mlcflow 1.2.5__tar.gz → 1.2.6__tar.gz

{mlcflow-1.2.5 → mlcflow-1.2.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mlcflow
-Version: 1.2.5
+Version: 1.2.6
 Summary: An automation interface tailored for CPU/GPU benchmarking
 Author-email: MLCommons <systems@mlcommons.org>
 License: 

{mlcflow-1.2.5 → mlcflow-1.2.6}/mlc/repo_action.py

@@ -2,6 +2,7 @@ from .action import Action
 import os
 import subprocess
 import re
+import shlex
 import yaml
 import json
 import shutil
@@ -322,8 +323,21 @@ class RepoAction(Action):
             return {"return": 0, "value": os.path.basename(
                 url).replace(".git", "")}
 
+    def _is_shallow_repo(self, repo_path):
+        """Return True if the git repository at *repo_path* is a shallow clone."""
+        try:
+            result = subprocess.run(
+                ['git', '-C', repo_path, 'rev-parse', '--is-shallow-repository'],
+                capture_output=True,
+                text=True,
+            )
+            return result.returncode == 0 and result.stdout.strip() == 'true'
+        except (subprocess.SubprocessError, FileNotFoundError, OSError):
+            return False
+
     def pull_repo(self, repo_url, branch=None, checkout=None, tag=None,
-                  pat=None, ssh=None, ignore_on_conflict=False, repo_path=None, force=False):
+                  pat=None, ssh=None, ignore_on_conflict=False, repo_path=None, force=False,
+                  shallow=False, depth=None, extra_git_args=None):
 
         # Determine the checkout path from environment or default
         repo_base_path = self.repos_path  # either the value will be from 'MLC_REPOS'
@@ -358,20 +372,41 @@ class RepoAction(Action):
                 repo_path = os.path.join(repo_base_path, repo_download_name)
 
         try:
+            # Compute depth argument: --shallow implies depth=1; explicit
+            # --depth=N takes precedence
+            clone_depth = None
+            if depth is not None:
+                try:
+                    clone_depth = int(depth)
+                except (TypeError, ValueError):
+                    return {
+                        "return": 1, "error": f"Invalid value for --depth: {depth!r}. Must be a positive integer."}
+                if clone_depth < 1:
+                    return {
+                        "return": 1, "error": f"Invalid value for --depth: {clone_depth}. Must be a positive integer."}
+            elif shallow:
+                clone_depth = 1
+
+            # Parse extra_git_args into a list
+            extra_args = []
+            if extra_git_args:
+                if isinstance(extra_git_args, list):
+                    extra_args = extra_git_args
+                else:
+                    extra_args = shlex.split(str(extra_git_args))
+
             # If the directory doesn't exist, clone it
             if not os.path.exists(repo_path):
                 logger.info(f"Cloning repository {repo_url} to {repo_path}...")
 
-                # Build clone command without branch if not provided
-                clone_command = ['git', 'clone', repo_url, repo_path]
+                # Build clone command
+                clone_command = ['git', 'clone']
                 if branch:
-                    clone_command = [
-                        'git',
-                        'clone',
-                        '--branch',
-                        branch,
-                        repo_url,
-                        repo_path]
+                    clone_command += ['--branch', branch]
+                if clone_depth is not None:
+                    clone_command += ['--depth', str(clone_depth)]
+                clone_command += extra_args
+                clone_command += [repo_url, repo_path]
 
                 subprocess.run(clone_command, check=True)
 
