"""
Adaptive Course Learner - Integrated system for learning from Canvas courses.
"""
import json
import pickle
from pathlib import Path
from adaptive_learner.canvas_fetcher import CanvasContentFetcher
from adaptive_learner.perceptron_model import PerceptronFeatureExtractor
from adaptive_learner.rl_agent import RLContextBuilder
class AdaptiveCourseLearner:
"""Orchestrates content fetching, perceptron learning, and RL context building."""
def __init__(self, knowledge_base_path: Path | None = None):
self.knowledge_base = knowledge_base_path or Path("knowledge_base")
self.knowledge_base.mkdir(parents=True, exist_ok=True)
self.models_path = self.knowledge_base / "models"
self.models_path.mkdir(exist_ok=True)
self.fetcher = CanvasContentFetcher()
self.perceptron = PerceptronFeatureExtractor()
self.rl_agent = RLContextBuilder()
self._load_models()
def learn_from_course(
self,
course_id: int,
iterations: int = 5,
course_content: dict | None = None,
) -> list[str]:
"""
Fetch course content (if not provided), learn patterns, and build optimal context.
Args:
course_id: Canvas course ID
iterations: Max RL iterations for context building
course_content: Optional pre-fetched content; if None, fetches from API
Returns:
List of context strings (module summaries)
"""
if course_content is None:
course_content = self.fetcher.fetch_course_content(course_id)
self.fetcher.save_to_folder(course_content, self.knowledge_base)
self.perceptron.learn_patterns([course_content])
context = self.rl_agent.build_context_iteratively(
course_content, self.perceptron, max_iterations=iterations
)
context_path = self.knowledge_base / f"course_{course_id}_context.json"
context_path.write_text(
json.dumps({"course_id": course_id, "context": context}, indent=2)
)
self._save_models()
return context
def get_course_content(self, course_id: int) -> dict:
"""Fetch and return raw course content (no learning)."""
return self.fetcher.fetch_course_content(course_id)
def _save_models(self) -> None:
"""Persist perceptron and RL agent."""
self.perceptron.save(self.models_path / "perceptron.pkl")
self.rl_agent.save(self.models_path / "rl_agent.json")
def _load_models(self) -> None:
"""Load persisted models if they exist."""
pkl = self.models_path / "perceptron.pkl"
if pkl.exists():
self.perceptron.load(pkl)
rl = self.models_path / "rl_agent.json"
if rl.exists():
self.rl_agent.load(rl)
