feat: Log commit sha (#537)

# Motivation

To be able to link a pipeline rune with the code that has executed it,
we want to log the git sha of the last commit in the pipeline logfile.

modyn/supervisor/internal/pipeline_executor/models.py

@@ -16,6 +16,7 @@
 from modyn.supervisor.internal.eval.handler import EvalRequest
 from modyn.supervisor.internal.grpc.enums import PipelineStage
 from modyn.supervisor.internal.utils.evaluation_status_reporter import EvaluationStatusReporter
+from modyn.supervisor.internal.utils.git_utils import get_head_sha
 from pydantic import BaseModel, Field, model_serializer, model_validator
 from typing_extensions import override
 
@@ -472,6 +473,7 @@ class PipelineLogs(BaseModel):
     # static logs
 
     pipeline_id: int
+    commit_sha: str | None = Field(None, description="The commit SHA that the pipeline was executed on.")
     pipeline_stages: dict[str, tuple[int, list[str]]]
     """List of all pipeline stages, their execution order index, and their parent stages."""
 
@@ -516,6 +518,11 @@ def materialize(self, log_dir_path: Path, mode: Literal["initial", "increment",
         # output logs
 
         if mode == "initial":
+            try:
+                self.commit_sha = get_head_sha()
+            except Exception as e:  # pylint: disable=broad-except
+                logger.warning(f"Failed to get commit SHA: {e}")
+
             with open(pipeline_logdir / "pipeline.part.log", "w", encoding="utf-8") as logfile:
                 logfile.write(self.model_dump_json(indent=2, exclude={"supervisor", "partial_idx"}, by_alias=True))
             return

modyn/supervisor/internal/utils/git_utils.py

@@ -0,0 +1,12 @@
+import subprocess
+from pathlib import Path
+
+
+def get_head_sha() -> str:
+    """When launched in a git repository, return the SHA of the current HEAD."""
+
+    result = subprocess.run(
+        ["git", "rev-parse", "HEAD"], capture_output=True, text=True, check=True, cwd=Path(__file__).parent
+    )
+    assert result.returncode == 0, f"Failed to get HEAD SHA: {result.stderr}, {result.stdout}"
+    return result.stdout.strip()