Merge pull request #440 from mlcommons/dev

Sync Dev

Author: anandhu-eng

.github/scripts/list_modified_files.py

@@ -8,7 +8,8 @@ def get_file_info(filepath):
     with open(filepath, 'r') as file:
         content = yaml.safe_load(file)
         tests = content.get('tests', [])
-        if tests:
+        needs_pat = content.get('needs_pat', False)
+        if tests and not needs_pat:
             num_tests = len(tests.get('run_inputs', []))
         else:
             num_tests = 0
@@ -17,7 +18,7 @@ def get_file_info(filepath):
 
 
 def process_files(files):
-    filenames = files.split()
+    filenames = files.split(",")
     return [
         {
             "file": file,
@@ -34,4 +35,4 @@ def process_files(files):
     changed_files = sys.stdin.read().strip()
     processed_files = process_files(changed_files)
     json_processed_files = json.dumps(processed_files)
-    print(f"::set-output name=processed_files::{json_processed_files}")
+    print(json_processed_files)

.github/workflows/build_wheels.yml

@@ -5,6 +5,7 @@ on:
     types: [published]
   workflow_dispatch: {}
 
+
 jobs:
   build_wheels:
     if: github.repository_owner == 'mlcommons'

.github/workflows/run-tests-on-modified-meta.yml

@@ -33,7 +33,7 @@ jobs:
       run: |
         git remote add upstream ${{ github.event.pull_request.base.repo.clone_url }}
         git fetch upstream
-        changed_files=$(git diff upstream/${{ github.event.pull_request.base.ref }} --name-only)
+        changed_files=$(git diff upstream/${{ github.event.pull_request.base.ref }} --name-only | paste -sd, -)
         echo "$changed_files" | python3 .github/scripts/list_modified_files.py
 
   process_modified_files:

.github/workflows/test-mlc-script-features.yml

@@ -50,18 +50,15 @@ jobs:
         mlcr get,wkhtmltopdf --quiet
 
     - name: Test versions
-      continue-on-error: true
-      if: runner.os == 'linux'
+      if: runner.os == 'linux' && matrix.python-version == '3.8'
       run: |
         mlcr get,generic-python-lib,_package.scipy --version=1.9.3 --quiet
         test $? -eq 0 || exit $?
         mlcr get,generic-python-lib,_package.scipy --version=1.9.2 --quiet
         test $? -eq 0 || exit $?
-        mlc find cache --tags=get,generic-python-lib,_package.scipy,version-1.9.3
+        mlc find cache --tags=get,generic-python-lib,_package.scipy,version-1.9.2
         test $? -eq 0 || exit $?
-        # Need to add find cache here
-        # mlcr get,generic-python-lib,_package.scipy --version=1.9.3 --quiet --only_execute_from_cache=True
-        # test $? -eq 0 || exit 0
+       
 
     - name: Test python install from src
       run: |
@@ -94,6 +91,34 @@ jobs:
       run: |
         mlcr run,docker,container --adr.compiler.tags=gcc --docker_mlc_repo=mlcommons@mlperf-automations --docker_mlc_repo_branch=dev --image_name=mlc-script-app-image-classification-onnx-py --env.MLC_DOCKER_RUN_SCRIPT_TAGS=app,image-classification,onnx,python --env.MLC_DOCKER_IMAGE_BASE=ubuntu:22.04 --env.MLC_DOCKER_IMAGE_REPO=local --quiet
 
+  test_experiment:
+    runs-on: ${{ matrix.os }}
+    strategy:
+      fail-fast: false
+      matrix:
+        python-version: ["3.12", "3.8"]
+        os: ["ubuntu-latest", "windows-latest", "macos-latest"]
+
+    steps:
+    - uses: actions/checkout@v4
+    - name: Set up Python ${{ matrix.python-version }}
+      uses: actions/setup-python@v3
+      with:
+        python-version: ${{ matrix.python-version }}
+    - name: Configure git longpaths (Windows)
+      if: matrix.os == 'windows-latest'
+      run: |
+        git config --system core.longpaths true
+    - name: Pull MLOps repository
+      run: |
+        pip install mlcflow
+        mlc pull repo ${{ github.event.pull_request.head.repo.html_url }} --branch=${{ github.event.pull_request.head.ref }}
+
+    - name: Test mlc experiment script
+      run: |
+        mlc experiment script --tags=detect,os --quiet --exp.repeat,=1,2,3
+        mlc experiment script --tags=detect,cpu --quiet --exp.explore=2:10:2
+
   test_mlperf_retinanet_cpp_venv:
     runs-on: ubuntu-latest
     strategy:

