From b92918581ef075f8b132f5e3d1992139c0d76252 Mon Sep 17 00:00:00 2001 From: matt Date: Sat, 18 Oct 2025 21:43:04 -0700 Subject: [PATCH] fix: use correct processed/ path for similarity cache building --- code/scripts/build_similarity_cache_parquet.py | 2 +- code/web/services/card_similarity.py | 7 ++++--- 2 files changed, 5 insertions(+), 4 deletions(-) diff --git a/code/scripts/build_similarity_cache_parquet.py b/code/scripts/build_similarity_cache_parquet.py index 1edf924..99d784d 100644 --- a/code/scripts/build_similarity_cache_parquet.py +++ b/code/scripts/build_similarity_cache_parquet.py @@ -155,7 +155,7 @@ def build_cache( """ Build similarity cache for all cards. - NOTE: Assumes card data (cards.csv, all_cards.parquet) and tagged data already exist. + NOTE: Assumes card data (card_files/processed/all_cards.parquet) and tagged data already exist. Run setup and tagging separately before building cache. Args: diff --git a/code/web/services/card_similarity.py b/code/web/services/card_similarity.py index c524da4..4c3e68a 100644 --- a/code/web/services/card_similarity.py +++ b/code/web/services/card_similarity.py @@ -31,12 +31,13 @@ class CardSimilarity: Initialize similarity calculator. Args: - cards_df: DataFrame with card data. If None, loads from all_cards.parquet + cards_df: DataFrame with card data. If None, loads from processed all_cards.parquet cache: SimilarityCache instance. If None, uses global singleton """ if cards_df is None: - # Load from default location - parquet_path = Path(__file__).parents[3] / "card_files" / "all_cards.parquet" + # Load from processed directory (M4 Parquet migration) + from path_util import get_processed_cards_path + parquet_path = get_processed_cards_path() logger.info(f"Loading cards from {parquet_path}") self.cards_df = pd.read_parquet(parquet_path) else: