Create two separate files

.github/workflows/check-reproducibility.yml

@@ -47,7 +47,7 @@ jobs:
       uses: jlumbroso/free-disk-space@main
     - name: make check-merges-reproducibility
       run: |
-        head -n 151 results/combined/result.csv > temp.csv && mv temp.csv results/combined/result.csv
+        head -n 151 results/combined/result_raw.csv > temp.csv && mv temp.csv results/combined/result_raw.csv
         make check-merges-reproducibility
       env:
           GITHUB_TOKEN: ${{ secrets.TOKEN_GITHUB }}

Makefile

@@ -8,9 +8,9 @@ SH_SCRIPTS   = $(shell grep --exclude-dir=build --exclude-dir=repos --exclude-di
 BASH_SCRIPTS = $(shell grep --exclude-dir=build --exclude-dir=repos --exclude-dir=cache -r -l '^\#! \?\(/bin/\|/usr/bin/env \)bash' * | grep -v /.git/ | grep -v '~$$' | grep -v '\.tar$$' | grep -v gradlew)
 PYTHON_FILES = $(shell find .  -name '*.py' ! -path './repos/*' -not -path "./.workdir/*" -not -path "./cache*/*" | grep -v '/__pycache__/' | grep -v '/.git/' | grep -v gradlew | grep -v git-hires-merge)
 
-CSV_RESULTS_COMBINED = results/combined/result.csv
-CSV_RESULTS_GREATEST_HITS = results/greatest_hits/result.csv
-CSV_RESULTS_REAPER = results/reaper/result.csv
+CSV_RESULTS_COMBINED = results/combined/result_raw.csv
+CSV_RESULTS_GREATEST_HITS = results/greatest_hits/result_raw.csv
+CSV_RESULTS_REAPER = results/reaper/result_raw.csv
 CSV_RESULTS = $(CSV_RESULTS_COMBINED)
 
 NUM_PROCESSES = 0

README.md

@@ -71,15 +71,15 @@ make small-test
 This runs the entire code on two small repos.
 The output data appears in `results/small/`.
 
-* `results/small/result.csv`: the final result
+* `results/small/result_adjusted.csv`: the final result
 
 * `results/small/merges/` contains all the merges.
 
 * `results/small/merges_compared/` contains all merges and indicates whether the merge results are different and thus need to be analyzed.
 
 * `results/small/merges_tested/` contains all merges that have been tested.
 
-* `results/small/result.csv` contains the final result.
+* `results/small/result_adjusted.csv` contains the final result.
 
 ### Perform full analysis
 
@@ -97,7 +97,7 @@ To run the stack on all repos and also diff the merges' outputs:
 
 This will run the entire code on all the repos and automatically decompress the cache if `cache/` does not exist.
 All the output data can be found in `results/`.
-The final result is found in `results/result.csv`.
+The final result is found in `results/result_adjusted.csv`.
 Directory `results/merges` contains all the merges for each repo.
 Directory `results/merges_tested` contains all the merges that have been tested.
 
@@ -248,7 +248,7 @@ To investigate differences between two mergers:
 * Run `src/python/utils/select_from_results.py` to create a .csv database containing only the differences.
 * Set `DELETE_WORKDIRS` to `false` in `src/python/variables.py`.
 * Run `src/python/replay_merge.py --idx INDEX` (maybe add `-test`) for the index of the merge you are interested in.
-  If the merge is in the small test, you may need to add `--merges_csv ./test/small-goal-files/result.csv`.
+  If the merge is in the small test, you may need to add `--merges_csv ./test/small-goal-files/result_adjusted.csv`.
 
 ## Overwriting results manually
 

src/python/delete_cache_entry_of_merge_tests.py

@@ -38,7 +38,7 @@ def delete_row(row):
     parser.add_argument(
         "--result",
         type=str,
-        default="results/combined/result.csv",
+        default="results/combined/result_adjusted.csv",
         help="The result csv file.",
     )
     parser.add_argument(

src/python/delete_tests_with_different_output.py

@@ -14,7 +14,7 @@ def main():
     parser.add_argument(
         "--result",
         type=str,
-        default="results/combined/result.csv",
+        default="results/combined/result_adjusted.csv",
         help="The result csv file.",
     )
     rows_affected = 0

src/python/latex_output.py

@@ -495,9 +495,9 @@ def main():
 
     print(f"CSV saved to: {csv_filename}")
     print(f"Rows: {len(filtered_df)}")
+
+    result_df.to_csv(args.output_dir / "result_raw.csv", index_label="idx")
 
-    for merge_tool in MERGE_TOOL:
-        result_df[merge_tool.name + "_raw"] = result_df[merge_tool.name]
     for idx, row in result_df.iterrows():
         for merge_tool in MERGE_TOOL:
             if "plus" in merge_tool.name:
@@ -507,8 +507,7 @@ def main():
             if result1.name == MERGE_STATE.Merge_failed.name \
                 and result2.name == TEST_STATE.Tests_failed.name:
                     result_df.loc[idx, merge_tool.name] = TEST_STATE.Tests_failed.name
-
-    result_df.to_csv(args.output_dir / "result.csv", index_label="idx")
+    result_df.to_csv(args.output_dir / "result_adjusted.csv", index_label="idx")
 
     main_df = result_df[result_df["branch_name"].isin(main_branch_names)]
     feature = result_df[~result_df["branch_name"].isin(main_branch_names)]

src/python/replay_merge.py

@@ -405,7 +405,7 @@ def merge_replay(
         "--merges_csv",
         help="CSV file with merges that have been tested",
         type=str,
-        default="results/combined/result.csv",
+        default="results/combined/result_adjusted.csv",
     )
     parser.add_argument(
         "--idx",

src/python/utils/build_inconsistent_merges.py

@@ -63,7 +63,7 @@ def main():
         "--result_csv",
         type=str,
         help="Path to the result CSV file",
-        default="results/combined/result.csv",
+        default="results/combined/result_adjusted.csv",
     )
     parser.add_argument(
         "--output_csv",

src/python/utils/select_from_results.py

@@ -52,7 +52,7 @@ def main():
     parser.add_argument(
         "--input",
         action="store",
-        default="results/combined/result.csv",
+        default="results/combined/result_adjusted.csv",
     )
     parser.add_argument("columns", nargs=argparse.REMAINDER)
     args = parser.parse_args()