Enhance TC-Diag to use tc_diag_driver version 0.11.0 #2769
Comments
JohnHalleyGotway
added
type: enhancement
|
After updating to tc_diag_driver version 0.11.0, I reran the @robertdemariacira I did a little digging but didn't see an obvious problem or solution. Can you please take a look and advise? |
|
@JohnHalleyGotway That is an odd bug. It looks like you might be attempting to create a |
|
@robertdemariacira sorry for the delay. The call to |
|
@JohnHalleyGotway Also sorry for the delay. I created a fresh environment, installed the |
|
@robertdemariacira sorry for the wild goose chase. It working again well after deleting and reinstalling my feature branch. Must have been an issue with my environment. |
|
@john Halley Gotway Does this mean things are ready for
me to generate retrospective data for Minna's development? If so, please
advise on whether any environment or setup needs to be done. Thanks.
|
|
@robertdemariacira, while I'm able to run tc_diag with the updated driver version, I'm still seeing missing data in the output that shouldn't be missing: SHR_MAG, SHR_HDG, 850TANG, 850VORT, and 200DVRG. Many of the computed diagnostic values also differ when compared against the CIRA diags. But I'm primarily concerned with the missing data being written. Running this test command with the I see the following in the output: Using compute_tc_diag.py, post_resample.yml, v2023-04-07_gdland_table.dat, and a sample NetCDF file I've sent previously. Any idea why this isn't working? I did check that the U/V names in the yml config file have already been updated. |
|
@JohnHalleyGotway Ran the provided script and config locally using the previously provided netCDF file: tmp_tc_diag_AL092022_GFSO_2022092400_f0_parent_13052_0_.nc I'm getting valid values for all of the variables you described. Would you mind providing the netCDF file you're using so I can compare? Maybe a name was changed between what I'm using and what you're using? |
|
@robertdemariacira thanks for sending over the NetCDF file you're using. Unfortunately, I'm still getting bad behavior: So its not an easy variable name difference. |
|
OK, I'm seeing the issue. I did NOT have scipy installed and get a warning message about that: Without scipy, I get bad data. After installing it, I get good data values: I believe we thought that scipy was NOT a requirement, but it appears that it actually is. Here's the GitHub comment where we discussed this. |
|
Sorry about that. None of my code is directly making use of scipy, but I am relying on pandas to interpolate between pressure levels and that, in turn, is relying on scipy. I wrote that part almost two years ago and forgot about that dependency. |
|
Great to see that you might be getting to the bottom of this. Thanks!
|
|
OK, so am I correct in interpreting this as meaning that scipy actually is a requirement for the driver code? @georgemccabe and @jprestop does this raise any red flags in adding scipy to list of required Python packages when running the TC-Diag tool? |
|
As far as I know, it should be OK to add scipy as a requirement for running TC-Diag. We should be careful to make sure to isolate this dependency so that the rest of the tools can be run without scipy available. We should also make sure that the error message that is reported when scipy is not found when running TC-Diag is clear. There are a few METplus wrappers that have additional Python dependencies beyond the normal requirements of the wrappers. I handle this by putting those import statements in a try/except block. I'm happy to go over the details of this if that is helpful. |
|
I am not aware of any red flags in adding scipy. It is on the list of allowed packages at WCOSS2, so this looks fine with regard to WCOSS2. |
|
@JohnHalleyGotway It would be possible to remove the dependency on scipy, but that would be a pretty time consuming change. |
|
@robertdemariacira sounds like I should just add scipy to our development environment. And should we just stick with its absence resulting in a warning message: And then let some of the diags be bad data? Or should we update the driver code to error out instead of providing partial results? |
|
@JohnHalleyGotway If we're considering scipy to no longer be an optional dependency, my preferred behavior would be to have the code instantly fail and raise an exception when it can't import scipy. This exception could be handled by compute_tc_diag.py in any way you wish. However, if you prefer letting the code continue but just generating missing data, the code currently generates error logs that could be captured by compute_tc_diag.py. If a variable fails to compute for any reason, the exception stack trace will appear in the logs. |
|
