[rocprofiler-compute] Add database output format to analyze mode (#748)

Analysis data dump

* Add `--output-format` and `--output-name` option to analyze mode

* Remove `--output` and `-save-dfs` option to analyze mode

* Add documentation on `rocpd` output format and analysis database file

* Create sqlite3 database using object relation mapping (ORM) provided
  by sqlalchemy library

* Fix metrics config to remove metrics marked as `null`, fix `Unit` header, add
  missing `title`

* Add test cases to ensure analysis data dump work

projects/rocprofiler-compute/tests/test_TCP_counters.py

@@ -147,7 +147,8 @@ def test_L1_cache_counters(
     base = Path(test_utils.get_output_dir())
 
     for app_name in app_names:
-        workload_dir = str(base / app_name)
+        workload_dir = f"{base}/{app_name}"
+        workload_dir_output = f"{base}_{app_name}"
 
         # 1. profile the app
         return_code = binary_handler_profile_rocprof_compute(
@@ -167,15 +168,17 @@ def test_L1_cache_counters(
             workload_dir,
             "-b",
             "16.3",
-            "--save-dfs",
-            workload_dir,
+            "--output-format",
+            "csv",
+            "--output-name",
+            workload_dir_output,
         ])
         assert return_code == 0
 
         # 3. save results in local
 
         # FIXME: customize file name to avoid hardcode
-        csv_path = workload_dir + "/16.3_vL1D_cache_access_metrics.csv"
+        csv_path = workload_dir_output + "/16.3_vL1D_cache_access_metrics.csv"
         data = load_metrics(csv_path)
 
         for metric in metrics:
@@ -185,6 +188,7 @@ def test_L1_cache_counters(
 
         # 4. clean local output
         test_utils.clean_output_dir(config["cleanup"], workload_dir)
+        test_utils.clean_output_dir(config["cleanup"], workload_dir_output)
     test_utils.clean_output_dir(config["cleanup"], base)
 
     # 5. check results are expected

projects/rocprofiler-compute/tests/test_analyze_commands.py

@@ -25,6 +25,7 @@
 
 import os
 import shutil
+from pathlib import Path
 from unittest.mock import Mock
 
 import pandas as pd
@@ -608,14 +609,16 @@ def test_decimal_3(binary_handler_analyze_rocprof_compute):
 
 @pytest.mark.misc
 def test_save_dfs(binary_handler_analyze_rocprof_compute):
-    output_path = "tests/workloads/vcopy/saved_analysis"
+    output_path = test_utils.get_output_dir()
     for dir in indirs:
         workload_dir = test_utils.setup_workload_dir(dir)
         code = binary_handler_analyze_rocprof_compute([
             "analyze",
             "--path",
             workload_dir,
-            "--save-dfs",
+            "--output-format",
+            "csv",
+            "--output-name",
             output_path,
         ])
         assert code == 0
@@ -627,6 +630,7 @@ def test_save_dfs(binary_handler_analyze_rocprof_compute):
 
         shutil.rmtree(output_path)
     test_utils.clean_output_dir(config["cleanup"], workload_dir)
+    test_utils.clean_output_dir(config["cleanup"], output_path)
 
 
 @pytest.mark.col
@@ -860,7 +864,6 @@ def test_dependency_MI100(binary_handler_analyze_rocprof_compute):
 def test_parser_utility_functions():
     """Test parser utility functions edge cases"""
     import sys
-    from pathlib import Path
 
     sys.path.insert(0, str(Path(__file__).parent.parent / "src"))
 
@@ -969,7 +972,6 @@ def test_parser_utility_functions():
 def test_parser_error_handling():
     """Test parser error handling paths"""
     import sys
-    from pathlib import Path
 
     sys.path.insert(0, str(Path(__file__).parent.parent / "src"))
 
@@ -1009,7 +1011,6 @@ def test_missing_file_handling(binary_handler_analyze_rocprof_compute):
 def test_ast_transformer_edge_cases():
     """Simplified test focusing on the actual code paths"""
     import sys
-    from pathlib import Path
 
     sys.path.insert(0, str(Path(__file__).parent.parent / "src"))
 
@@ -1051,7 +1052,6 @@ def test_ast_transformer_edge_cases():
 def test_analyze_with_debug_mode(binary_handler_analyze_rocprof_compute):
     """Test analyze to cover debug paths in eval_metric - using direct function call"""
     import sys
-    from pathlib import Path
 
     sys.path.insert(0, str(Path(__file__).parent.parent / "src"))
 
@@ -1138,7 +1138,6 @@ def test_filter_combinations_coverage(binary_handler_analyze_rocprof_compute):
 def test_apply_filters_direct():
     """Test apply_filters function directly to cover filter branches"""
     import sys
-    from pathlib import Path
 
     sys.path.insert(0, str(Path(__file__).parent.parent / "src"))
 
@@ -1213,7 +1212,6 @@ def test_missing_files_scenarios(binary_handler_analyze_rocprof_compute):
 def test_pc_sampling_basic_coverage():
     """Test PC sampling functions with minimal data"""
     import sys
-    from pathlib import Path
 
     sys.path.insert(0, str(Path(__file__).parent.parent / "src"))
 
@@ -1245,7 +1243,6 @@ def test_pc_sampling_basic_coverage():
 def test_build_dfs_edge_cases():
     """Test build_dfs and gen_counter_list with various configurations"""
     import sys
-    from pathlib import Path
 
     sys.path.insert(0, str(Path(__file__).parent.parent / "src"))
 
@@ -1275,7 +1272,6 @@ def test_build_dfs_edge_cases():
 def test_update_functions_coverage():
     """Test update_denom_string and update_normUnit_string branches"""
     import sys
-    from pathlib import Path
 
     sys.path.insert(0, str(Path(__file__).parent.parent / "src"))
 

projects/rocprofiler-compute/tests/test_profile_general.py

@@ -766,6 +766,32 @@ def test_roof_rocpd(binary_handler_profile_rocprof_compute):
     test_utils.clean_output_dir(config["cleanup"], workload_dir)
 
 
+@pytest.mark.misc
+def test_analyze_rocpd(
+    binary_handler_profile_rocprof_compute, binary_handler_analyze_rocprof_compute
+):
+    workload_dir = test_utils.get_output_dir()
+    options = ["--device", "0", "--format-rocprof-output", "rocpd"]
+    binary_handler_profile_rocprof_compute(config, workload_dir, options, roof=True)
+
+    db_name = "test"
+    code = binary_handler_analyze_rocprof_compute([
+        "analyze",
+        "--output-format",
+        "db",
+        "--output-name",
+        f"{db_name}",
+        "--path",
+        workload_dir,
+    ])
+    assert code == 0
+    assert os.path.isfile(f"{db_name}.db")
+
+    # Remove test.db
+    os.remove(f"{db_name}.db")
+    test_utils.clean_output_dir(config["cleanup"], workload_dir)
+
+
 @pytest.mark.misc
 def test_roofline_workload_dir_not_set_error():
     """