generate module overview table for DAOS

also, cleanup file/object terminology in job summary
darshan-hpc · Nov 12, 2024 · e44ea71 · e44ea71
1 parent 24b1891
commit e44ea71
Show file tree

Hide file tree

Showing 3 changed files with 15 additions and 10 deletions.
diff --git a/darshan-util/pydarshan/darshan/backend/cffi_backend.py b/darshan-util/pydarshan/darshan/backend/cffi_backend.py
@@ -70,6 +70,8 @@
     "APXC",
     "APMPI",
     "HEATMAP",
+    "DFS",
+    "DAOS",
 ]
 def mod_name_to_idx(mod_name):
     return _mod_names.index(mod_name)

diff --git a/darshan-util/pydarshan/darshan/cli/summary.py b/darshan-util/pydarshan/darshan/cli/summary.py
@@ -508,7 +508,7 @@ def register_figures(self):
                 sect_title = f"Per-Module Statistics: {mod}"
 
             try:
-                if mod in ["POSIX", "MPI-IO", "STDIO"]:
+                if mod in ["POSIX", "MPI-IO", "STDIO", "DFS", "DAOS"]:
                     # get the module's record dataframe and then pass to
                     # Darshan accumulator interface to generate a cumulative
                     # record and derived metrics
@@ -526,9 +526,10 @@ def register_figures(self):
                             fig_grid_area="overview")
                     self.figures.append(mod_overview_fig)
 
+                    data_type = "File" if mod != "DAOS" else "Object"
                     file_count_summary_fig = ReportFigure(
                             section_title=sect_title,
-                            fig_title=f"File Count Summary <br> (estimated by {mod} I/O access offsets)",
+                            fig_title=f"{data_type} Count Summary <br> (estimated by {mod} I/O access offsets)",
                             fig_func=log_file_count_summary_table,
                             fig_args=dict(derived_metrics=acc.derived_metrics,
                                           mod_name=mod),

diff --git a/darshan-util/pydarshan/darshan/lib/accum.py b/darshan-util/pydarshan/darshan/lib/accum.py
@@ -10,18 +10,19 @@
 
 def log_file_count_summary_table(derived_metrics,
                                  mod_name: str):
+    data_type = "files" if mod_name != "DAOS" else "objects"
     # the darshan_file_category enum is not really
     # exposed in CFFI/Python layer, so we effectively
     # re-export the content indices we need here
     # so that we can properly index the C-level data
-    darshan_file_category = {"total files":0,
-                             "read-only files":1,
-                             "write-only files":2,
-                             "read/write files":3}
+    darshan_file_category = {f"total {data_type}":0,
+                             f"read-only {data_type}":1,
+                             f"write-only {data_type}":2,
+                             f"read/write {data_type}":3}
     df = pd.DataFrame.from_dict(darshan_file_category, orient="index")
     df.rename(columns={0:"index"}, inplace=True)
     df.index.rename('type', inplace=True)
-    df["number of files"] = np.zeros(4, dtype=int)
+    df[f"number of {data_type}"] = np.zeros(4, dtype=int)
     df["avg. size"] = np.zeros(4, dtype=str)
     df["max size"] = np.zeros(4, dtype=str)
 
@@ -59,9 +60,10 @@ def log_module_overview_table(derived_metrics,
     mod_overview = []
     total_cat = derived_metrics.category_counters[0]
 
-    total_files = total_cat.count
-    indices = ["files accessed", "bytes read", "bytes written", "I/O performance estimate"]
-    mod_overview.append(f"{total_files}")
+    total_count = total_cat.count
+    data_type = "files" if mod_name != "DAOS" else "objects"
+    indices = [f"{data_type} accessed", "bytes read", "bytes written", "I/O performance estimate"]
+    mod_overview.append(f"{total_count}")
     total_bytes_read = total_cat.total_read_volume_bytes
     total_bytes_read_str = humanize.naturalsize(total_bytes_read, binary=True, format="%.2f")
     total_bytes_written = total_cat.total_write_volume_bytes