@robertdemariacira thanks for shaking out the scipy issue. With it installed, I am getting valid diagnostic values but the results differ considerably from the previously generated CIRA output. Here's the full set of diagnostic output for CIRA (from @musgrave-kate) and MET for the same case: But let's start with a single timestep. Here's a call to the Python driver code for forecast hour 0: The input/output from this command can be found in this diag_example.tgz file. In particular, look in the Comparing them to the values in this CIRA diagnostics file the numbers are VERY different: Can you please run this example through your tc_diag_driver code and confirm that your results match the output in If not, some issue remains in updating the tc_diag_driver Python code in MET. I'm just try to figure out next steps in debugging these diffs. |
|
@JohnHalleyGotway Ran the provided compute_tc_diag.py script in my environment using the provided input files and it produced output identical to your diag_f0.txt. |
|
@robertdemariacira great, thanks for confirming that. So the issue is likely in the NetCDF cylindrical coordinates data we're passing as input to the tc_diag_driver code. Any suggestions on how we can debug that further? @jvigh perhaps I should ask you to run a test case with the existing TC-Diag Python code so that we can be sure we're making fair comparisons? |
|
@JohnHalleyGotway It seemed like you had a fairly easy way to switch the resampling from nearest neighbor to bilinear interpolation? If you have the ability to control the number of neighbors it uses, then it should just use the 4 nearest neighbors. If you haven't already done so, try switching that and see if the results get closer. ( The U and V fields are pretty noisy, so small changes to the interpolation method are likely to result in large changes to the contents of the cylindrical grid.) I think an apples to apples comparison among |
|
@musgrave-kate and @jvigh I wanted to call your attention to this very long commit message: This change enforces consist diagnostic name and units string field widths. But it also introduces the potential for a downstream parsing problem. The MET If someone were to add a custom diagnostic with units like And you can see the obvious problem. There's no whitespace between columns 2 and 3... so parsing it as being whitespace delimited won't work. For now, I'll just ignore this potential problem and assume that the priority is exactly matching the spacing of the existing output. I just wanted to point out the potential issue. Please let me know if I should handle this differently. |
|
@JohnHalleyGotway, thanks for highlighting this. My guess is that the lack of whitespace could
be a problem in some settings. But readers using fixed column positions
(e.g., Fortran) probably would be okay. But changing the default might mess
those readers up.
Kate probably has a better idea of what programs would be reading these, so
I'll defer to her.
|
…favor of the MET_KEEP_TEMP_FILE environment variable option that Howard is adding on his feature_2772_python_embedding_json branch. I did add documentation about that option here even though the functionality is coming from Howard's branch.
JohnHalleyGotway
changed the title
JohnHalleyGotway
changed the title
…e Python environment which is needed by TC-Diag.
…nce we didn't get it out on the 6th.
* Per #2769, update to tc_diag_driver version 0.11.0. * Per #2769, add tmp_nc_diag_flag config option to retain temporary files. * Per #2769, change default regrdding method from NEAREST to BILIN. * Per #2769, add checks to make sure input files exist. * Fix the handling of diagnostic name and units strings. Both are limited to 7 characters but the units are enclosed in parantheses, increasing the field width to 9. If the strings are too long, truncate them and print warnings like this: WARNING: OutFileInfo::write_cira_diag_vals() -> long diagnostic name "850TANGXXX" truncated to "850TANG"! WARNING: OutFileInfo::write_cira_diag_vals() -> long diagnostic units string "(10^7C/MXXX)" truncated to "(10^7C/M)"! Pad the first 2 columns out to widths of 7 and 9 and set the inter-column spacing between columns 2 and 3 to 0. This could technically introduce parsing problems when the units are 7 characters long and the values have 5 digits, but this is the logic needed to exactly replicate the existing output. * Just whitespace * Per #2769, remove the tmp_nc_diag_flag TC-Diag config file option in favor of the MET_KEEP_TEMP_FILE environment variable option that Howard is adding on his feature_2772_python_embedding_json branch. I did add documentation about that option here even though the functionality is coming from Howard's branch. * Per #2769, remove tmp_nc_diag_flag that has been removed. * Per #2769, remove mention of tmp_nc_diag_flag option that has been removed. * Per #2769, upgrade to METbaseimage version 3.2 to provide SciPy in the Python environment which is needed by TC-Diag. * Per #2769, add SciPy to the list of required Python packages. * Per #2769, tweak AppendixF details. * Per #2769, update release date for MET-12.0.0-beta3 to be 20240207 since we didn't get it out on the 6th.
