7693e97d36
docs/superpowers/specs/2026-09-22-llm-tagging-design.md lays out P35: a versioned hand-curated vocabulary, embedding shortlist + per-theme yes/no judgement with an evidence chunk on the largest live host, resumable docket-scoped runs with state in bib extra_json, write-back that replaces only llm: tags, and a golden-set gate before fan-out. Slice 1: src/llm/vocab/themes.yaml — 53 themes (conversion factor, practice expense, telehealth, care management, behavioral health, Part B drugs, MSSP, QPP, …), each with a label, one-sentence definition, letter-language synonyms and the FR section stems it was seeded from (dev/scripts/llm_vocab_candidates.py mines heading paragraphs from the CY2018+ rules' fr_anchors). llm.vocab.load() validates unique kebab-case slugs, non-empty definitions, integer version ≥ 1 and that retired slugs are not still active; Theme.card is the text embedded per theme; split_tags() separates llm: tags from hand tags for the write-back. stack llm vocab [--path] [--slug].
115 lines
3.6 KiB
Python
115 lines
3.6 KiB
Python
"""llm.vocab — the closed theme vocabulary (#574)."""
|
|
|
|
from __future__ import annotations
|
|
|
|
import pytest
|
|
|
|
from llm.vocab import Theme, Vocab, VocabError, is_llm_tag, load, parse, split_tags
|
|
|
|
GOOD = """
|
|
version: 3
|
|
themes:
|
|
- slug: telehealth
|
|
label: Telehealth
|
|
definition: Medicare telehealth services.
|
|
synonyms: [telehealth, audio-only]
|
|
sections: [Medicare Telehealth Services]
|
|
- slug: care-management
|
|
definition: Monthly care management codes.
|
|
retired: [old-theme]
|
|
"""
|
|
|
|
|
|
class TestParse:
|
|
def test_reads_themes_and_version(self):
|
|
v = parse(GOOD, source="t")
|
|
assert isinstance(v, Vocab) and v.version == 3 and len(v) == 2
|
|
assert v.slugs == ("telehealth", "care-management")
|
|
t = v.get("telehealth")
|
|
assert isinstance(t, Theme) and t.synonyms == ("telehealth", "audio-only")
|
|
assert (
|
|
v.get("care-management").label == "care-management"
|
|
) # label defaults to slug
|
|
assert v.retired == ("old-theme",) and "telehealth" in v and "nope" not in v
|
|
|
|
def test_cards_and_tags(self):
|
|
v = parse(GOOD)
|
|
assert (
|
|
v.cards["telehealth"]
|
|
== "Telehealth. Medicare telehealth services. Also: telehealth, audio-only."
|
|
)
|
|
assert (
|
|
v.cards["care-management"]
|
|
== "care-management. Monthly care management codes."
|
|
)
|
|
assert (
|
|
v.tag("telehealth") == "llm:telehealth"
|
|
and v.get("telehealth").tag == "llm:telehealth"
|
|
)
|
|
|
|
@pytest.mark.parametrize(
|
|
"text, msg",
|
|
[
|
|
(
|
|
"version: 1\nthemes:\n - slug: Bad Slug\n definition: x\n",
|
|
"kebab-case",
|
|
),
|
|
(
|
|
"version: 1\nthemes:\n - slug: a\n definition: ''\n",
|
|
"definition is empty",
|
|
),
|
|
(
|
|
"version: 1\nthemes:\n - slug: a\n definition: x\n - slug: a\n definition: y\n",
|
|
"duplicate slug",
|
|
),
|
|
(
|
|
"version: 0\nthemes:\n - slug: a\n definition: x\n",
|
|
"version must be >= 1",
|
|
),
|
|
("version: x\nthemes: []\n", "version must be an integer"),
|
|
("version: 1\nthemes: []\n", "no themes"),
|
|
("- a\n- b\n", "expected a mapping"),
|
|
(
|
|
"version: 1\nthemes:\n - slug: a\n definition: x\nretired: [a]\n",
|
|
"retired slugs still active",
|
|
),
|
|
],
|
|
)
|
|
def test_rejects_bad_files(self, text, msg):
|
|
with pytest.raises(VocabError, match=msg):
|
|
parse(text, source="t")
|
|
|
|
|
|
class TestPackagedVocab:
|
|
def test_loads_and_is_sane(self):
|
|
v = load()
|
|
assert v.version >= 1 and 40 <= len(v) <= 150
|
|
assert all(t.definition.endswith(".") for t in v.themes)
|
|
assert all(len(t.synonyms) >= 1 for t in v.themes)
|
|
for must in (
|
|
"telehealth",
|
|
"care-management",
|
|
"conversion-factor",
|
|
"practice-expense",
|
|
"palliative-care",
|
|
"skin-substitutes",
|
|
):
|
|
assert must in v
|
|
|
|
def test_load_from_path(self, tmp_path):
|
|
p = tmp_path / "v.yaml"
|
|
p.write_text(GOOD)
|
|
assert load(p).source == str(p)
|
|
|
|
|
|
class TestTagHelpers:
|
|
def test_split(self):
|
|
ours, rest = split_tags(
|
|
["llm:telehealth", "topic:palliative", "llm:care-management", "docket:x"]
|
|
)
|
|
assert ours == ["llm:telehealth", "llm:care-management"] and rest == [
|
|
"topic:palliative",
|
|
"docket:x",
|
|
]
|
|
assert is_llm_tag("llm:x") and not is_llm_tag("llmx:y")
|