debug_retrieval.py

import sys
import os
import uuid

sys.stdout.reconfigure(encoding='utf-8')

from app.services import rag_service, retrieval_service, pdf_processor
from app.services.tutor_context import build_tutor_context

timings = {}
doc_name = 'SCHAND PHYSICS.pdf'
q = "Let's start learning Chapter 1"

print('=== 1. Testing DocumentIndex & Chapter 1 ===')
doc_idx = pdf_processor.load_document_index('uploads', doc_name)
if doc_idx:
    ch1 = doc_idx.get_chapter('1')
    if ch1:
        print(f"Chapter 1: title='{ch1.title}', pages={ch1.start_page}-{ch1.end_page}, status={ch1.status}")
        print(f"Key concepts ({len(ch1.key_concepts)}): {ch1.key_concepts[:6]}")
    else:
        print("Chapter 1 NOT found in docindex!")
else:
    print("DocumentIndex NOT found!")

print('\n=== 2. Testing parse_book_scope ===')
scope = retrieval_service.parse_book_scope(q)
print('Scope:', scope)

print('\n=== 3. Testing vector_store_service / retrieval_service ===')
results = retrieval_service.retrieve(
    query=q,
    document_name=doc_name,
    scope=scope,
    timings=timings,
    upload_dir='uploads',
)
print(f'Retrieved {len(results)} chunks:')
for i, r in enumerate(results[:8]):
    print(f'  Chunk {i+1}: page={r.page_number}, score={r.score:.3f}, source={r.source}')
    print(f'    Snippet: {r.text[:120]}...')

print('\n=== 4. Testing rag_service.ask ===')
res = rag_service.ask(
    question=q,
    document_name=doc_name,
    session_id=str(uuid.uuid4()),
    timings=timings,
    upload_dir='uploads',
)
print(f'Sources count: {len(res.sources)}')
for s in res.sources:
    print(f'  Source: page={s.page}, score={s.score:.3f}')
print(f'Answer ({len(res.answer)} chars):')
print(res.answer)