* Store change of sonar.login to sonar.token. * Hotfix to develop, updating aclocal.m4 and config.h.in to what is created when running bootstrap inside the dtcenter/met-base:v3.1 Docker image. Updates autoconf 2.69 output to 2.71. * Adding -lnetcdf to configure_lib_args for NetCDF-CXX compilation * Added eckit and atlas loads and paths * Added atlas and eckit loads and paths * Changing -j to "-j 5" as the recommended value * Update ensemble-stat.rst Fixed incorrect path to obs error table * Updated values for GRIB2CLIB_NAME and BUFRLIB_NAME * Updated for proj, eckit, and atlas * Feature #2761 develop seneca (#2762) * Per #2761, define runtime python version for testing rather than using the default version which no longer exists in /usr/local * Per #2761, fix setting ci-skip-all * Per #2761, patching test_util.R to use the -C command line option for ncdiff. I did test running comp_dir.sh with this change and confirmed that it now runs to completion. * #2652 Added find_var_by_standard_name and separated common codes to find_xy_vars * #2757 Get the email list from the environment variable MET_CRON_EMAIL_LIST__MET (or MET_CRON_EMAIL_LIST_) * #2757 THe SonarQube token and URL are replaced by ujsing the environment variable SONAR_TOKEN_VALUE and SONAR_SERVER_URL * #2757 The SonarQube token and URL are replaced with the pre-defined strings, SONAR_TOKEN_VALUE and SONAR_SERVER_URL * Bugfix #2670 develop --enable-python (#2768) * #2755 Added a header count and checking header count instead of using header id (hid) * Update install_met_env.acorn_py3.10 * Update install_met_env.wcoss2_py3.10 * Feature #2776 cleanup (#2777) * Per #2776, delete stale prob_pair_info.h/.cc, wwmca_utils.h/.cc, and rmw_analysis_utils.h/.cc files that were not being compiled by their Makefiles. * Per #2776, update Makefile to compile the test_read_rmw utility. * Per #2776, add Makefile for the vx_python3_utils test utility directory. * add rpath for atlas and eckit lib dirs so dynamic libraries can be found when running on seneca * only set -L and -I arguments for atlas and eckit if the appropriate environment variable is set -- this matches how it is handled for other external library dependencies * Per #2776, just whitespace and capitalization --------- Co-authored-by: George McCabe <[email protected]> * Bugfix #2782 develop MASSDEN (#2783) * Per #2782, port over fixes from the bugfix_2782_main_v11.1_MASSDEN branch over to the bugfix branch for develop. Will also add new GRIB2 filtering options in this branch. * Per #2782, add support for 4 new GRIB2 filtering options for GRIB2_aerosol_type, GRIB2_aerosol_interval_type, GRIB2_aerosol_size_lower, and GRIB2_aerosol_size_upper. These are useful in filtering the MASSDEN records in the RRFS smoke output files. * Per #2782, fix aerosol_size_lower() and aersol_size_upper() inline definitions. * Per #2782, add a unit test for GRIB2 table 4.48. * Per #2782, switch from strict equality to using the is_eq() function when checking the GRIB2_aerosol_size_lower and upper values since they're doubles and not ints. * Feature #2701 ismn (#2758) * Add support for ISMN soil moisture data * Per #2701, update ascii2nc.cc with the ismn file option. * Per #2701, update ascii2nc docs * Per #2701, store the depth for precip as 0. * Per #2701, parse the ISMN observation value from the correct column. * Per #2701, every time I run ascii2nc to I see a log message for 'DEBUG 1: Number of NDBC skipped files due to no lookup 0'. This is printed by the NdbcHandler destructor. Update the logic to only print a debug level 3 log message when the number of missing locations is greater than 0, 'DEBUG 3: Skipped 5 NDBC files whose locations are not defined in 'MET_BASE/table_files/ndbc_stations.xml'. Set the MET_NDBC_STATIONS environment variable to override this file.' * Per #2701, as instructed by @anewman89, store the average of the depth values rather than the maximum value. * Per #2701, add a test of processing the ISMN data through ascii2nc. * Per #2701, fix parsing of year and month to subtract 1900 and 1, respectively. * Per #2701, set use_var_id to true for ISMN inputs * Per #2701, even though we're mapping