fixed all hydra related errors

wenting-zhao · wenting-zhao · commit ca5f87470dc3 · 2024-09-10T21:39:05.000Z
diff --git a/commit0/__main__.py b/commit0/__main__.py
@@ -1,18 +1,40 @@
 import commit0.harness.run_pytest_ids
 import commit0.harness.build
 import commit0.harness.setup
+import copy
 import sys
+import hydra
 
 
 def main() -> None:
     command = sys.argv[1]
+    # have hydra to ignore all command-line arguments
+    sys_argv = copy.deepcopy(sys.argv)
+    sys.argv = [sys.argv[0]]
+    hydra.initialize(version_base=None, config_path="../configs")
+    config = hydra.compose(config_name="base")
+    # after hydra gets all configs, put command-line arguments back
+    sys.argv = sys_argv
 
     if command == "clone":
-        commit0.harness.setup.main()
+        commit0.harness.setup.main(
+            config.dataset_name, config.dataset_split, config.base_dir
+        )
     elif command == "build":
-        commit0.harness.build.main()
+        commit0.harness.build.main(config.dataset_name, config.dataset_split)
     elif command == "test":
-        commit0.harness.run_pytest_ids.main()
+        repo = sys.argv[2]
+        test_ids = sys.argv[3]
+        commit0.harness.run_pytest_ids.main(
+            config.dataset_name,
+            config.dataset_split,
+            config.base_dir,
+            repo,
+            config.branch,
+            test_ids,
+            config.backend,
+            config.timeout,
+        )
 
 
 if __name__ == "__main__":
diff --git a/commit0/harness/build.py b/commit0/harness/build.py
@@ -4,21 +4,18 @@
 from datasets import load_dataset
 from typing import Iterator
 
-from omegaconf import DictConfig
 from commit0.harness.docker_build import build_repo_images
 from commit0.harness.spec import make_spec
 from commit0.harness.constants import RepoInstance
-import hydra
 
 logging.basicConfig(
     level=logging.INFO, format="%(asctime)s - %(name)s - %(levelname)s - %(message)s"
 )
 logger = logging.getLogger(__name__)
 
 
-@hydra.main(version_base=None, config_path="configs", config_name="base")
-def main(config: DictConfig) -> None:
-    dataset: Iterator[RepoInstance] = load_dataset(hf_name, split="test")  # type: ignore
+def main(dataset_name: str, dataset_split: str) -> None:
+    dataset: Iterator[RepoInstance] = load_dataset(dataset_name, split=dataset_split)  # type: ignore
     specs = []
     for example in dataset:
         spec = make_spec(example)
diff --git a/commit0/harness/docker_utils.py b/commit0/harness/docker_utils.py
@@ -330,7 +330,6 @@ def log_error(x: str) -> None:
         def log_info(x: str) -> None:
             print(x)
 
-        raise_error = True
     elif logger == "quiet":
         # if logger is "quiet", don't print anything
         def log_info(x: str) -> None:
diff --git a/commit0/harness/run_pytest_ids.py b/commit0/harness/run_pytest_ids.py
@@ -6,9 +6,6 @@
 from pathlib import Path
 import logging
 
-from omegaconf import DictConfig, OmegaConf
-import hydra
-
 from commit0.harness.constants import RUN_PYTEST_LOG_DIR
 from commit0.harness.docker_build import (
     close_logger,
@@ -196,39 +193,46 @@ def run_modal(
                 )
 
 
-@hydra.main(version_base=None, config_path="configs", config_name="base")
-def main(config: DictConfig) -> None:
-    OmegaConf.to_yaml(config)
-    dataset = load_dataset(config.dataset_name, split="test")
+def main(
+    dataset_name: str,
+    dataset_split: str,
+    base_dir: str,
+    repo: str,
+    branch: str,
+    test_ids: str,
+    backend: str,
+    timeout: int,
+) -> None:
+    dataset = load_dataset(dataset_name, split="test")
     spec = None
     for example in dataset:
-        if example["repo"].endswith(config.repo):
+        if example["repo"].endswith(repo):
             spec = make_spec(example)
             break
     assert spec is not None, "No spec available"
 
-    hashed_test_ids = get_hash_string(config.test_ids)
+    hashed_test_ids = get_hash_string(test_ids)
     # set up logging
-    log_dir = RUN_PYTEST_LOG_DIR / config.repo / hashed_test_ids
+    log_dir = RUN_PYTEST_LOG_DIR / repo / hashed_test_ids
     log_dir.mkdir(parents=True, exist_ok=True)
     log_file = log_dir / "run_pytest.log"
-    logger = setup_logger(config.repo, log_file)
+    logger = setup_logger(repo, log_file)
 
     # make eval file
     eval_script = spec.eval_script.format(
-        local_repo=f"{config.base_dir}/{config.repo}",
-        branch_name=config.branch,
-        test_ids=config.test_ids,
-        ip=get_ip(config.backend),
+        local_repo=f"{base_dir}/{repo}",
+        branch_name=branch,
+        test_ids=test_ids,
+        ip=get_ip(backend),
         user=get_user(),
     )
     eval_file = Path(log_dir / "eval.sh")
     eval_file.write_text(eval_script)
 
-    if ExecutionBackend(config.backend) == ExecutionBackend.LOCAL:
-        run_docker(spec, logger, eval_file, config.timeout, log_dir)
-    elif ExecutionBackend(config.backend) == ExecutionBackend.MODAL:
-        run_modal(spec, logger, eval_file, config.timeout, log_dir)
+    if ExecutionBackend(backend) == ExecutionBackend.LOCAL:
+        run_docker(spec, logger, eval_file, timeout, log_dir)
+    elif ExecutionBackend(backend) == ExecutionBackend.MODAL:
+        run_modal(spec, logger, eval_file, timeout, log_dir)
 
 
 __all__ = []
diff --git a/commit0/harness/setup.py b/commit0/harness/setup.py
@@ -2,9 +2,7 @@
 import os
 
 import docker
-import hydra
 from datasets import load_dataset
-from omegaconf import DictConfig
 
 from typing import Iterator
 from commit0.harness.utils import clone_repo
@@ -19,9 +17,8 @@
 logger = logging.getLogger(__name__)
 
 
-@hydra.main(version_base=None, config_path="configs", config_name="base")
-def main(config: DictConfig) -> None:
-    dataset: Iterator[RepoInstance] = load_dataset(hf_name, split="test")  # type: ignore
+def main(dataset_name: str, dataset_split: str, base_dir: str) -> None:
+    dataset: Iterator[RepoInstance] = load_dataset(dataset_name, split="test")  # type: ignore
     out = dict()
     specs = []
     for example in dataset:
