Stop excluding dockerignored files from uploads

Author: benjibc

eval_protocol/evaluation.py

@@ -33,7 +33,7 @@ def __init__(
 
     @staticmethod
     def _parse_ignore_file(ignore_path: str) -> List[str]:
-        """Parse .gitignore or .dockerignore and return patterns."""
+        """Parse .gitignore and return patterns."""
         patterns = []
         if not os.path.exists(ignore_path):
             return patterns
@@ -128,8 +128,7 @@ def _create_tar_gz_with_ignores(output_path: str, source_dir: str) -> int:
 
         source_path = Path(source_dir)
         gitignore_patterns = Evaluator._parse_ignore_file(str(source_path / ".gitignore"))
-        dockerignore_patterns = Evaluator._parse_ignore_file(str(source_path / ".dockerignore"))
-        all_ignore_patterns = gitignore_patterns + dockerignore_patterns
+        all_ignore_patterns = gitignore_patterns
 
         logger.info(f"Creating tar.gz with {len(all_ignore_patterns)} ignore patterns")
 

tests/test_ep_upload_e2e.py

@@ -13,6 +13,7 @@
 import os
 import shutil
 import sys
+import tarfile
 import tempfile
 from pathlib import Path
 from unittest.mock import MagicMock, patch
@@ -566,6 +567,38 @@ async def test_math_correctness(row: EvaluationRow) -> EvaluationRow:
         shutil.rmtree(test_project_dir, ignore_errors=True)
 
 
+def test_create_tar_includes_dockerignored_files(tmp_path):
+    from eval_protocol.evaluation import Evaluator
+
+    project_dir = tmp_path / "project"
+    project_dir.mkdir()
+    (project_dir / "requirements.txt").write_text("")
+    (project_dir / "Dockerfile").write_text("FROM python:3.11\n")
+    (project_dir / ".dockerignore").write_text("Dockerfile\nignored_dir/\n")
+
+    ignored_dir = project_dir / "ignored_dir"
+    ignored_dir.mkdir()
+    (ignored_dir / "data.txt").write_text("package me\n")
+
+    tar_path = tmp_path / "archive.tar.gz"
+    archive_size = Evaluator._create_tar_gz_with_ignores(str(tar_path), str(project_dir))
+
+    assert archive_size > 0
+    with tarfile.open(tar_path, "r:gz") as tar:
+        names = tar.getnames()
+
+    project_prefix = project_dir.name
+    expected_paths = [
+        f"{project_prefix}/Dockerfile",
+        f"{project_prefix}/.dockerignore",
+        f"{project_prefix}/ignored_dir/data.txt",
+        f"{project_prefix}/requirements.txt",
+    ]
+
+    for expected_path in expected_paths:
+        assert expected_path in names, f"Expected {expected_path} in archive"
+
+
 def test_ep_upload_force_flag_triggers_delete_flow(
     mock_env_variables,
     mock_gcs_upload,