debug_quiz.py

from src.preprocessor import Preprocessor
from src.extractor import ConceptExtractor
from src.hierarchy import HierarchyBuilder
from src.quiz import QuizGenerator
from src.utils import HeuristicNLP

with open("notes.txt", "r", encoding="utf-8") as f:
    text = f.read()

doc = Preprocessor().process(text)
concepts = ConceptExtractor().extract(doc)
graph = HierarchyBuilder().build(doc, concepts)
generator = QuizGenerator(graph)

print(f"Total nodes: {len(graph.nodes)}")
print(f"\nChecking high-quality concepts:")

for node in list(graph.nodes.values())[:15]:
    print(f"\n'{node.term}':")
    print(f"  Has def: {node.definition_sentence is not None}")
    if node.definition_sentence:
        print(f"  Def: {node.definition_sentence[:80]}...")
        print(f"  Is def sentence: {HeuristicNLP.is_definition_sentence(node.definition_sentence)}")
        
        s_lower = node.definition_sentence.lower()
        t_lower = node.term.lower()
        prefixes = [t_lower, f"a {t_lower}", f"an {t_lower}", f"the {t_lower}"]
        starts_ok = any(s_lower.startswith(p) for p in prefixes)
        print(f"  Starts with term: {starts_ok}")
    
    is_hq = generator._is_high_quality_concept(node)
    print(f"  ✓ HIGH QUALITY: {is_hq}")