GitLab

vera/vera-dev#3638

Adds full support and testing. Also adds support in HDF5 where PLs are read and written.

* auto-open HDF5 and use REQUIRE more

* Fix some formatting issues in HDF5

* Add vs code to gitignore

* Update HDF5 test

* Add's 5D, 6D, and 7D support for SSK, SLK, SNK, and SDK parameters

* Add support for higher order int/real arrays when reading/writing HDF5 parameter lists

* Fix testTimes to work with virtual systems and WSL

* Add space in error message for HDF5 file

* Address review comments

* Add support to ParameterLists for 4D SNK, SLK, SSK, and SDK arrays

* Add PLPtr wrapper class to ParameterLists

* Consolidate reallocation statements in ParameterLists

* Add support to HDF5 for reading rank 4 arrays into a PL

* Add capability to HDF5 to get data type and shape

* Add some protection for unneeded / in HDF5

* Remove spurious write statement

* Add missing preproc symbols

* Address unused variables

* Automatically create parent directories in HDF5 file

This eliminates the need explicitly create directories
in many places, an unneeded burden on client code.

* Improve robustness of recursive mkdir

* Fix issues with pathExists

Graham, Aaron authored Jan 15, 2020 and Stimpson, Shane G committed Jan 15, 2020

* Address code that does not adhere to the code standard

Fixes 3 things:
1. Removes author, data, revision, and dependencies from module headers
2. Removes unneeded indentation inside MODULE statments
3. Removes unneeded indentation inside module CONTAINS statements

* Fix coinuation line indentations

* Fix SELECTTYPE and SELECTCASE indentations

* Fix indentation for PROGRAM and MODULE blocks in unit tests

* Fix continuation lines and SELECT TYPE/CASE in unit tests

HendersonSC authored Jan 13, 2020 and Stimpson, Shane G committed Jan 13, 2020

* Removes unused parameter definitions

* Re-add parameter inside preproc

* Add test for unused_parameters to CI

* Last few unused_parameters

* Rearrange preproc to remove unused variable

HendersonSC authored Jan 08, 2020 and Stimpson, Shane G committed Jan 08, 2020

* Implements string replace method and testing

* Clean up StringType doxygen, add check for clear method

* Add testing. Fix doxygen comments

HendersonSC authored Nov 06, 2019 and COLLINSBS email committed Nov 06, 2019

* Implement string library like functionality

* Adds a check to ensure strings are allocated correctly

* Remove double colon from module procedure

* Adding character string array as attribute

* Adding read attribute for character and updates unit test

Also adds error checks to ensure that new data sill fit into the
current data type and data set sizes when overwriting pre-existing
data sets.  Updated the test accordingly.

Needed to init a bunch of varables to -1 to avoid uninit variable
Valgrind errors.

bkochuna authored Apr 11, 2018 and COLLINSBS email committed Apr 11, 2018

Description:
Also updating dependencies. This is done to avoid
potential namespace collisions when integrating with other
repos/packages since "DBC" may be a common header or module.

CASL Ticket # - #5114

Salko Jr, Robert authored Feb 05, 2018 and COLLINSBS email committed Feb 05, 2018

* Add new file status to allow overwriting HDF5 datasets

Description:
Github issue 74.
Also add procedure for quieting HDF5 exceptions for cleaner testing.

CASL Ticket # - 5261

* Making test on FileType_HDF5 overwrite abilty more robust

Description:
Gitub issue 74.

CASL Ticket # - 5261

Stimpson, Shane G authored Jan 30, 2018 and COLLINSBS email committed Jan 30, 2018

CASL Ticket $ - 5341

Description:
Adding some logic so that very small non-scalar data sets are
not compressed. Any dataset < 1 MB should not get compressed.

CASL Ticket # - #5200

Description:
The existing interface just had a true/false logical for
an option argument of whether or not to use compression. This
option is changed to an integer so that the level of compression
can be specified.

CASL Ticket # - #5200

Description:
Makefile misses this module dependency and therefore build
errors may arise from compiling things out of order.

Description:
Interface had a dummy argument with an HDF5 kind being used.
Modified to SLK.

CASL Ticket # - #5200

Description:
* Fixed typo in comment
* Clarified description of getChunkSize
* Added REQUIRE() to getChunkSize instead of IF block
* Fixed valgrind error in compute_chunk_size
* Added assertion to large array write
* Fixed error behavior in postRead with compression

CASL Ticket # - #5200

Description:
This should make debugging a lot easier. Also added a warning that
tries to identify the case of OOM when reading in large compressed
data sets.

CASL Ticket # - 5200

Description:
Actually implemented some logic to compute the chunk size for
compressed data sets. This should reduce the excessive memory
overhead used by HDF5 when writing large compressed datasets.

Added writing of a large dataset and some performance measurements
to the testHDF5FileType

Detailed discussion on HDF5 website here:
https://support.hdfgroup.org/HDF5/doc/Advanced/Chunking/index.html

CASL Ticket # - #5200

Description:
Added two methods for querying an HDF5 or data set to get the
chunk size and whether or not it is compressed. This is to be used
in facilitating error checking of large memory overhead in the
HDF5 library for compressed data sets.

CASL Ticket # - #5200

Description:
The ultimate issue for the extra characters showing up in the state
variables was that they happened during a restart read.  The issue comes
from HDF5 reading in an empty string, where the C_NULL_CHAR is returned
if the string is empty.  Since we do not need that functionality, adding
a string find/replace after the data is read in so that C_NULL_CHAR is
replaced with ''.  Added scalar and array unit tests to check that this
change works properly.

CASL Ticket # - PHI #5241