initial run bug run

Author: cadddr

.gitmodules

@@ -13,3 +13,6 @@
 [submodule "cache"]
 	path = cache
 	url = https://github.com/ASSERT-KTH/elle-elle-aime-cache.git
+[submodule "benchmarks/run_bug_run"]
+	path = benchmarks/run_bug_run
+	url = https://github.com/giganticode/run_bug_run.git

benchmarks/run_bug_run

@@ -0,0 +1,56 @@
+from pathlib import Path
+from unidiff import PatchSet
+from elleelleaime.core.benchmarks.benchmark import Benchmark
+from elleelleaime.core.benchmarks.runbugrun.runbugrunbug import RunBugRunBug
+
+import subprocess
+import logging
+
+import pandas as pd
+
+class RunBugRun(Benchmark):
+    """
+    The class for representing the RunBugRun benchmark.
+    """
+
+    def __init__(
+        self, path: Path = Path("benchmarks/run_bug_run").absolute()
+    ) -> None:
+        super().__init__("runbugrun", path)
+        
+    def initialize(self) -> None:
+        """
+        Initializes the RunBugRun benchmark object by collecting all the bugs.
+        """
+        logging.info("Initializing RunBugRun benchmark...")
+        
+        python_path = Path(self.get_path(), 'python_valid0.jsonl')
+        # test_path = Path(self.get_path(), 'tests_all.jsonl')
+        
+        python_df = pd.read_json(python_path, lines=True).set_index('problem_id')
+        
+        for prob_id, (buggy_submission_id, buggy_code, fixed_submission_id, fixed_code) \
+        in python_df.drop_duplicates(subset=['buggy_submission_id'])[
+            ['buggy_submission_id','buggy_code', 'fixed_submission_id', 'fixed_code']
+        ].iterrows():
+            
+            buggy_file = Path(self.path, f'{prob_id}_{buggy_submission_id}.py')
+            fixed_file = Path(self.path, f'{prob_id}_{fixed_submission_id}.py')
+
+            run = subprocess.run(
+                f"""cd {self.get_path()} && 
+                echo '''{buggy_code}''' > {buggy_file} && 
+                echo '''{fixed_code}''' > {fixed_file} && 
+                diff --unified {fixed_file.relative_to(self.path)} {buggy_file.relative_to(self.path)}""",
+                shell=True,
+                capture_output=True
+            )
+            if run.returncode:
+                print (run)
+
+            diff = PatchSet(run.stdout.decode("utf-8"))
+            # Change the source file path to point to the buggy version
+            diff[0].source_file = f"{buggy_file.relative_to(self.path)}"
+    
+            self.add_bug(RunBugRunBug(self, f"{prob_id}_{buggy_submission_id}", str(diff)))
+        

elleelleaime/core/benchmarks/runbugrun/__init__.py

@@ -0,0 +1,25 @@
+import subprocess
+import shutil
+import os
+from elleelleaime.core.benchmarks.benchmark import Benchmark
+
+from elleelleaime.core.benchmarks.bug import Bug
+from elleelleaime.core.benchmarks.test_result import TestResult
+from elleelleaime.core.benchmarks.compile_result import CompileResult
+
+class RunBugRunBug(Bug):
+    """
+    The class for representing RunBugRun bugs
+    """
+
+    def __init__(self, benchmark: Benchmark, bid: str, ground_truth: str) -> None:
+        super().__init__(benchmark, bid, ground_truth, True)
+        
+    def checkout(self, path: str, fixed: bool = False) -> bool:
+        pass
+
+    def compile(self, path: str) -> CompileResult:
+        pass
+
+    def test(self, path: str) -> TestResult:
+        pass

elleelleaime/core/benchmarks/runbugrun/runbugrun.py

@@ -1,24 +1,41 @@
 #!/bin/bash
 
 ### Submodules
-git submodule init;
-git submodule update;
-
-### Java and Maven images
-docker pull openjdk:11;
-docker pull maven:3.9.8-eclipse-temurin-8;
-
-### Defects4J image
-cd benchmarks/defects4j;
-cpanm --installdeps .;
-./init.sh;
-cd ../..;
-
-### GitBug-Java
-cd benchmarks/gitbug-java;
-chmod +x gitbug-java;
-poetry install --no-root;
-# Skip setup if in CI
-if [ -z "$CI" ]; then
- poetry run ./gitbug-java setup;
-fi
+# git submodule init;
+# git submodule update;
+
+# ### Java and Maven images
+# docker pull openjdk:11;
+# docker pull maven:3.9.8-eclipse-temurin-8;
+
+# ### Defects4J image
+# cd benchmarks/defects4j;
+# cpanm --installdeps .;
+# ./init.sh;
+# cd ../..;
+
+# ### GitBug-Java
+# cd benchmarks/gitbug-java;
+# chmod +x gitbug-java;
+# poetry install --no-root;
+# # Skip setup if in CI
+# if [ -z "$CI" ]; then
+#  poetry run ./gitbug-java setup;
+# fi
+
+# mkdir benchmarks/run_bug_run
+# cd benchmarks/run_bug_run
+# wget https://github.com/giganticode/run_bug_run_data/releases/download/v0.0.1/python_valid0.jsonl.gz
+# wget https://github.com/giganticode/run_bug_run_data/releases/download/v0.0.1/tests_all.jsonl.gz
+
+# gzip -d python_valid0.jsonl.gz
+# gzip -d tests_all.jsonl.gz
+# cd ../..
+
+apt-get install php-cli nodejs gcc g++ default-jdk ruby python3 golang-go bubblewrap
+cd benchmarks/run_bug_run
+gem install bundler
+bundle install
+bundle exec rbugr download 0.0.1
+
+