Orchestrata Ultimate › Tests › Reference › orchestrator › test_intent_grammar_accuracy

Tests reference — `tests/orchestrator/test_intent_grammar_accuracy.py`

Module of tests/orchestrator, split out because the directory page would be too large.

Prose pending.

`tests/orchestrator/test_intent_grammar_accuracy.py`

source · 14 tests · 429 lines

Test Standards Compliance (see AGENTS.md § Test Standards):

Covers: core.base.traces, core.paths, infra.prolog_bridge, orchestrator.intent_types, orchestrator.intent_vocabulary, orchestrator.pipeline_router

Kind	Name	Line	Description
function	`test_lexicon_has_no_cross_category_stem_collisions`	103	Two keywords in different categories that share a Porter stem make
function	`test_inflected_forms_reach_the_grammar_tier`	147
function	`test_decisive_word_outranks_generic_verb`	185
function	`test_work_type_is_resolved_by_majority_vote`	218	REQ-003-US-001-065. Measured across the three corpora: leftmost 22/76,
function	`test_generic_verbs_stay_in_the_scoring_tier_registries`	234	The generic-verb map changes GRAMMAR precedence only. A generic verb
function	`test_generic_verbs_round_trip_through_dict`	250
function	`test_generic_verb_defaults_are_real_work_types`	259
function	`test_every_work_type_has_keywords`	271
function	`test_every_target_has_keywords`	281
function	`test_the_four_measured_misleading_keywords_stay_out`	291	`memory`, `trace`, `integration` and `version` each name a SUBSYSTEM
function	`test_grammar_accuracy_on_labelled_corpus`	313	The regression floor. Written alongside the vocabulary, so its score
function	`test_grammar_accuracy_on_heldout_corpus`	355	The generalization estimate. These sentences were written after the
function	`test_grammar_accuracy_on_raw_user_turns`	384	The production-input benchmark: verbatim user turns, typos and all.
function	`test_every_raw_turn_carries_a_label_and_verbatim_text`	415	The corpus is only worth its floor if the text is untouched. A