phoenix/litellm-mirror
0
0
Fork 1
mirror of https://github.com/BerriAI/litellm.git synced 2025-04-26 11:14:04 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

(feat) redis-semantic cache on proxy

Browse source
This commit is contained in:
ishaan-jaff 2024-02-06 10:35:21 -08:00 • committed by Krrish Dholakia
parent 698cee89fa
commit 4265c7affb
2 changed files with 5 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

5
litellm/proxy/proxy_server.py
View file

@ -1168,7 +1168,7 @@ class ProxyConfig:
verbose_proxy_logger.debug(f"passed cache type={cache_type}") verbose_proxy_logger.debug(f"passed cache type={cache_type}")
if cache_type == "redis": if cache_type == "redis" or cache_type == "redis-semantic":
cache_host = litellm.get_secret("REDIS_HOST", None) cache_host = litellm.get_secret("REDIS_HOST", None)
cache_port = litellm.get_secret("REDIS_PORT", None) cache_port = litellm.get_secret("REDIS_PORT", None)
cache_password = litellm.get_secret("REDIS_PASSWORD", None) cache_password = litellm.get_secret("REDIS_PASSWORD", None)
@ -1195,6 +1195,9 @@ class ProxyConfig:
f"{blue_color_code}Cache Password:{reset_color_code} {cache_password}" f"{blue_color_code}Cache Password:{reset_color_code} {cache_password}"
) )
print() # noqa print() # noqa
if cache_type == "redis-semantic":
# by default this should always be async
cache_params.update({"redis_semantic_cache_use_async": True})
# users can pass os.environ/ variables on the proxy - we should read them from the env # users can pass os.environ/ variables on the proxy - we should read them from the env
for key, value in cache_params.items(): for key, value in cache_params.items():

1
requirements.txt
View file

@ -10,6 +10,7 @@ gunicorn==21.2.0 # server dep
boto3==1.28.58 # aws bedrock/sagemaker calls boto3==1.28.58 # aws bedrock/sagemaker calls
redis==4.6.0 # caching redis==4.6.0 # caching
redisvl==0.0.7 # semantic caching redisvl==0.0.7 # semantic caching
numpy==1.24.3 # semantic caching
prisma==0.11.0 # for db prisma==0.11.0 # for db
mangum==0.17.0 # for aws lambda functions mangum==0.17.0 # for aws lambda functions
google-generativeai==0.3.2 # for vertex ai calls google-generativeai==0.3.2 # for vertex ai calls