Allow registry fields with initial_value to not error on not found in…

… input data (#365)

Tag name (required for release branches):
Originator(s): @jimmielin (with extensive guidance from @peverwhee)

Description (include the issue title, and the keyword ['closes',
'fixes', 'resolves'] followed by the issue number):
* Fixes #359 

Should enable CAM7 / CAM4 regression tests to run again despite no PBL
scheme, if `pbuf_pblh` includes
`<initial_value>0.0_kind_phys</initial_value>` in registry.xml -
separate PR will include this.

Describe any changes made to build system: 
* Adds `vars_init_value` in generation of physics_inputs.F90 to record
which variables have an `<initial_value>` set in `registry.xml`

Describe any changes made to the namelist: N/A

List any changes to the defaults for the input datasets (e.g. boundary
datasets): N/A

List all files eliminated and why: N/A

List all files added and what they do:
```
A       test/unit/python/sample_files/write_init_files/phys_vars_init_check_initial_value.F90
A       test/unit/python/sample_files/write_init_files/physics_inputs_initial_value.F90
A       test/unit/python/sample_files/write_init_files/simple_reg_initial_value.xml
  - new test sample files for write_init_files
```

List all existing files that have been modified, and describe the
changes:
(Helpful git command: `git diff --name-status
development...<your_branch_name>`)
```
M       cime_config/cam_autogen.py
M       cime_config/cam_build_cache.py
M       cime_config/cam_config.py
M       src/data/generate_registry_data.py
  - pass variables having initial_value element in registry.xml

M       src/data/write_init_files.py
  - add error_on_not_found=.false. to variables having initial_value element
  - or write info message for variables that cannot be read (i.e., no horizontal dimension)

M       test/unit/python/sample_files/build_cache_files/update_reg_build_cache.xml
M       test/unit/python/test_build_cache.py
M       test/unit/python/test_cam_autogen.py
M       test/unit/python/test_registry.py
M       test/unit/python/test_write_init_files.py
  - fix compatibility of existing tests with new vars_init_value argument
  - add test_simple_initial_value_write_init unit test
```


If there are new failures (compared to the
`test/existing-test-failures.txt` file),
have them OK'd by the gatekeeper, note them here, and add them to the
file.
If there are baseline differences, include the test and the reason for
the
diff. What is the nature of the change? Roundoff?

derecho/intel/aux_sima:

derecho/gnu/aux_sima:

If this changes climate describe any run(s) done to evaluate the new
climate in enough detail that it(they) could be reproduced:

CAM-SIMA date used for the baseline comparison tests if different than
latest:

cime_config/cam_autogen.py

@@ -392,7 +392,7 @@ def generate_registry(data_search, build_cache, atm_root, bldroot,
                                    gen_fort_indent, source_mods_dir, atm_root,
                                    logger=_LOGGER, schema_paths=data_search,
                                    error_on_no_validate=True)
-            retcode, reg_file_list, ic_names, registry_constituents = retvals
+            retcode, reg_file_list, ic_names, registry_constituents, vars_init_value = retvals
             # Raise error if gen_registry failed:
             if retcode != 0:
                 emsg = "ERROR:Unable to generate CAM data structures from {}, err = {}"
@@ -406,15 +406,16 @@ def generate_registry(data_search, build_cache, atm_root, bldroot,
         # Save build details in the build cache
         reg_file_paths = [x.file_path for x in reg_file_list if x.file_path]
         build_cache.update_registry(gen_reg_file, registry_files, dycore,
-                                    reg_file_paths, ic_names, registry_constituents)
+                                    reg_file_paths, ic_names, registry_constituents, vars_init_value)
     else:
         # If we did not run the registry generator, retrieve info from cache
         reg_file_paths = build_cache.reg_file_list()
         ic_names = build_cache.ic_names()
         registry_constituents = build_cache.constituents()
+        vars_init_value = build_cache.vars_init_value()
     # End if
 
-    return genreg_dir, do_gen_registry, reg_file_paths, ic_names, registry_constituents
+    return genreg_dir, do_gen_registry, reg_file_paths, ic_names, registry_constituents, vars_init_value
 
