feature: importing HPCToolkit data into Hatchet (#126)

* feature: importing HPCToolkit data into Hatchet * update: issues addressed in the pr fixed * fix: treating PMPI/MPI differently * remove: graphframe test file * remove: hpctoolkit test files * tests: unit tests for importing HPCToolkit data into Hatchet * update: storing depth as Node attribute * fix: update tests * update: license comments inside test filed * delete: remove src directory from the test performance db * update: gitignore file * update: formating + more tests * update: mapping inclusive/exlusive metrics and time metric * format: graphframe * remove: __init__ file inside tests dir --------- Co-authored-by: Dragana Grbic <grbic1@quartz2300.llnl.gov> Co-authored-by: Dragana Grbic <grbic1@quartz770.llnl.gov> Co-authored-by: Dragana Grbic <grbic1@quartz1538.llnl.gov> Co-authored-by: Dragana Grbic <grbic1@quartz764.llnl.gov>
LLNL · May 30, 2024 · fc47f96 · fc47f96
1 parent 9a4e09f
commit fc47f96
Show file tree

Hide file tree

Showing 10 changed files with 3,286 additions and 11 deletions.
diff --git a/hatchet/graphframe.py b/hatchet/graphframe.py
@@ -4,29 +4,28 @@
 # SPDX-License-Identifier: MIT
 
 import copy
+import json
 import sys
 import traceback
-
 from collections import defaultdict
 
-import pandas as pd
-import numpy as np
 import multiprocess as mp
-import json
+import numpy as np
+import pandas as pd
 
-from .node import Node
-from .graph import Graph
+from .external.console import ConsoleRenderer
 from .frame import Frame
+from .graph import Graph
+from .node import Node
 from .query import (
-    is_hatchet_query,
+    AbstractQuery,
     ObjectQuery,
-    parse_string_dialect,
     QueryEngine,
-    AbstractQuery,
+    is_hatchet_query,
+    parse_string_dialect,
 )
-from .external.console import ConsoleRenderer
-from .util.dot import trees_to_dot
 from .util.deprecated import deprecated_params
+from .util.dot import trees_to_dot
 
 try:
     from .cython_modules.libs import graphframe_modules as _gfm_cy
@@ -110,6 +109,35 @@ def from_hpctoolkit(dirname):
 
         return HPCToolkitReader(dirname).read()
 
+    @staticmethod
+    def from_hpctoolkit_latest(
+        dirname: str,
+        max_depth: int = None,
+        min_percentage_of_application_time: int = None,
+        min_percentage_of_parent_time: int = None,
+    ):
+        """
+        Read an HPCToolkit database directory into a new GraphFrame
+
+        Arguments:
+            dirname (str): directory of an HPCToolkit performance database
+            max_depth (int): maximum depth that nodes in the CCT can have to be imported in Hatchet
+            min_percentage_of_application_time (int): minimum percentage of application time that nodes in the CCT must have to be imported in Hatchet
+            min_percentage_of_parent_time (int): minimum percentage of parent time that nodes in the CCT must have to be imported in Hatchet
+
+        Returns:
+            (GraphFrame): new GraphFrame containing HPCToolkit profile data
+        """
+        # import this lazily to avoid circular dependencies
+        from .readers.hpctoolkit_reader_latest import HPCToolkitReaderLatest
+
+        return HPCToolkitReaderLatest(
+            dirname,
+            max_depth=max_depth,
+            min_application_percentage_time=min_percentage_of_application_time,
+            min_parent_percentage_time=min_percentage_of_parent_time,
+        ).read()
+
     @staticmethod
     def from_caliper(filename_or_stream, query=None):
         """Read in a Caliper .cali or .json file.