matt 505bbdf166 fix: handle numpy arrays in card_similarity parse_theme_tags ... The similarity cache build was failing because parse_theme_tags() was checking isinstance(tags, list) but Parquet files return numpy.ndarray objects. This caused all cards to be flagged as having no theme tags, resulting in an empty cache. Changed to use hasattr(__len__) check instead, which works for both lists and numpy arrays.		2025-10-19 08:26:20 -07:00
..
models	feat(preview): sampling, metrics, governance, server mana data	2025-09-23 09:19:23 -07:00
routes	feat: migrate to unified Parquet format with instant GitHub setup and 4x faster tagging	2025-10-18 21:32:12 -07:00
services	fix: handle numpy arrays in card_similarity parse_theme_tags	2025-10-19 08:26:20 -07:00
static	feat(ui): add similar cards refresh button and reduce sidebar animation distractions	2025-10-17 18:40:15 -07:00
templates	feat: migrate to unified Parquet format with instant GitHub setup and 4x faster tagging	2025-10-18 21:32:12 -07:00
__init__.py	Web UI: setup progress + logs folding, Finished Decks library, commander search UX (debounce, keyboard, highlights, color chips), ranking fixes (first-word priority, substring include), optional auto-select; setup start reliability (POST+GET), force runs, status with percent/ETA/timestamps; stepwise builder with added stage reporting and sidecar summaries; keyboard grid wrap-around; restrict commander search to eligible rows	2025-08-26 09:48:25 -07:00
app.py	Add card browser with similar cards and performance optimizations	2025-10-17 16:17:36 -07:00