Unify evaluation caching: remove redundant adapter-level cache #301
lukedhlee
force pushed
from
4d8d6af1
to
9f9a3251
103 days ago
lukedhlee
force pushed
from
9f9a3251
to
b276da65
103 days ago
Unify evaluation caching into core EvaluationCache, remove adapter-le…
a89b0295
Fix test_refiner.py: remove references to deleted cache_mode parameter
61903574
Add disk write-through to core EvaluationCache, restore cache_evaluat…
7a6d2cfb
Use atomic write for disk cache entries
660340f1
Enable disk cache in gepa.optimize() when run_dir is set
296b3c26
Revert candidate_hash rename — keep _candidate_hash private
d9eff3c8
Cache trajectories in EvaluationCache so Step A benefits from caching
82551cb2
Mock LLM tests: add llm_call marker and auto-mock fixture to prevent …
a7ae1a62
Move LLM mocks into individual test files instead of shared conftest
5043c9b0
Consolidate cache-aware evaluation into evaluate_batch_with_cache()
36397cae
Add end-to-end tests for disk cache, trajectory caching, evaluate_bat…
abcaa82c
lukedhlee
force pushed
from
ac1a8564
to
abcaa82c
103 days ago
Add inline docs for cache_evaluation_storage explaining auto/memory/disk
0a5cde01
Add disk_cache_dir property to EvaluationCache for encapsulation
f4efee91
Remove unused cached_evaluate stub, rename cached_evaluate_full → cac…
a7a24790
Add e2e tests proving caching actually reduces evaluate calls
18e5cae3
Unify EvaluationCache: remove evaluate_with_cache_full, route cached_…
e793ba5b
Merge remote-tracking branch 'origin/main' into refactor/unify-evalua…
e45de923
Fix flaky CI: use deterministic scores in E2E cache test
542a22b1
Login to write a write a comment.
Login via GitHub