Move aggregation operator to cli argument

Thomsch · Dec 9, 2023 · fd42341 · fd42341
1 parent 2aaa1fe
commit fd42341
Show file tree

Hide file tree

Showing 3 changed files with 50 additions and 12 deletions.
diff --git a/analysis/generate_paper.sh b/analysis/generate_paper.sh
@@ -47,7 +47,7 @@ main() {
   #
   # Untangling performance
   #
-  python src/python/main/analysis/print_median_performance.py --d4j "$D4J_SCORE_FILE" --lltc4j "$LLTC4J_SCORE_FILE" > "${PAPER_REPOSITORY}/tables/tool-performance.tex" 2> "${PAPER_REPOSITORY}/lib/tool-performance.tex"
+  python src/python/main/analysis/print_performance.py --d4j "$D4J_SCORE_FILE" --lltc4j "$LLTC4J_SCORE_FILE" --aggregator 'mean' > "${PAPER_REPOSITORY}/tables/tool-performance.tex" 2> "${PAPER_REPOSITORY}/lib/tool-performance.tex"
 
   #
   # Untangling statistics

diff --git a/...main/analysis/print_median_performance.py → ...python/main/analysis/print_performance.py b/...main/analysis/print_median_performance.py → ...python/main/analysis/print_performance.py
@@ -19,7 +19,7 @@
 PRECISION = 2
 
 
-def main(d4j_file:str, lltc4j_file:str):
+def main(d4j_file:str, lltc4j_file:str, aggregator:str):
     """
     Implementation of the script's logic. See the script's documentation for details.
 
@@ -31,18 +31,17 @@ def main(d4j_file:str, lltc4j_file:str):
     df_scores = load_dataframes(d4j_file, lltc4j_file, names=["Defects4J", "LLTC4J"])
 
     # calculate performance
-    performance_operator='median'
     df_performance = df_scores.groupby(["dataset"]).agg(
         {
-            "smartcommit_rand_index": performance_operator,
-            "flexeme_rand_index": performance_operator,
-            "filename_rand_index": performance_operator,
+            "smartcommit_rand_index": aggregator,
+            "flexeme_rand_index": aggregator,
+            "filename_rand_index": aggregator,
         }
     )
     df_performance = clean_labels(df_performance)
 
     # print performance in latex format
-    print_performance_commands(df_performance, performance_operator)
+    print_performance_commands(df_performance, aggregator)
     print_performance_table(df_performance)
 
 def print_performance_table(dataframe: pd.DataFrame):
@@ -135,5 +134,12 @@ def clean_labels(dataframe: pd.DataFrame):
         metavar="LLTC4J_SCORE_FILE",
     )
 
+    parser.add_argument(
+        "--aggregator",
+        help="The aggregator operation used to calculate the performance. e.g., median, mean",
+        required=True,
+        metavar="AGGREGATOR",
+    )
+
     args = parser.parse_args()
-    main(args.d4j, args.lltc4j)
+    main(args.d4j, args.lltc4j, args.aggregator)
diff --git a/...analysis/test_print_median_performance.py → ...n/test/analysis/test_print_performance.py b/...analysis/test_print_median_performance.py → ...n/test/analysis/test_print_performance.py
@@ -1,9 +1,9 @@
 """
-Tests for print_median_performance.py
+Tests for print_performance.py
 """
 import pytest
 
-import src.python.main.analysis.print_median_performance as print_median_performance
+import src.python.main.analysis.print_performance as print_median_performance
 
 
 @pytest.fixture
@@ -35,12 +35,11 @@ def sample_lltc4j_scores(tmpdir):
     return str(file)
 
 
-
 def test_calculate_performance(sample_d4j_scores, sample_lltc4j_scores, capfd):
     """
     Tests that the performance metrics are calculated correctly.
     """
-    print_median_performance.main(sample_d4j_scores, sample_lltc4j_scores)
+    print_median_performance.main(sample_d4j_scores, sample_lltc4j_scores, "median")
 
     captured = capfd.readouterr()
 
@@ -68,3 +67,36 @@ def test_calculate_performance(sample_d4j_scores, sample_lltc4j_scores, capfd):
 
     assert captured.out == expected_standard_output
     assert captured.err == expected_error_output
+
+def test_calculate_performance(sample_d4j_scores, sample_lltc4j_scores, capfd):
+    """
+    Tests that the performance metrics are calculated correctly.
+    """
+    print_median_performance.main(sample_d4j_scores, sample_lltc4j_scores, "mean")
+
+    captured = capfd.readouterr()
+
+    expected_standard_output = (
+        "\\begin{tabular}{lrrr}\n"
+        "\\toprule\n"
+        " & Flexeme & SmartCommit & File-based \\\\\n"
+            "Dataset &  &  &  \\\\\n"
+            "\\midrule\n"
+        "Defects4J & 0.65 & \\bfseries 0.75 & 0.55 \\\\\n"
+            "LLTC4J & 0.40 & \\bfseries 0.68 & 0.52 \\\\\n"
+            "\\bottomrule\n"
+        "\\end{tabular}\n"
+        "\n"
+    )
+
+    expected_error_output = (
+        "\\newcommand\\defectsfjFlexemeMean{0.65\\xspace}\n"
+        "\\newcommand\\defectsfjSmartcommitMean{0.75\\xspace}\n"
+        "\\newcommand\\defectsfjFilebasedMean{0.55\\xspace}\n"
+        "\\newcommand\\lltcfjFlexemeMean{0.4\\xspace}\n"
+        "\\newcommand\\lltcfjSmartcommitMean{0.68\\xspace}\n"
+        "\\newcommand\\lltcfjFilebasedMean{0.52\\xspace}\n"
+    )
+
+    assert captured.out == expected_standard_output
+    assert captured.err == expected_error_output