obs data to GRIB code names and units, we want to encode them with use_var_id = true and so we need to keep track of those var_id values. * Per #2701, fix typo in the name of the ascii2nc netcdf output file. --------- Co-authored-by: MET Tools Test Account <[email protected]> * Feature 2753 comp script config (#2771) * set dynamic library file extension to .dylib if running on MacOS and .so otherwise * Added disabling of jasper documentation for compiliation on Hera * Updated * remove extra export of compiler env vars * include full path to log file so it is easier to file the log file to examine when a command fails * send cmake output to a log file * remove redundant semi-colon * use full path to log file so it is easier to examine on failure * use run_cmd to catch if rm command fails * Modifications for compilation on hera, gaea, and orion * Updating * fixed variable name * clean up if/else statements * set TIFF_LIBRARY_RELEASE argument to use full path to dynamic library file to prevent failure installing proj library * set LDFLAGS so that LDFLAGS value set in the user's environment will also be used * Updated based on gaea, orion, and hera installs * Updated * change extension of dynamic library files only if architecture is arm64 because older Macs still use .so * added netcdf library to args to prevent error installing NetCDF-CXX when PROJ has been installed in the same run of the script -- PATH is set in the COMPILE_PROJ if block that causes this flag from being added automatically * clean up how rpath and -L are added to LDFLAGS so that each entry is separate -- prevents errors installing on Mac arm64 because multiple rpath values aren't read using :. Also use MET_PROJLIB * Updated * removed -ltiff from MET libs * only add path to rpath and -L arguments if they are not already included in LDFLAGS * changed from using LIB_TIFF (full path to tiff lib file) to use TIFF_LIB_DIR (dir containing tiff lib file). Added TIFF_INCLUDE_DIR to proj compilation and -DJAS_ENABLE_DOC to jasper compliation taken from @jprestop branch * update comments * ensure all MET_* and MET_*LIB variables are added to the rpath for consistency * remove unnecessary if block and only export LDFLAGS at the end of setting locally * Updated * Added section for adding <VALUE>/lib64 and rearranged placement of ADDTL_DIR * Commenting out the running of the Jasper lib tests * Updating and/or removing files * Updating and/or removing files * Latest udpates which include the addition of the tiff library for proj * Remove commented out line. Co-authored-by: John Halley Gotway <[email protected]> * Make indentation consistent. Co-authored-by: John Halley Gotway <[email protected]> * Make indentation consistent. Co-authored-by: John Halley Gotway <[email protected]> * Make indentation consistent. Co-authored-by: John Halley Gotway <[email protected]> --------- Co-authored-by: George McCabe <[email protected]> Co-authored-by: John Halley Gotway <[email protected]> * #2697 Share the temporary file for blocking * #2697 Reduced the temporary files for pb2nc * Moved typedef unixtime from vx_cal.h to time_array.h * #2697 Changed static const to constexpr for SonarQube (code smell) * Removing ${MAKE_ARGS} in some locations Removing ${MAKE_ARGS} from "make install" and "make test" for MET. Removing "met" prefix from met.configure.log because we really need config.log for any detail. It is confusing to have met.configure.log when that does not contain useful information. * #2673 Removed unused variable grid_x and grid_y * #2673 Renamed the shadowed variable n to nd * #2673 Use nullptr instead of literal 0 * #2673 Moved down namespace declarations. Removed unused variables * #2673 Reduced the scope of variables * #2673 Use nullptr instead of literal 0 * #2673 Removed unused variables * #2673 Renamed ex to ex2 which becomes shadowed variable * #2673 Removed always true condition * #2673 Moved down namespace declarations. Use nullptr instead of literal 0 * #2673 Use nullptr instead of literal 0. Removed unused variables * #2673 Moved down namespace declarations. * #2673 Added namespace std to string * #2673 Removed unsued variables. Check