation to be consistent with conductance calculation * fprettify * add unit test for get_perturbation

…pt (#1777) MSVS files for mf5to6 and zonebudget were out of date, update them and include them in the validation script we run in CI

…cases in parallel (#1780) - update petsc matrix before multiplication - backtracking now properly synchronized

…chInfo chapter (#1781) * fix(mf6io): synchronize mf6io with descriptions in forthcoming SuppTechInfo chapter * Move the discussion about tortuosity to the definition file * Tweak description of CND package, specifically its relationship with XT3D

…1782) * separate scripts for formatting and spellchecking * rename msvs_vfproj_check.py -> check_vfproj.py * rename fortran_format_check.py -> check_format.py * rename codespell_docs_check.py -> check_spelling.py * include unit tests in Fortran format checks, reformat where needed * use multiprocessing for significant speedup (~80s to ~15s for format checks on my machine)

* support binary for regular, numeric list input blocks * some cleanup * more cleanup * refactor to remove exchange reference in static loader --------- Co-authored-by: mjreno <mjreno@IGSAAA071L01144.gs.doi.net>

* Manager() creates a server process — as recommended, avoid starting new processes outside the main function so the module can safely be imported * avoid EOF errors

* previously corresponded to mp7 'terminated for unknown reason'

* limit zeromethod options to 1 brent, 2 chandrupatla in ternary solution * remove test method root function from MathUtil * cleanup, use named constants

…th CND write-up (#1787)

* update test_gwfgwf to read exchange data from binary input * add gwfgwf_lgr binary input tests * add disu binary exchange input test --------- Co-authored-by: mjreno <mjreno@IGSAAA071L01144.gs.doi.net>

… arrays (#1792) * minimum tracking level used to be 0 (simulation) but this is not necessary, move minimum tracking level to 1 (model) * allocate/resize particle arrays with memory manager, no longer need special handling for non-standard array bounds * update exclusions in check_format.py

* in progress: add hpc config file * - add concept for test * - fix merge error * - update from dfn * - load load balance from HPC file - connect UTL HPC subpackage - add ifind for charstring * fix msvs * - follow up on review comments * - generate new makefile * - also makefiles for zbud * - msvs for zonebudget * - isolated load balance in separate container (Distributed Simulation) * - add to msvs * - fprettify * - fprettify2 * - codespell * - fix PRT test * - HPC input: check load balance validity - increase autotest coverage * - validate hpc file only when hpc file exists * - one more test for hpc file (and splitter) * - remove metis dependency from autotest

* - suppress BUY message for interface model - add safety check to avoid bound exception in comparison * - format

…1796)

* autotest/build_mfio_tex.py is no longer used and out of date, remove it * distribution/build_docs.py::test_build_mf6io_tex_from_dfn instead — this is the script used in the release automation

* reimplement vertex velocity calculations such that local mass is conserved * fix for non-rectangular 4-sided cells erroneously solved via pollock's method * fix multiple floating point exceptions in voronoi subcells: div-0, underflow * add option dev_forceternary to force ternary solution for quad grids * refactor/optimize routine shared_face() for finding neighbor faces, move to GeomUtilModule * switch create/destroy to allocate/deallocate naming for derived type setup/teardown routines * define named parameters for magic numbers e.g. MAX_POLY_CELLS * use KindModule types throughout * use snapshot test fixtures to check pathlines and/or endpoints for select tests * test_prt_release_timing.py * test_prt_triangle.py * test_prt_voronoi1.py * remove development notes, miscellaneous cleanup * add syrupy and gitpython to environment.yml and pixi.toml, update pixi.lock Co-authored-by: Alden Provost <aprovost@usgs.gov>

…es (#1801) * remove unneeded get_npolyverts() routines from MethodDisv and CellDefn * refactor routine to get rectangular dimensions/rotation without unneeded args * rough benchmark comparing shared_face() with shared_edge() performance

* docs(mf6io): preparing docs for release * include reserved doi for version 6.5.0 * minor updates to mf6io to reflect idm changes * updates to mf6io tex files in response to recent changes * other minor updates * Undo my hack to remove swf from mf6io

