Commit graph

43 commits

Author SHA1 Message Date
Ashwin Bharambe
6e5ca1350e bug fixes to make this work, trace creation worked - spans dont yet 2024-09-22 22:02:25 -07:00
Ashwin Bharambe
84ebed9c9f opentelemetry -> jaeger 2024-09-22 21:05:33 -07:00
Ashwin Bharambe
5d75c2437b enhance the tracing span utility to make a context manager 2024-09-22 20:55:15 -07:00
Ashwin Bharambe
484dc2e4f5 nuke safety/list_shields, we don't need it now 2024-09-22 19:49:08 -07:00
Ashwin Bharambe
ef042e1605 Add more sample content 2024-09-22 17:09:03 -07:00
Ashwin Bharambe
6cbc387851 Add sample adapters 2024-09-22 16:44:57 -07:00
Ashwin Bharambe
b153a67a3e refactor persistence into another file 2024-09-22 16:34:46 -07:00
Ashwin Bharambe
cd4880126b kill with_safety(), not needed 2024-09-22 16:34:46 -07:00
Ashwin Bharambe
59f1fe5af8 Agent persistence works 2024-09-22 16:34:46 -07:00
Ashwin Bharambe
4eb0f30891 initial cut at using kvstores for agent persistence 2024-09-22 16:34:44 -07:00
Ashwin Bharambe
61974e337f kvstore impls for redis / sqlite moved 2024-09-22 16:33:55 -07:00
Ashwin Bharambe
c1ab66f1e6
Further generalize Xi's changes (#88)
* Further generalize Xi's changes

- introduce a slightly more general notion of an AutoRouted provider
- the AutoRouted provider is associated with a RoutingTable provider
- e.g. inference -> models
- Introduced safety -> shields and memory -> memory_banks
  correspondences

* typo

* Basic build and run succeeded
2024-09-22 16:31:18 -07:00
Xi Yan
b8914bb56f add safety/list_shields to query available shields 2024-09-22 01:08:32 -07:00
Xi Yan
b5217fe6fc fix configure 2024-09-22 00:19:24 -07:00
Hardik Shah
8e757ed274 Respect user sent instructions in agent config and add them to system prompt 2024-09-22 00:05:17 -07:00
Xi Yan
0348f26e00 models endpoint testing 2024-09-22 00:01:35 -07:00
Xi Yan
c0199029e5 supported models wip 2024-09-21 18:37:22 -07:00
Xi Yan
f0580251a3 memory routers working 2024-09-21 16:40:23 -07:00
Xi Yan
85d927adde skeleton unified routing table, api routers 2024-09-21 13:44:33 -07:00
Xi Yan
39c27a3d8c Revert "migrate router for memory wip"
This reverts commit 08379f5214.
2024-09-21 12:46:02 -07:00
Xi Yan
665ab1f812 Revert "delete router from providers"
This reverts commit d8fab77a4f.
2024-09-21 12:46:02 -07:00
Xi Yan
ee77431b64 Revert "clean up"
This reverts commit bc4ac2ceb4.
2024-09-21 12:46:01 -07:00
Ashwin Bharambe
a57411b4b3 Further bug fixes 2024-09-21 12:46:01 -07:00
Ashwin Bharambe
9252e81a7b test safety against safety client 2024-09-21 12:46:01 -07:00
Ashwin Bharambe
d6a41d98d2 Update safety implementation inside agents 2024-09-21 12:46:01 -07:00
Ashwin Bharambe
82ddd851c8 Update the meta reference safety implementation to match new API 2024-09-21 12:46:01 -07:00
Ashwin Bharambe
7e40eead4e safety API cleanup part 1
Sample adapter implementation for Bedrock implementation of Guardrails
2024-09-21 12:46:01 -07:00
Ashwin Bharambe
32beecb20d Add a special header per-client call to parser provider data 2024-09-21 12:46:00 -07:00
Xi Yan
3787408863 clean up 2024-09-20 13:02:10 -07:00
Xi Yan
9c33587003 delete router from providers 2024-09-20 12:21:44 -07:00
Xi Yan
cda61119ce migrate router for memory wip 2024-09-20 12:19:33 -07:00
Hardik Shah
8fa49593e0
Allow TGI adaptor to have non-standard llama model names (#84)
Co-authored-by: Hardik Shah <hjshah@fb.com>
2024-09-19 21:42:15 -07:00
Hardik Shah
42d29f3a5a Allow TGI adaptor to have non-standard llama model names 2024-09-19 21:37:02 -07:00
Xi Yan
59af1c8fec
fix memory url parsing (#81) 2024-09-19 13:35:03 -07:00
Ashwin Bharambe
132f9429b1 Add a test for CLI, but not fully done so disabled 2024-09-19 13:27:07 -07:00
Ashwin Bharambe
8b3ffa33de Add another test case 2024-09-19 13:02:57 -07:00
Ashwin Bharambe
abb43936ab Add a test runner and 2 very simple tests for agents 2024-09-19 12:22:48 -07:00
Ashwin Bharambe
f5eda1decf Add default for max_seq_len 2024-09-18 21:59:10 -07:00
Ashwin Bharambe
8cdc2f0cfb No RunShieldRequest 2024-09-18 20:38:21 -07:00
Xi Yan
e6fdb9df29
fix context retriever (#75) 2024-09-18 08:24:36 -07:00
Ashwin Bharambe
9fd431e710 make shield imports more lazy 2024-09-17 21:27:37 -07:00
Ashwin Bharambe
25adc83de8 Fix for safety 2024-09-17 19:56:58 -07:00
Ashwin Bharambe
9487ad8294
API Updates (#73)
* API Keys passed from Client instead of distro configuration

* delete distribution registry

* Rename the "package" word away

* Introduce a "Router" layer for providers

Some providers need to be factorized and considered as thin routing
layers on top of other providers. Consider two examples:

- The inference API should be a routing layer over inference providers,
  routed using the "model" key
- The memory banks API is another instance where various memory bank
  types will be provided by independent providers (e.g., a vector store
  is served by Chroma while a keyvalue memory can be served by Redis or
  PGVector)

This commit introduces a generalized routing layer for this purpose.

* update `apis_to_serve`

* llama_toolchain -> llama_stack

* Codemod from llama_toolchain -> llama_stack

- added providers/registry
- cleaned up api/ subdirectories and moved impls away
- restructured api/api.py
- from llama_stack.apis.<api> import foo should work now
- update imports to do llama_stack.apis.<api>
- update many other imports
- added __init__, fixed some registry imports
- updated registry imports
- create_agentic_system -> create_agent
- AgenticSystem -> Agent

* Moved some stuff out of common/; re-generated OpenAPI spec

* llama-toolchain -> llama-stack (hyphens)

* add control plane API

* add redis adapter + sqlite provider

* move core -> distribution

* Some more toolchain -> stack changes

* small naming shenanigans

* Removing custom tool and agent utilities and moving them client side

* Move control plane to distribution server for now

* Remove control plane from API list

* no codeshield dependency randomly plzzzzz

* Add "fire" as a dependency

* add back event loggers

* stack configure fixes

* use brave instead of bing in the example client

* add init file so it gets packaged

* add init files so it gets packaged

* Update MANIFEST

* bug fix

---------

Co-authored-by: Hardik Shah <hjshah@fb.com>
Co-authored-by: Xi Yan <xiyan@meta.com>
Co-authored-by: Ashwin Bharambe <ashwin@meta.com>
2024-09-17 19:51:35 -07:00