 ###############################################################################
 def generate_physics_suites(build_cache, preproc_defs, host_name,
@@ -659,7 +660,8 @@ def generate_physics_suites(build_cache, preproc_defs, host_name,
 ###############################################################################
 def generate_init_routines(build_cache, bldroot, force_ccpp, force_init,
                            source_mods_dir, gen_fort_indent,
-                           cap_database, ic_names, registry_constituents):
+                           cap_database, ic_names, registry_constituents,
+                           vars_init_value):
 ###############################################################################
     """
     Generate the host model initialization source code files
@@ -697,7 +699,7 @@ def generate_init_routines(build_cache, bldroot, force_ccpp, force_init,
         #   within write_init_files (so that write_init_files can be the place
         #   where the source include files are stored).
         source_paths = [source_mods_dir, _REG_GEN_DIR]
-        retmsg = write_init_files(cap_database, ic_names, registry_constituents,
+        retmsg = write_init_files(cap_database, ic_names, registry_constituents, vars_init_value,
                                   init_dir, _find_file, source_paths,
                                   gen_fort_indent, _LOGGER)
 

cime_config/cam_build_cache.py

@@ -227,6 +227,7 @@ def __init__(self, build_cache):
         self.__reg_gen_files = []
         self.__ic_names = {}
         self.__constituents = []
+        self.__vars_init_value = []
         if os.path.exists(build_cache):
             # Initialize build cache state
             _, cache = read_xml_file(build_cache)
@@ -257,6 +258,9 @@ def __init__(self, build_cache):
                             stdname = item.get('standard_name')
                             itext = clean_xml_text(item)
                             self.__constituents.append(itext)
+                        elif item.tag == 'vars_init_value_entry':
+                            itext = clean_xml_text(item)
+                            self.__vars_init_value.append(itext)
                         else:
                             emsg = "ERROR: Unknown registry tag, '{}'"
                             raise ValueError(emsg.format(item.tag))
@@ -318,7 +322,7 @@ def __init__(self, build_cache):
         # end if
 
     def update_registry(self, gen_reg_file, registry_source_files,
-                        dycore, reg_file_list, ic_names, constituents):
+                        dycore, reg_file_list, ic_names, constituents, vars_init_value):
         """Replace the registry cache data with input data
         """
         self.__dycore = dycore
@@ -334,6 +338,7 @@ def update_registry(self, gen_reg_file, registry_source_files,
         # and should already be of type dict:
         self.__ic_names = ic_names
         self.__constituents = constituents
+        self.__vars_init_value = vars_init_value
 
     def update_ccpp(self, suite_definition_files, scheme_files, host_files,
                     xml_files, namelist_meta_files, namelist_groups,
@@ -410,6 +415,9 @@ def write(self):
             const_entry = ET.SubElement(registry, 'constituent_entry')
             const_entry.text = stdname
         # end for
+        for stdname in self.__vars_init_value:
+            var_entry = ET.SubElement(registry, 'vars_init_value_entry')
+            var_entry.text = stdname
         # CCPP
         ccpp = ET.SubElement(new_cache, 'CCPP')
         for sfile in self.__sdfs.values():
@@ -617,5 +625,9 @@ def constituents(self):
         """Return a copy of the registry constituents list"""
         return list(self.__constituents)
 
+    def vars_init_value(self):
+        """Return a copy of the list of variables with initial_value"""
+        return list(self.__vars_init_value)
+
 #############
 # End of file

cime_config/cam_config.py

@@ -854,7 +854,7 @@ def generate_cam_src(self, gen_fort_indent):
         retvals = generate_registry(data_search, build_cache, self.__atm_root,
                                     self.__bldroot, source_mods_dir,
                                     dyn, gen_fort_indent)
-        reg_dir, force_ccpp, reg_files, ic_names, registry_constituents = retvals
+        reg_dir, force_ccpp, reg_files, ic_names, registry_constituents, vars_init_value = retvals
 
         #Add registry path to config object:
         reg_dir_desc = "Location of auto-generated registry code."
@@ -887,7 +887,8 @@ def generate_cam_src(self, gen_fort_indent):
         init_dir = generate_init_routines(build_cache, self.__bldroot,
                                           force_ccpp, force_init,
                                           source_mods_dir, gen_fort_indent,
-                                          capgen_db, ic_names, registry_constituents)
+                                          capgen_db, ic_names, registry_constituents,
+                                          vars_init_value)
 
         #Add registry path to config object:
         init_dir_desc = "Location of auto-generated physics initialization code."

src/data/generate_registry_data.py

@@ -1758,6 +1758,30 @@ def _create_constituent_list(registry):
     # end for
     return constituent_list
 
+###############################################################################
+def _create_variables_with_initial_value_list(registry):
+###############################################################################
+    """
+    Create a list of all variables with initial_value defined in the registry.
+    To be used by write_init_files.py to allow these variables to
+    not error when not found in the initial conditions file.
+    """
+    vars_init_value_list = []
+    for section in registry:
+        if section.tag == 'file':
+            for obj in section:
+                if obj.tag == 'variable':
+                    for subobj in obj:
+                        if subobj.tag == 'initial_value':
+                            stdname = obj.get('standard_name')
+                            vars_init_value_list.append(stdname)
+                        # end if (only if initial_value node is found)
+                    # end for
+                # end if (ignore other node types)
+            # end for
+        # end if (ignore other node types)
+    # end for
+    return vars_init_value_list
 
