Build review_data: key = (uid, iid, timestamp) value = {review, title}
(reviews, user2index, item2index)
| 455 | |
| 456 | |
| 457 | def build_review_data_amazon23(reviews, user2index, item2index): |
| 458 | """ |
| 459 | Build review_data: |
| 460 | key = (uid, iid, timestamp) |
| 461 | value = {review, title} |
| 462 | """ |
| 463 | |
| 464 | review_data = {} |
| 465 | |
| 466 | for r in tqdm(reviews, desc="Building review_data"): |
| 467 | user = r["user_id"] |
| 468 | asin = r["asin"] |
| 469 | |
| 470 | if user not in user2index or asin not in item2index: |
| 471 | continue |
| 472 | |
| 473 | uid = user2index[user] |
| 474 | iid = item2index[asin] |
| 475 | ts = r["timestamp"] |
| 476 | |
| 477 | key = str((uid, iid, ts)) |
| 478 | |
| 479 | review_data[key] = { |
| 480 | "review": clean_text(r.get("review_text", "")), |
| 481 | "summary": clean_text(r.get("review_title", "")), |
| 482 | "helpful_votes": r.get("helpful_votes", 0), |
| 483 | "verified": r.get("verified", False), |
| 484 | } |
| 485 | |
| 486 | print(f"[ReviewData] Stored {len(review_data)} review entries") |
| 487 | return review_data |
| 488 | |
| 489 | |
| 490 | def parse_args(): |