* Update mf6core.f90 Fixed a typo * Revert "Update mf6core.f90" This reverts commit 98d27a0. * Update mf6core.f90 2 typos * Update src/mf6core.f90 Co-authored-by: Eric Morway <emorway@usgs.gov> --------- Co-authored-by: Eric Morway <emorway@usgs.gov>

…#1803) * factor out routines for cleaner MethodDisType and MethodDisvType * optimize cell definition load: merge loops when setting cell faceflows * rename destroy() -> deallocate() in method types * misc cleanup, improve docstrings

Package logging wasn't connected to iprpak in base type, wire it up. Also idm mempath was used where it should not be. And add some tests

…1806)

Store endpoint snapshot for mp7/prt examples and compare in test_prt_notebooks.py.

* give MethodDisvType a neighboring CellDefn, don't reallocate on each cell-cell transition * also remove example name workaround in test_prt_notebooks.py

* close #1804

* - fix matching of connection/interface models between GWF and GWT * - add parallel test case too

* docs(releasenotes): update release notes for 6.5.0 * codespell for the win! * grammar

* doc(supptechinfo): Update with new GWE chapter * Update doc/SuppTechInfo/groundwater-energy-model.tex Co-authored-by: langevin-usgs <langevin@usgs.gov> * More update for chapter 10 and an additional reference added to .bib --------- Co-authored-by: langevin-usgs <langevin@usgs.gov>

@scharlton2

…6io (#1813) * intel fortran preprocessor `fpp` substitutes "1" for "unix" on linux/mac * https://www.intel.com/content/www/us/en/docs/cpp-compiler/developer-guide-reference/2021-8/additional-predefined-macros.html, thanks to @scharlton2 for pointing this out * rename `unix`/`uniy` and related variables (interior/exterior unit normal components) to `unintx`/`uninty` etc * update options * move `zeromethod` option to prp, rename to `dev_exit_solve_method` * add `exit_solve_tolerance` option * update prt autotests and snapshots * add comment to brent's root-finding method * remove `dev_feature` blocking prt in `SimulationCreate` * remove prt obs for now, plan to support for next release * revise mf6io, regenerate mf6io files

@aprovost-usgs

Supplemental Technical Info chapter for PRT, courtesy of @aprovost-usgs --------- Co-authored-by: aprovost-usgs <aprovost@usgs.gov> Co-authored-by: wpbonelli <wpbonelli@ucar.edu> Co-authored-by: langevin-usgs <langevin@usgs.gov>

Add sections describing particle track binary and CSV output files Co-authored-by: aprovost-usgs <aprovost@usgs.gov>

Co-authored-by: wpbonelli <wpbonelli@ucar.edu>

* make EXIT_SOLVE_TOLERANCE required for DISV grids for all grids, motivation: one should be aware the generalized method is semi-analytical and subcells are solved numerically for grid cells we can't solve with Pollock's method * accept any value for DIS grids, following convention in e.g. rnb in gwf-csub.dfn packagedata where a sometimes-required option can have any value when not required * fix double comma in bib file

Mark some tests that need to be skipped in release mode

Fix test selection in large.yml, marker no longer exists with test: remove unneeded marker #1820

) * track_all isn't necessary as all events are enabled by default * "need" -> "needed` in supplemental technical info document

On rectangular grids, the generalized particle tracking method can find itself in an endless loop on macOS 14 compiled with gfortran with optimization=2, confirmed with gfortran 11-14. It's difficult to say where things are going wrong since on ARM macs neither gdb nor lldb can be used so we are stuck with print debugging. An architecture/compiler problem seems plausible since we'd presumably see the same behavior on other platforms with opt=2 if it were our bug, but I'm not sure. This should not occur typically since rectangular cells are detected and solved with Pollock's method, but we use dev_forceternary in test_prt_disv1.py to check that the generalized method reduces to Pollock's. The test has been hanging in CI recently — we might have caught it sooner but there was a separate setup-fortran issue which concealed this one at first. While it might be safe to release with optimization=2 (with the assumption developers are the only ones using dev options) I don't think we can be sure the same or similar issues could not affect other model grids. So I think we should drop the optimization level pending more information.

