cleanup; remove checkout during test

Author: wenting-zhao

commit0/harness/constants.py

@@ -15,6 +15,7 @@ class Files(TypedDict):
     eval_script: Dict[str, Path]
     patch: Dict[str, Path]
 
+BASE_BRANCH = "commit0"
 
 # Constants - Evaluation Log Directories
 BASE_IMAGE_BUILD_DIR = Path("logs/build_images/base")

commit0/harness/run_pytest_ids.py

@@ -85,12 +85,24 @@ def main(
     if branch == "reference":
         commit_id = example["reference_commit"]
     else:
-        try:
-            local_repo.git.checkout(branch)
-            local_branch = local_repo.branches[branch]
-            commit_id = local_branch.commit.hexsha
-        except Exception as e:
-            raise Exception(f"Problem checking out branch {branch}.\n{e}")
+        # Check if it's a local branch
+        if branch in local_repo.branches:
+            commit_id = local_repo.commit(branch)
+        else:
+            found_remote_branch = False
+            for remote in local_repo.remotes:
+                remote.fetch()  # Fetch latest updates from each remote
+
+                # Check if the branch exists in this remote
+                for ref in remote.refs:
+                    if ref.remote_head == branch:  # Compare branch name without remote prefix
+                        commit_id = local_repo.commit(ref.name)
+                        found_remote_branch = True
+                        break  # Branch found, no need to keep checking this remote
+                if found_remote_branch:
+                    break  # Stop checking other remotes if branch is found
+            if not found_remote_branch:
+                raise Exception(f"Branch {branch} does not exist locally or remotely.")
     patch = generate_patch_between_commits(
         local_repo, example["base_commit"], commit_id
     )

commit0/harness/setup.py

@@ -7,7 +7,7 @@
 from commit0.harness.utils import (
     clone_repo,
 )
-from commit0.harness.constants import RepoInstance, SPLIT
+from commit0.harness.constants import BASE_BRANCH, RepoInstance, SPLIT
 
 
 logging.basicConfig(
@@ -29,7 +29,12 @@ def main(
             continue
         clone_url = f"https://github.com/{example['repo']}.git"
         clone_dir = os.path.abspath(os.path.join(base_dir, repo_name))
-        clone_repo(clone_url, clone_dir, example["base_commit"], logger)
+        branch = dataset_name.split('/')[-1]
+        repo = clone_repo(clone_url, clone_dir, branch, logger)
+        if BASE_BRANCH in repo.branches:
+            repo.git.branch('-d', BASE_BRANCH)
+        repo.git.checkout('-b', BASE_BRANCH)
+        logger.info("Checked out the base commit: commit 0")
 
 
 __all__ = []

commit0/harness/utils.py

@@ -85,7 +85,7 @@ def extract_test_output(ss: str, pattern: str) -> str:
 
 
 def clone_repo(
-    clone_url: str, clone_dir: str, commit: str, logger: logging.Logger
+    clone_url: str, clone_dir: str, branch: str, logger: logging.Logger
 ) -> git.Repo:
     """Clone repo into the specified directory if it does not already exist.
 
@@ -98,8 +98,8 @@ def clone_repo(
         URL of the repository to clone.
     clone_dir : str
         Directory where the repository will be cloned.
-    commit : str
-        The commit hash or branch/tag name to checkout.
+    branch : str
+        The branch/tag name to checkout.
     logger : logging.Logger
         The logger object.
 
@@ -129,11 +129,10 @@ def clone_repo(
         except git.exc.GitCommandError as e:
             raise RuntimeError(f"Failed to clone repository: {e}")
 
-    logger.info(f"Checking out {commit}")
     try:
-        repo.git.checkout(commit)
+        repo.git.checkout(branch)
     except git.exc.GitCommandError as e:
-        raise RuntimeError(f"Failed to check out {commit}: {e}")
+        raise RuntimeError(f"Failed to check out {branch}: {e}")
 
     return repo