 ###############################################################################
 def gen_registry(registry_file, dycore, outdir, indent,
@@ -1767,7 +1791,7 @@ def gen_registry(registry_file, dycore, outdir, indent,
     """Parse a registry XML file and generate source code and metadata.
     <dycore> is the name of the dycore for DP coupling specialization.
     <config> is a dictionary containing other configuration items for
-       souce code customization.
+       source code customization.
     Source code and metadata is output to <outdir>.
     <src_mod> is the location of the builds SourceMods/src.cam directory
     <src_root> is the top of the component tree
@@ -1819,7 +1843,8 @@ def gen_registry(registry_file, dycore, outdir, indent,
         retcode = 1
         files = None
         ic_names = None
-        constituents = None
+        registry_constituents = None
+        vars_init_value = None
     else:
         library_name = registry.get('name')
         emsg = f"Parsing registry, {library_name}"
@@ -1830,9 +1855,10 @@ def gen_registry(registry_file, dycore, outdir, indent,
         # See comment in _create_ic_name_dict
         ic_names = _create_ic_name_dict(registry)
         registry_constituents = _create_constituent_list(registry)
+        vars_init_value = _create_variables_with_initial_value_list(registry)
         retcode = 0 # Throw exception on error
     # end if
-    return retcode, files, ic_names, registry_constituents
+    return retcode, files, ic_names, registry_constituents, vars_init_value
 
 def main():
     """Function to execute when module called as a script"""

src/data/write_init_files.py

@@ -47,8 +47,8 @@
 #Main function
 ##############
 
-def write_init_files(cap_database, ic_names, registry_constituents, outdir,
-                     file_find_func, source_paths, indent, logger,
+def write_init_files(cap_database, ic_names, registry_constituents, vars_init_value,
+                     outdir, file_find_func, source_paths, indent, logger,
                      phys_check_filename=None, phys_input_filename=None):
 
     """
@@ -239,7 +239,8 @@ def write_init_files(cap_database, ic_names, registry_constituents, outdir,
 
         # Write physics_read_data subroutine:
         write_phys_read_subroutine(outfile, host_dict, host_vars, host_imports,
-                                   phys_check_fname_str, constituent_set)
+                                   phys_check_fname_str, constituent_set,
+                                   vars_init_value)
 
         outfile.blank_line()
 
@@ -810,7 +811,8 @@ def get_dimension_info(hvar):
     return vdim_name, legal_dims, fail_reason
 
 def write_phys_read_subroutine(outfile, host_dict, host_vars, host_imports,
-                               phys_check_fname_str, constituent_set):
+                               phys_check_fname_str, constituent_set,
+                               vars_init_value):
 
     """
     Write the "physics_read_data" subroutine, which
@@ -866,10 +868,20 @@ def write_phys_read_subroutine(outfile, host_dict, host_vars, host_imports,
             if levnm is not None:
                 call_str += f"'{levnm}', "
             # end if
-            call_str += f"timestep, {var_locname})"
+            err_on_not_found_string = ""
+            if var_stdname in vars_init_value:
+                # if initial value is available, do not throw error when not found in initial condition file.
+                err_on_not_found_string = ", error_on_not_found=.false."
+            # end if
+            call_str += f"timestep, {var_locname}{err_on_not_found_string})"
         else:
-            call_str = f"call endrun('Cannot read {var_locname} from file'" + \
+            # if initial value is assigned, then it can be ignored
+            if var_stdname in vars_init_value:
+                call_str = f"if(masterproc) write(iulog,*) '{var_locname} already has an initial_value, using it now. It also cannot be read from file: {reason}'"
+            else:
+                call_str = f"call endrun('Cannot read {var_locname} from file'" + \
                 f"//', {reason}')"
+            # end if
         # end if
 
         # Add string to dictionary:

test/unit/python/sample_files/build_cache_files/update_reg_build_cache.xml

@@ -10,6 +10,7 @@
       <ic_name_entry standard_name="Only_had_a">brain</ic_name_entry>
       <constituent_entry>cnst_1</constituent_entry>
       <constituent_entry>cnst_2</constituent_entry>
+      <vars_init_value_entry>Only_had_a_brain</vars_init_value_entry>
    </registry>
    <CCPP>
       <SDF file_path="/behind/curtain/suite_wizard.xml" hash="ade45dea32a6f279957735596a147efddd061f9c"/>