nullptr of gt * #2673 Moved down namespace declarations. * #2673 Added namespace std to string * #2772 Added quit_msg * Feature #2547 Read WRF output files natively (#2790) * Per #2547, add the file_type = NETCDF_WRF configuration option. * Per #2547, rename vx_data2d_nc_interp as vx_data2d_nc_wrf and vx_data2d_nccf as vx_data2d_nc_cf. * Per #2547, rename nccf files as nc_cf for consistency throughout * Per #2547, rename Pinterp classes as Wrf * only set -L and -I arguments for atlas and eckit if the appropriate environment variable is set -- this matches how it is handled for other external library dependencies * add rpath for atlas and eckit lib dirs so dynamic libraries can be found when running on seneca * ignore directory automatically created by CLion * rename pinterp variables and classes to Wrf * more rename and fix typos * per #2547, add support for reading WRF files by adding support for different X/Y/Z dimensions, read X/Y/Z dimension names when reading a field instead of storing them per file, track if data is staggered in X/Y/Z dimension (still need to add DataPlane function to destagger data), handle pressure field that also includes time dimension * per #2547, add DataPlane function to handle staggering of grid * per #2547, change file type returned from pinterp to wrf * per #2547, call function to stagger DataPlane if necessary * per #2547, keep store if variable is on pressure levels and only read pressure field if it is * ignore cmake directory * per #2547, support U_PL, U_ZL, U, and U<x> where <x> is an integer in is_u_wind and the same for is_v_wind * change regex to match zero or more instances of a number after U or V instead of 1 or more to include the fields named U and V * error and exit if var with unsupported subgrid dimension is processed * clean up logging * only destagger Z dimension for bottom_top_stag because other _stag Z dimensions don't need to be destaggered, clean up logic to read X/Y/Z dimension info * Per #2547, instead of checking dimensions of all vars in file, check dimensions for the var that is being read. Only read pressure var if the z dimension matches the var to read. Set class variable DimNames instead of local gDimNames so dump function will work as expected * per #2547, support reading variable that corresponds to dimension with different name as dimension and includes a time dimension, e.g. P_PL(Time, num_press_levels_stag) * per #2547, added unit tests to process WRF out data with plot_data_plane and grid_stat * merged develop and resolved conflicts * rename lib from pinterp to wrf * rename lib * fixed incorrect merge changes * fixed broken unit test format * Per #2547, restore the mention of NAK.poly to Appendix B. * Change model to WRF Co-authored-by: John Halley Gotway <[email protected]> * change obtype to WRF Co-authored-by: John Halley Gotway <[email protected]> * plot higher vertical level to get non-zero output Co-authored-by: John Halley Gotway <[email protected]> * indentation Co-authored-by: John Halley Gotway <[email protected]> --------- Co-authored-by: John Halley Gotway <[email protected]> * #2772 Use JSON for attrubutes and numpy serialization dor 2D data instead of NetCDF * Hotfix to the develop branch after PR #2790 merged the feature_2547_wrf branch into develop. Rerunning bootstrap on the project machine seneca produces 2 small differences that should also be incorporated into develop. * Feature 2588 install rewrite (#2791) * updated first 3 sections * updated spacing * adding the Environment Variables to Run Script section * adding in the remainder of the google doc changes * trying to fix duplicate explicit target name issue * fixing required_external_libraries_to_build_met links * HDF4 testing double underscore * adding double underscores for HDF5 and HDF-EOS2 and changing library back * fixing HDF5 web link name * Added formatting, made minor edits, removed manual build instructions sections * Removed reference to sample-test-case which no longer exists * Resolving errors * Minor formatting change * Updated apptainer instructions, new version references * Formatting updates and removal of first person language * fixed code