@@ -433,8 +468,17 @@ class RepoAction(Action):
                     logger.info(
                         "Pulling latest changes...")
                     try:
+                        pull_command = ['git', '-C', repo_path, 'pull']
+                        if clone_depth is not None:
+                            if self._is_shallow_repo(repo_path):
+                                pull_command += ['--depth', str(clone_depth)]
+                            else:
+                                logger.warning(
+                                    f"--depth/--shallow ignored for pull on non-shallow repo {repo_path}. "
+                                    "Re-clone with --shallow to create a shallow copy."
+                                )
                         subprocess.run(
-                            ['git', '-C', repo_path, 'pull'],
+                            pull_command,
                             capture_output=True,
                             text=True,
                             check=True)
@@ -492,8 +536,16 @@ class RepoAction(Action):
                 else:
                     logger.info(
                         "No local changes detected. Pulling latest changes...")
-                    subprocess.run(
-                        ['git', '-C', repo_path, 'pull'], check=True)
+                    pull_command = ['git', '-C', repo_path, 'pull']
+                    if clone_depth is not None:
+                        if self._is_shallow_repo(repo_path):
+                            pull_command += ['--depth', str(clone_depth)]
+                        else:
+                            logger.warning(
+                                f"--depth/--shallow ignored for pull on non-shallow repo {repo_path}. "
+                                "Re-clone with --shallow to create a shallow copy."
+                            )
+                    subprocess.run(pull_command, check=True)
                     logger.info("Repository successfully pulled.")
 
             if tag:
@@ -562,6 +614,9 @@ class RepoAction(Action):
     - `--tag <release_tag>`: Checks out a particular release tag.
     - `--pat <access_token>` or `--ssh`: Clones a private repository using a personal access token or SSH.
     - `--force`: For existing repositories with local tracked changes, stashes changes before pull and reapplies them after pull.
+    - `--shallow`: Perform a shallow clone with `--depth=1` (fastest for a fresh copy without history). For existing repos, only applied if the repo is already shallow; otherwise ignored with a warning.
+    - `--depth=N`: Perform a shallow clone/pull with the specified history depth (e.g. `--depth=5`). For existing repos, `--depth` is only applied when the repository is already a shallow clone; passing `--depth` to a full-history clone would corrupt it and is therefore ignored with a warning.
+    - `--extra_git_args=<args>`: Pass additional arguments to the `git clone` command (e.g. `--extra_git_args="--filter=blob:none"`). Only applies when cloning a new repository; not used for pull on existing repos. Accepts only trusted input — arguments are passed directly to git without further validation.
 
     Example Output:
 
@@ -592,7 +647,11 @@ class RepoAction(Action):
                         repo_object.path, os.W_OK):
                     repo_folder_name = os.path.basename(repo_object.path)
                     res = self.pull_repo(
-                        repo_folder_name, repo_path=repo_object.path, force=run_args.get('force'))
+                        repo_folder_name, repo_path=repo_object.path, force=run_args.get(
+                            'force'),
+                        shallow=run_args.get('shallow', False),
+                        depth=run_args.get('depth'),
+                        extra_git_args=run_args.get('extra_git_args'))
                     if res['return'] > 0:
                         return res
         else:
@@ -604,6 +663,9 @@ class RepoAction(Action):
             ssh = run_args.get('ssh')
             force = run_args.get('force')
             ignore_on_conflict = run_args.get('ignore_on_conflict')
+            shallow = run_args.get('shallow', False)
+            depth = run_args.get('depth')
+            extra_git_args = run_args.get('extra_git_args')
 
             if sum(bool(var) for var in [branch, checkout, tag]) > 1:
                 return {
@@ -617,7 +679,10 @@ class RepoAction(Action):
                 pat,
                 ssh,
                 ignore_on_conflict=ignore_on_conflict,
-                force=force)
+                force=force,
+                shallow=shallow,
+                depth=depth,
+                extra_git_args=extra_git_args)
             if res['return'] > 0:
                 return res
 

{mlcflow-1.2.5 → mlcflow-1.2.6}/mlcflow.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mlcflow
-Version: 1.2.5
+Version: 1.2.6
 Summary: An automation interface tailored for CPU/GPU benchmarking
 Author-email: MLCommons <systems@mlcommons.org>
 License: 

{mlcflow-1.2.5 → mlcflow-1.2.6}/mlcflow.egg-info/SOURCES.txt

@@ -26,4 +26,5 @@ mlcflow.egg-info/requires.txt
 mlcflow.egg-info/top_level.txt
 tests/test_action_invalid_meta_entries.py
 tests/test_cache_mark_tmp.py
+tests/test_mlcflow_unix_installer.py
 tests/test_script_action_apptainer.py

