Summary
LangCacheSemanticCache is part of the public redisvl.extensions.cache.llm API and is covered by tests, but it
is not clearly documented in the public docs beyond the redisvl[langcache] install extra.
Gaps
- No dedicated how-to guide for using LangCache as the LLM cache backend
- Missing from the cache API reference
- Semantic cache docs do not explain when to use
LangCacheSemanticCache vs SemanticCache
- Attribute prerequisites and unsupported operations are not called out
Proposed Work
- Add
LangCacheSemanticCache to docs/api/cache.rst
- Add a task-oriented guide covering installation, initialization, read-through caching, attributes, TTL, async
usage, and delete operations
- Link the guide from
docs/user_guide/how_to_guides/index.md
- Mention the managed-service option in
docs/concepts/extensions.md
Acceptance Criteria
LangCacheSemanticCache appears in the published API reference
- A user can follow a dedicated guide to configure and use the wrapper
- Docs explain that attributes must already be configured in LangCache
- Docs explain current limitations: no direct
vector search, no filter_expression, no update()/aupdate()
- Docs build successfully for the changed pages
Summary
LangCacheSemanticCacheis part of the publicredisvl.extensions.cache.llmAPI and is covered by tests, but itis not clearly documented in the public docs beyond the
redisvl[langcache]install extra.Gaps
LangCacheSemanticCachevsSemanticCacheProposed Work
LangCacheSemanticCachetodocs/api/cache.rstusage, and delete operations
docs/user_guide/how_to_guides/index.mddocs/concepts/extensions.mdAcceptance Criteria
LangCacheSemanticCacheappears in the published API referencevectorsearch, nofilter_expression, noupdate()/aupdate()