Add the developmode marker to parallel tests that reuse models from tests that use dev options

Apply 'not developmode' marker when testing parallel mf6 in test-par-win action... * on master branch, generally * if developmode is false in the release workflow

…ut (#1835)

* docs(MODFLOW6References.bib): correct url links * fix minor typos in OC * Add comment line to gwe apt example observation .tex files for improved readability * add missing comma in *.bib file

@rbwinst-usgs

Incorporate fixes and suggestions from @rbwinst-usgs' review. Co-authored-by: aprovost-usgs <aprovost@usgs.gov>

* unneeded since we are removing the list of distribution contents from release notes * remove mk_folder_struct.py, pixi task in pixi.toml, and steps from release.yml

* docs(releasenotes): more cleanup for 6.5.0 release * fix ftype table links in mf6io

* add optional tol dummy arg to nudge() routine in MethodSubcellTernary, better dummy arg naming, improve docstring * use LENBOUNDNAME instead of hardcoded 40 in mf6io document * clarify docstring for shared_face() routine in GeomUtil

Remove break statement likely leftover from development/debugging, which caused only the first example model to get a test run. This is why the examples script issue addressed by MODFLOW-USGS/modflow-devtools#156, where models were written to runall.[sh/bat] in an order which did not respect flow model dependencies, was not caught before. After this PR, the runall.[sh/bat] script will be tested, as well as 3 of the separate example scripts picked at random.

…ames (#1775) * check input length of model,exg,pacakge names * delay model package name error until all names checked --------- Co-authored-by: mjreno <mjreno@IGSAAA071L01144.gs.doi.net>

* need to consider whether necessary to test all individual example scripts too * logic to pick them randomly was broken, didn't respect gwf model dependencies

developmode, run_tests and models inputs were not passed to release.yml correctly

Release 6.5.0 #1845

Release 6.5.0 #1845

Commits on Feb 8, 2024

Commits on Feb 9, 2024

Commits on Feb 12, 2024

Commits on Feb 13, 2024

Commits on Feb 14, 2024

Commits on Feb 15, 2024

Commits on Feb 16, 2024

Commits on Feb 18, 2024

Commits on Feb 19, 2024

Commits on Feb 20, 2024

Commits on Feb 21, 2024

Commits on Feb 22, 2024

Commits on Feb 23, 2024

Commits on Feb 26, 2024

Commits on Feb 27, 2024

Commits on Feb 28, 2024

Commits on Mar 4, 2024

Commits on Mar 6, 2024

Commits on Mar 7, 2024

Commits on Mar 8, 2024

Commits on Mar 9, 2024

Commits on Mar 10, 2024

Commits on Mar 11, 2024

Commits on Mar 14, 2024

Commits on Mar 15, 2024

Commits on Mar 17, 2024

Commits on Mar 18, 2024

Commits on Mar 19, 2024

Commits on Mar 22, 2024

Commits on Mar 24, 2024

Commits on Mar 26, 2024

Commits on Mar 27, 2024

Commits on Mar 28, 2024

Commits on Apr 2, 2024

Commits on Apr 3, 2024

Commits on Apr 4, 2024

Commits on Apr 5, 2024

Commits on Apr 8, 2024

Commits on Apr 11, 2024

Commits on Apr 12, 2024

Commits on Apr 15, 2024

Commits on Apr 16, 2024

Commits on Apr 17, 2024

Commits on Apr 18, 2024

Commits on Apr 19, 2024

Commits on Apr 20, 2024

Commits on Apr 21, 2024

Commits on Apr 22, 2024

Commits on Apr 23, 2024

Commits on Apr 25, 2024

Commits on May 1, 2024

Commits on May 3, 2024

Commits on May 5, 2024

Commits on May 6, 2024

Commits on May 7, 2024

Commits on May 8, 2024

Commits on May 13, 2024

Commits on May 14, 2024

Commits on May 15, 2024

Commits on May 16, 2024

Commits on May 17, 2024

Commits on May 18, 2024

Commits on May 20, 2024

Commits on May 21, 2024

Commits on May 22, 2024

Commits on May 23, 2024