fix: remove background model preload (degraded perf on cpu-basic)

perf: shared cached model + background preload at home page

feat: re-land vocab demo on main via PyPI latincy-vocab

feat: bump la_core_web_{lg,md,sm} to v3.9.6

fix: bump latincy-lexicon to >=0.5.0 (supports Python 3.13)

fix: pin HF Space to Python 3.12

fix: bump la_core_web_md and sm to v3.9.5 (3.9.4 wheels removed from HF)

chore: remove vocab demo from HF Space deployment

feat: bump la_core_web_lg to v3.9.5, update parsing demo scores

feat: drop similarity/morphology/paradigm demos, add vocab demo to home

chore: add macron_morph to home page, bump model version in parsing demo

fix: lowercase entire headword at render time, not just initial

fix: lowercase headword initial in vocab render

fix: remove st.rerun() from sort buttons to preserve checkbox state

feat: add DCC Core filter to vocab demo

fix: remove latincy-words from About tab, attribute glosses to latincy-lexicon

fix: trim default text to first 3 sentences of Ritchie's §1

fix: add citation forms to vocab demo via whitakers_words pipe

fix: vocab demo layout, spacing, and data quality

feat: add vocab builder demo page (latincy-vocab)

Bump latincy-lexicon to PyPI 0.2.0

Simplify sentence picker to a single default example

Add latincy-lexicon demos: lexicon lookup and paradigm generator

Trigger HF Spaces rebuild

Use CPU-only torch to reduce HF Spaces install size

Update dashboard for v3.9 models and fix documentation

Trigger HF Spaces rebuild

Switch la-latincy-lookups to PyPI install

Trigger HF Spaces rebuild

Update lookups package name and bump latincy-preprocess to >=0.2.0

Trigger HF Spaces rebuild

Trigger HF Spaces rebuild

Update lg model to v3.9.0, bump spaCy to 3.8.11

Add normalize_greek preprocessing to diacritics demo

Update diacritics demo: add model attribution, simplify About text

Add latincy-diacritics dependency

Add CITATION.cff

Add MIT license

Fix evaluate tab: colorize missed diacritics, handle case

Add Greek diacritics restorer demo page

Migrate UV and long-s demos to latincy-preprocess package

Add .claude/ and .DS_Store to .gitignore, untrack if present

Add parsing About tab to demos

Add About tab to NER demo

Add About tab to parsing demo

Add About tabs to all 7 remaining Streamlit demos

Add U/V normalizer demo (page 8) and latincy-uv dependency

Update README with current demos and model versions

Update demo texts and fix morph word selector persistence

Adjust similarity results slider to 1-10, default 3