blocks * Resolving error * Removed first person references * Removed reference to met_directory_structure * updated bolding, removed end section * Changing X.Y.X references to X.Y.Z * updating thru Using the compile_MET_all.sh Script * fixing spacing for ATLAS web address * Fixed broken internal reference to installation due to typo * adding version numbers * fixing Recommended-Components MET * 3.4.2 section updates * fixing spacing * adding a period and more dropdown menus * fixing spacing * loose ends * removing space * removing bold * External Library handling in compile_MET_all.sh section updates * installation directory view * installation directory after * attempting to add figures via the web updating Executing the compile_MET_all.sh script section * fixing bolds, italics etc. * updating last half of doc * loose ends * loose ends * adding more dropdown menus * fixing dropdown menu and testing automatic version * testing automatic version * confirmed automatic version cannot be added to italics or code blocks * another round of updates * fixing indents * creating a table * fixing table * fixing table 2 * more loose end changes * more changes * hopefully final updates * trying a fancy new csv table * troubleshooting csv table * correcting table name * adding a grid table * fixing grid table * removing csv table * Per #2588, update to the environment variables table. * Delete docs/Users_Guide/environment_variables_table.csv * Per #2785, add new example installation files * updating small changes in the first half of the documentation * new dropdown menu and fixing some links * trying to fix line spacing in dropdown * take 2 fixing spacing * take 3 fix spacing * env var order * removing "s" * small changes * un capitalizing S * A few minor changes --------- Co-authored-by: j-opatz <[email protected]> Co-authored-by: Julie Prestopnik <[email protected]> Co-authored-by: j-opatz <[email protected]> Co-authored-by: John Halley Gotway <[email protected]> * Minor hotfix to develop to fix a typo in the comments of the PB2NC config files. * Feature #2796 develop gha node20 (#2797) * Per #2796, update actions/checkout@v3 to actions/checkout@v4. * Per #2796, update to actions/upload-artifact@v4 and actions/setup-python@v5. * Per #2796, update to download-artifact@v4 and actions/github-script@v7 * Feature #2796 develop gha_node20, fix artifact names (#2799) * Per #2796, update actions/checkout@v3 to actions/checkout@v4. * Per #2796, update to actions/upload-artifact@v4 and actions/setup-python@v5. * Per #2796, update to download-artifact@v4 and actions/github-script@v7 * Per #2796, update testing workflow for update-artifact v4 version which requires unique artifact names. ci-run-unit * Per #2796, merge the logs into a single artifact. ci-run-unit * Per #2796, renaming rc_leads and rc as 1c and 2c, respectively for consistency purposes. Note that this commit should result in just a logs_compile artifact being created... and not doing the log merging step. * Per #2796, update the compilation workflow to make the log artifact names unique. * #2772 Change nan and inf to -9999 on reading ASCII input if failed to parse * Feature #2801 warnings (#2802) * Per #2801, add new time_offset_warning config option. * Per #2810, update MetConfig class to parse and check the time_offset_warning config option. * Per #2801, update Grid-Stat, MODE, PCP-Combine, Series-Analysis, Wavelet-Stat, and the WWMCA tool to call MetConfig::time_offset_warning() to decide whether to print a Warning or Debug(3) log message about initialization or valid time differences. * Per #2801, no change to content. Just removing unneeded whitespace. * Per #2801, remove extra empty line foromw m warning message. * Per #2801, remove extra empty line from MODE warning messages. --------- Co-authored-by: MET Tools Test Account <[email protected]> * #2772 Initial release, Separated from point.py * #2772 Added point_nc.py * #2772 Changed write_tmp_nc and read_tmp_nc to write_tmp_py and read_tmp_py * #2772 Removed python_key_point_data & python_key_point_data_list and replaced them to tmp_point_var_name and tmp_point_data * #2772 Renamed