31 lines
1.2 KiB
Python
31 lines
1.2 KiB
Python
"""Pure cache and dataset-freshness policies for offline and real evaluations."""
|
|
|
|
import hashlib
|
|
import json
|
|
|
|
|
|
VALID, NEEDS_REVALIDATION, CONFLICT = "VALID", "NEEDS_REVALIDATION", "CONFLICT"
|
|
|
|
|
|
def cache_key(input_fingerprint, candidate_set_hash, model, prompt_version, dataset_versions):
|
|
"""Return a stable key covering every input that can change an answer."""
|
|
payload = {
|
|
"input_fingerprint": str(input_fingerprint),
|
|
"candidate_set_hash": str(candidate_set_hash),
|
|
"model": str(model),
|
|
"prompt_version": str(prompt_version),
|
|
"dataset_versions": dict(sorted((dataset_versions or {}).items())),
|
|
}
|
|
return hashlib.sha256(json.dumps(payload, sort_keys=True, separators=(",", ":")).encode()).hexdigest()
|
|
|
|
|
|
def classify_knowledge_freshness(record_versions, active_versions):
|
|
"""Classify versioned knowledge without deleting or rewriting history."""
|
|
recorded = dict(record_versions or {})
|
|
active = dict(active_versions or {})
|
|
if any(key in recorded and key in active and recorded[key] != active[key] for key in recorded):
|
|
return CONFLICT
|
|
if recorded != active:
|
|
return NEEDS_REVALIDATION
|
|
return VALID
|