{mlcflow-1.2.5 → mlcflow-1.2.6}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "mlcflow"
-version = "1.2.5"
+version = "1.2.6"
 
 description = "An automation interface tailored for CPU/GPU benchmarking"
 authors = [

mlcflow-1.2.6/tests/test_mlcflow_unix_installer.py

@@ -0,0 +1,83 @@
+import subprocess
+import platform
+import sys
+from pathlib import Path
+
+
+REPO_ROOT = Path(__file__).resolve().parents[1]
+INSTALLER_SCRIPT = REPO_ROOT / "docs" / "install" / "mlcflow_unix_installer.sh"
+
+
+def _resolve_venv_dir(tmp_path, setup_snippet):
+    default_dir = tmp_path / "mlcflow"
+    marker = "__RESULT__"
+    command = f"""
+set -euo pipefail
+source "{INSTALLER_SCRIPT}"
+DEFAULT_VENV_DIR="{default_dir}"
+VENV_DIR="$DEFAULT_VENV_DIR"
+PY_MAJOR_MINOR="$(python3 -c 'import sys; print("{{}}.{{}}".format(*sys.version_info[:2]))')"
+PY_ARCH="$(python3 -c 'import platform; print(platform.machine())')"
+{setup_snippet}
+resolve_default_venv_dir
+echo "{marker}$VENV_DIR"
+"""
+    result = subprocess.run(
+        ["bash", "-lc", command],
+        text=True,
+        capture_output=True,
+        check=True,
+    )
+
+    for line in reversed(result.stdout.splitlines()):
+        if line.startswith(marker):
+            return line[len(marker):]
+
+    raise AssertionError(
+        f"Could not parse resolved venv path from output:\n{result.stdout}")
+
+
+def test_resolve_default_venv_dir_prefers_compatible_default(tmp_path):
+    setup_snippet = """
+python3 -m venv "$DEFAULT_VENV_DIR"
+"""
+    resolved = _resolve_venv_dir(tmp_path, setup_snippet)
+    assert resolved == str(tmp_path / "mlcflow")
+
+
+def test_resolve_default_venv_dir_uses_suffix_for_incompatible_default(
+        tmp_path):
+    setup_snippet = """
+mkdir -p "$DEFAULT_VENV_DIR"
+"""
+    resolved = _resolve_venv_dir(tmp_path, setup_snippet)
+    assert resolved.startswith(str(tmp_path / "mlcflow_"))
+    assert "_py" in resolved
+
+
+def test_resolve_default_venv_dir_removes_stale_shared_env_when_default_incompatible(
+        tmp_path):
+    """When the default venv is incompatible and the suffixed venv already exists
+    but is itself stale/broken, the stale suffixed dir must be removed so that
+    setup_venv can create a fresh one rather than silently reusing a broken env."""
+    setup_snippet = """
+mkdir -p "$DEFAULT_VENV_DIR"
+SHARED_VENV_DIR="${DEFAULT_VENV_DIR}_${PY_ARCH}_py${PY_MAJOR_MINOR}"
+mkdir -p "$SHARED_VENV_DIR"
+"""
+    resolved = _resolve_venv_dir(tmp_path, setup_snippet)
+    expected_suffix = f"{tmp_path / 'mlcflow'}_{platform.machine()}_py{sys.version_info[0]}.{sys.version_info[1]}"
+    assert resolved == expected_suffix
+    # The stale shared dir should have been removed so setup_venv creates it
+    # fresh
+    assert not Path(expected_suffix).exists()
+
+
+def test_resolve_default_venv_dir_reuses_compatible_suffixed_env(tmp_path):
+    setup_snippet = """
+SHARED_VENV_DIR="${DEFAULT_VENV_DIR}_${PY_ARCH}_py${PY_MAJOR_MINOR}"
+python3 -m venv "$SHARED_VENV_DIR"
+"""
+    resolved = _resolve_venv_dir(tmp_path, setup_snippet)
+    expected = f"{tmp_path / 'mlcflow'}_{platform.machine()}_py{sys.version_info[0]}.{sys.version_info[1]}"
+    assert resolved == expected