tmp_nc_base_name, tmp_nc_file_var_name & tmp_nc_point_var_name to tmp_py_base_name, tmp_file_var_name, & tmp_point_var_name * #2772 More log messages for error * #2772 Changed API (log_msg to log_message) * #2772 Use met_point_nc_tools instead of met_point_tools * #2772 Changed APIs * #2772 Changed API * #2772 Changed default temp output format to JSON ande numpoy serialization * #2772 Allow to keep the temporary files by using the environment variable * Added log message if the temprary fiule was not deleted * Feature #2745 mvmode enhancements (#2779) * Working mvmode with percentile thresholds and grid_type fixes, plus more unit tests * separated multivar and traditional mode frontend classes to simplify the mode executive class design * documentation changes to go with changes to percentiles and data_type * Cleanup and bugfix to doc * Bug fix to doc * Another bugfix attempt * synching up development.seneca with what is in the develop branch * added clone() method to all the vx_data2d var_info classes * Fix multivar_name/level definitions in UG * bug fix regarding accessing the config to create a verification grid * Update docs/Users_Guide/mode.rst * Update docs/Users_Guide/mode.rst * Per #2745, commit changes to Makefile.am/.in files after running bootstrap on this feature branch on seneca. * Per #2745, no real code changes. Just whitespace. --------- Co-authored-by: MET Tools Test Account <[email protected]> Co-authored-by: Tracy Hertneky <[email protected]> Co-authored-by: John Halley Gotway <[email protected]> * Per #2772, added MET_PYTHON_EXE to various test cases and removed what would have been a duplicate after adding MET_PYTHON_EXE * Per #2772, add documentation about 3 new environment variables. * Per #2772, tweak the wording. * Fixing typo * Feature #2772 python_embedding_tmp_file (#2807) Co-authored-by: Howard Soh <[email protected]> * Feature #2805 filter_set_hdr (#2806) * Per #2805, update STATAnalysisJob::dump_stat_line() to take an optional argument to indicate whether the set_hdr options should be applied. Update Stat-Analysis to only call dump_stat_line() with set_hdr for filter jobs. * Per #2805, update User's Guide with -set_hdr info. * Per #2805, the -set_hdr option doesn't apply to tc_stat * Per #2805, modify one Stat-Analysis filter job to demonstrate that the -set_hdr job command option is now applied to the -dump_row output. * Per #2805, enhance the R differencing logic to handle NA strings in the header columns. Previously, differences that include the NA string were ignored. This enhances the compareStatLty() function to handle NA strings properly. * Per #2805, no real code change. Just whitespace. --------- Co-authored-by: MET Tools Test Account <[email protected]> * Feature 2772 python embedding fill value (#2810) * #2772 Update masked value to -9999. Renamed error_messageg to error_message * #2772 Renamed error_messageg to error_message * #2772 Accept 4th argument as user defined fill value. * #2772 Added python_numpy_plot_data_plane_missing which processes user defined fill value --------- Co-authored-by: Howard Soh <[email protected]> * Feature #2809 nc_laea (#2811) * Per #2809 work in progress parsing LAEA grids from MET NetCDF files. * Per #2809, update the LaeaData struct to store the spheroid_name as a character array. I tried very hard to store this as a string instead but ran into lots of problems with segfaults when deleting the allocated structs. Using a fixed length character array while making calls to the m_strncpy() and m_strlen() utility functions enables the code to work properly while hopefully minimizing SonarQube findings. * Per #2809, call m_strncpy() to set LaeaData::spheroid_name * Per #2809, add tests to pcp_combine and plot_data_plane to demonstrate the reading the LAEA data from MET NetCDF files. * Per #2809, correct the expected PostScript file output name in unit_plot_data_plane.xml * Feature #2809 v12.0.0_beta3 (#2814) * Per #2809, update copyright to 2024 * Per #2809, add release v12.0.0-beta3 release notes. * Feature #2769 tc_diag_driver_v0.11.0 (#2812) * Per #2769, update to tc_diag_driver version 0.11.0. * Per #2769, add tmp_nc_diag_flag config option