.github/workflows/test-mlperf-automotive.yml

@@ -0,0 +1,43 @@
+name: Test MLPerf automotive
+
+on:
+  pull_request_target:
+    branches: [ "main", "dev" ]
+    paths:
+      - 'script/**meta.yaml'
+
+jobs:
+  fetch-secret:
+    runs-on: ubuntu-latest
+    outputs:
+      gdrive_secret: ${{ steps.op-load-secret.outputs.GDRIVE_SERVICE_ACCOUNT_KEY }}
+    steps:
+      - name: Load secret from 1Password
+        id: op-load-secret
+        uses: 1password/load-secrets-action@v2
+        with:
+          export-env: false
+        env:
+          OP_SERVICE_ACCOUNT_TOKEN: ${{ secrets.OP_SERVICE_ACCOUNT_TOKEN }}
+          GDRIVE_SERVICE_ACCOUNT_KEY: op://7basd2jirojjckncf6qnq3azai/bzbaco3uxoqs2rcyu42rvuccga/credential
+
+  run-mlperf:
+    runs-on: ubuntu-latest
+    needs: 
+      - fetch-secret
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 2
+      - name: Set RCLONE Service account env var from secret
+        shell: bash
+        run: |
+          echo "::add-mask::${{ needs.fetch-secret.outputs.gdrive_secret }}"
+          echo "RCLONE_CONFIG_MLC_COGNATA_SERVICE_ACCOUNT_CREDENTIALS=${{ needs.fetch-secret.outputs.gdrive_secret }}" >> $GITHUB_ENV
+          echo "RCLONE_CONFIG_MLC_NUSCENES_SERVICE_ACCOUNT_CREDENTIALS=${{ needs.fetch-secret.outputs.gdrive_secret }}" >> $GITHUB_ENV
+      - name: Run MLPerf
+        run: |
+          pip install mlcflow
+          mlc pull repo ${{ github.event.pull_request.head.repo.html_url }} --branch=${{ github.event.pull_request.head.ref }}
+          mlc test script app,mlperf,automotive

automation/script/docker.py

@@ -36,7 +36,6 @@ def dockerfile(self_module, input_params):
     state_data = input_params.get('state', {})
     constant_vars = input_params.get('const', {})
     constant_state = input_params.get('const_state', {})
-    dockerfile_env = input_params.get('dockerfile_env', {})
     tag_values = input_params.get('tags', '').split(",")
     variation_tags = [tag[1:] for tag in tag_values if tag.startswith("_")]
 
@@ -56,7 +55,6 @@ def dockerfile(self_module, input_params):
             'script_variation_tags': variation_tags
         }
         docker_settings = metadata.get('docker', {})
-        docker_settings['dockerfile_env'] = dockerfile_env
         state_data['docker'] = docker_settings
         add_deps_recursive = input_params.get('add_deps_recursive', {})
 
@@ -94,8 +92,6 @@ def dockerfile(self_module, input_params):
 
         # Set Docker-specific configurations
         docker_settings = state_data['docker']
-        docker_settings['dockerfile_env'] = dockerfile_env
-        dockerfile_env['MLC_RUN_STATE_DOCKER'] = True
 
         if not docker_settings.get('run', True) and not input_params.get(
                 'docker_run_override', False):
@@ -171,13 +167,18 @@ def dockerfile(self_module, input_params):
         if input_params.get('docker_push_image') in [True, 'True', 'yes']:
             env['MLC_DOCKER_PUSH_IMAGE'] = 'yes'
 
+        dockerfile_env = docker_inputs.get('env', {})
+        dockerfile_build_env = docker_inputs.get('build_env', {})
+
+        dockerfile_env['MLC_RUN_STATE_DOCKER'] = True
         # Generate Dockerfile
         mlc_docker_input = {
             'action': 'run', 'automation': 'script', 'tags': 'build,dockerfile',
             'fake_run_option': " " if docker_inputs.get('real_run') else " --fake_run",
             'comments': comments, 'run_cmd': f"{run_command_string} --quiet",
             'script_tags': input_params.get('tags'), 'env': env,
             'dockerfile_env': dockerfile_env,
+            'dockerfile_build_env': dockerfile_build_env,
             'quiet': True, 'real_run': True
         }
 
@@ -377,6 +378,9 @@ def docker_run(self_module, i):
             env.update({docker_input_mapping[key]: i[key]
                        for key in docker_input_mapping if key in i})
 
+        if docker_inputs.get('user'):
+            docker_settings['user'] = docker_inputs['user']
+
         # Handle environment variable-based mounts
         res = process_mounts(
             mounts,
@@ -412,6 +416,7 @@ def docker_run(self_module, i):
             'quiet': True, 'real_run': True, 'add_deps_recursive': {'build-docker-image': {'dockerfile': dockerfile_path}},
             **docker_inputs
         }
+
         r = self_module.action_object.access(mlc_docker_input)
         if r['return'] > 0:
             return r

automation/script/docker_utils.py

@@ -134,7 +134,7 @@ def prepare_docker_inputs(input_params, docker_settings,
     keys = [
         "mlc_repo", "mlc_repo_branch", "base_image", "os", "os_version",
         "mlc_repos", "skip_mlc_sys_upgrade", "extra_sys_deps", "image_name",
-        "gh_token", "fake_run_deps", "run_final_cmds", "real_run", "copy_files", "path", "user"
+        "gh_token", "fake_run_deps", "run_final_cmds", "real_run", "copy_files", "path", "user", "env", "build_env"
     ]
 
     if run_stage:
@@ -452,7 +452,10 @@ def get_container_path(value, username="mlcuser", extract_parent_folder=False):
 
     new_value = ''
     if "cache" in path_split and "local" in path_split:
-        new_path_split = ["", "home", username, "MLC", "repos"]
+        if username == "root":
+            new_path_split = ["", "root", "MLC", "repos"]
+        else:
+            new_path_split = ["", "home", username, "MLC", "repos"]
         repo_entry_index = path_split.index("local")
         if len(path_split) >= repo_entry_index + 3:
             new_path_split1 = new_path_split + \