Merge branch 'update-import' of github.com:benedikt-schesch/AST-Mergi…

…ng-Evaluation into update-import

.github/workflows/small-test.yml

@@ -4,7 +4,7 @@ jobs:
   test:
     strategy:
       matrix:
-        maven: [ '3.9.2']
+        maven: [ '3.9.2' ]
     runs-on: ubuntu-latest
     defaults:
       run:
@@ -52,6 +52,7 @@ jobs:
         maven-version: ${{ matrix.maven }}
     - name: Clean caches & workspace
       run: make clean
+    - run: echo "LOGURU_COLORIZE=NO" >> $GITHUB_ENV
     - name: Run small test
       run: |
         git config --global user.email "[email protected]"

README.md

@@ -215,6 +215,5 @@ To run style checking run `make style`.
 ## Comparing merge algorithms
 
 To investigate differences between two mergers:
- * run `src/python/select_from_results.py` to produce a CSV file containing only the differences you are interested in.
- * run `` to create a .csv database containing only the differences.
+ * run `src/python/select_from_results.py` to produce a CSV file containing only the differences you are interested in.  See its [documentation (at top of file)](src/python/select_from_results.py) for how to run it.
  * run `src/python/replay_merge.py --merges_csv CSV_FILE --idx INDEX` (maybe add `-test`) for the index of the merge you are interested in.

src/python/replay_merge.py

@@ -52,10 +52,9 @@ def merge_replay(
 
             if (WORKDIR_DIRECTORY / workdir).exists():
                 # Ask the user if they want to delete the workdir
-                print(
-                    f"workdir {workdir} already exists. Do you want to delete it? (y/n)"
+                answer = input(
+                    f"workdir {workdir} exists. Delete it (n=reuse it)? (y/n)"
                 )
-                answer = input()
                 if answer == "y":
                     shutil.rmtree(WORKDIR_DIRECTORY / workdir)
                 else:
@@ -107,7 +106,10 @@ def merge_replay(
                 log_path,
                 repo.local_repo_path,
             ]
-            if merge_data[f"{merge_tool.name}_merge_fingerprint"] != merge_fingerprint:
+            if (
+                merge_data[f"{merge_tool.name}_merge_fingerprint"] != merge_fingerprint
+                and not arguments.dont_check_fingerprints
+            ):
                 raise Exception(
                     f"fingerprints differ: after merge of {workdir} with {merge_tool}, found"
                     + f" {merge_fingerprint} but expected "
@@ -170,7 +172,7 @@ def merge_replay(
         "--merges_csv",
         help="CSV file with merges that have been tested",
         type=str,
-        default="results/small/result.csv",
+        default="results/combined/result.csv",
     )
     parser.add_argument(
         "--idx",
@@ -183,11 +185,17 @@ def merge_replay(
         help="Test the replay of a merge",
         action="store_true",
     )
+    parser.add_argument(
+        "--dont_check_fingerprints",
+        help="Don't check the fingerprint of a merge",
+        default=False,
+        action="store_true",
+    )
     arguments = parser.parse_args()
 
     # Setup for imports
     os.system(
-        "./src/scripts/merge_tools/merging/gradlew -p src/scripts/merge_tools/merging shadowJar"
+        "./src/scripts/merge_tools/merging/gradlew -q -p src/scripts/merge_tools/merging shadowJar"
     )
     os.environ["PATH"] = os.environ["PATH"] + os.getcwd() + "/src/scripts/merge_tools/:"
     os.environ["PATH"] = (

src/python/select_from_results.py

@@ -13,6 +13,10 @@
 Here are example invocations:
   select_from_results.py '(gitmerge_ort == "Merge_failed") and (spork != "Merge_failed")'
   select_from_results.py '(gitmerge_ort == "Merge_failed") != (spork == "Merge_failed")'
+
+The resulting .csv is useful for manual examination but cannot be passed to
+`replay_merge.py` because that requires a .csv file with all tools and all
+fingerprints.
 """
 
 import argparse
@@ -62,9 +66,20 @@ def main():
 
     # Select some columns
     columns_to_select = (
-        ["idx", "repo-idx", "merge-idx", "branch_name", "merge", "left", "right"]
+        [
+            "idx",
+            "repo-idx",
+            "merge-idx",
+            "branch_name",
+            "merge",
+            "left",
+            "left_tree_fingerprint",
+            "right",
+            "right_tree_fingerprint",
+        ]
         + columns_in_query(args.query)
         + args.columns
+        + ["repository"]
     )
     df = df[columns_to_select]