to retain temporary files. * Per #2769, change default regrdding method from NEAREST to BILIN. * Per #2769, add checks to make sure input files exist. * Fix the handling of diagnostic name and units strings. Both are limited to 7 characters but the units are enclosed in parantheses, increasing the field width to 9. If the strings are too long, truncate them and print warnings like this: WARNING: OutFileInfo::write_cira_diag_vals() -> long diagnostic name "850TANGXXX" truncated to "850TANG"! WARNING: OutFileInfo::write_cira_diag_vals() -> long diagnostic units string "(10^7C/MXXX)" truncated to "(10^7C/M)"! Pad the first 2 columns out to widths of 7 and 9 and set the inter-column spacing between columns 2 and 3 to 0. This could technically introduce parsing problems when the units are 7 characters long and the values have 5 digits, but this is the logic needed to exactly replicate the existing output. * Just whitespace * Per #2769, remove the tmp_nc_diag_flag TC-Diag config file option in favor of the MET_KEEP_TEMP_FILE environment variable option that Howard is adding on his feature_2772_python_embedding_json branch. I did add documentation about that option here even though the functionality is coming from Howard's branch. * Per #2769, remove tmp_nc_diag_flag that has been removed. * Per #2769, remove mention of tmp_nc_diag_flag option that has been removed. * Per #2769, upgrade to METbaseimage version 3.2 to provide SciPy in the Python environment which is needed by TC-Diag. * Per #2769, add SciPy to the list of required Python packages. * Per #2769, tweak AppendixF details. * Per #2769, update release date for MET-12.0.0-beta3 to be 20240207 since we didn't get it out on the 6th. --------- Co-authored-by: MET Tools Test Account <[email protected]> Co-authored-by: root <root@83062d57c5dd> Co-authored-by: Julie Prestopnik <[email protected]> Co-authored-by: Christina Kalb <[email protected]> Co-authored-by: John Halley Gotway <[email protected]> Co-authored-by: Howard Soh <[email protected]> Co-authored-by: Howard Soh <[email protected]> Co-authored-by: George McCabe <[email protected]> Co-authored-by: metplus-bot <[email protected]> Co-authored-by: lisagoodrich <[email protected]> Co-authored-by: j-opatz <[email protected]> Co-authored-by: j-opatz <[email protected]> Co-authored-by: davidalbo <[email protected]> Co-authored-by: Tracy Hertneky <[email protected]>
Describe the Enhancement
On Dec 8, 2023, @robertdemariacira provided an updated version v0.11.0 of the tc_diag_driver (private repo). That update addresses the issues discussed in this comment for issue #2729.
This issue is basically a subtask of #2729, to be done immediately during the MET-12.0.0-beta3 development cycle. Further TC-Diag development will occur in the beta4 and beta5 cycles.
The tasks for this issue include:
AVNOtoGFSO. Yes, confirmed with @musgrave-kate on Nov 29, 2023 that this is desired.SOUNDING DATAoutput section. This is a problem and a space is needed.TGRDdiagnostic is missing from the MET output.unit_tc_diag.xmlto run a realistic HWRF parent + nest example. Be sure to test with with and withoutMET_PYTHON_EXEset. Use Hurricane Ian example for which we have diagnostics output and @KathrynNewman is providing gridded HWRF test data. Note that d01 should be used as the "parent" and d03 as the "nest" data source.MET_KEEP_TEMP_FILEenvironment variable to retain temp files.Time Estimate
1 day.
Sub-Issues
Consider breaking the enhancement down into sub-issues.
None at this time.
Relevant Deadlines
List relevant project deadlines here or state NONE.
Funding Source
2770043
Define the Metadata
Assignee
Labels
Milestone and Projects
Define Related Issue(s)
Consider the impact to the other METplus components.
No impacts currently.
Enhancement Checklist
See the METplus Workflow for details.
Branch name:
feature_<Issue Number>_<Description>Pull request:
feature <Issue Number> <Description>Select: Reviewer(s) and Development issue
Select: Milestone as the next official version
Select: MET-X.Y.Z Development project for development toward the next official release
The text was updated successfully, but these errors were encountered: