New Architecture - Pipeline & Sandbox Pool & Database

autograder/autograder.py

@@ -0,0 +1,57 @@
+from autograder.services.report.reporter_service import ReporterService
+from autograder.services.upstash_driver import UpstashDriver
+from autograder.pipeline import AutograderPipeline
+from autograder.steps.export_step import ExporterStep
+from autograder.steps.feedback_step import FeedbackStep
+from autograder.steps.grade_step import GradeStep
+from autograder.steps.load_template_step import TemplateLoaderStep
+from autograder.steps.pre_flight_step import PreFlightStep
+from autograder.steps.build_tree_step import BuildTreeStep
+
+
+def build_pipeline(
+                 template_name,
+                 include_feedback,
+                 grading_criteria,
+                 feedback_config,
+                 setup_config = None,
+                 custom_template = None,
+                 feedback_mode = None):
+    """
+    Build an autograder pipeline based on configuration.
+
+    Args:
+        template_name: Name of the template to use
+        include_feedback: Whether to include feedback generation
+        grading_criteria: Criteria configuration dictionary
+        feedback_config: Configuration for feedback generation
+        setup_config: Pre-flight setup configuration
+        custom_template: Custom template object (if any)
+        feedback_mode: Mode for feedback generation
+    Returns:
+        Configured AutograderPipeline
+    """
+    pipeline = AutograderPipeline()
+
+    # Load template
+    pipeline.add_step(TemplateLoaderStep(template_name, custom_template)) # Passes the template to the next step
+
+    pipeline.add_step(BuildTreeStep(grading_criteria)) # Uses template to match selected tests in criteria and builds tree
+
+    # Pre-flight checks (if configured)
+    if setup_config:
+        pipeline.add_step(PreFlightStep(setup_config))
+
+    pipeline.add_step(GradeStep()) # Generates GradingResult with final score and result tree
+
+    # Feedback generation (if configured)
+    if include_feedback:
+        reporter_service = ReporterService(feedback_mode=feedback_mode)
+        pipeline.add_step(FeedbackStep(reporter_service, feedback_config)) # Uses GradingResult to generate feedback and appends it to GradingResult
+
+    # Export results
+    pipeline.add_step(ExporterStep(UpstashDriver)) # Exports final results and feedback
+
+    return pipeline
+
+