# What does this PR do?
In the Responses API, we convert incoming response requests to chat
completion requests. When streaming the resulting chunks of those chat
completion requests, inference providers that use OpenAI clients will
often return a `type=None` value in the tool call parts of the response.
This causes issues when we try to dump and load that response into our
pydantic model, because type cannot be None in the Responses API model
we're loading these into.
So, strip the "type" field, if present, off those chat completion tool
call results before dumping and loading them as our typed pydantic
models, which will apply our default value for that type field.
## Test Plan
This was found via manual testing of the Responses API with codex, where
I was getting errors in some tool call situations. I added a unit test
to simulate this scenario and verify the fix, as well as manual codex
testing to verify the fix.
Signed-off-by: Ben Browning <bbrownin@redhat.com>
# What does this PR do?
This is a combination of what was previously 3 separate PRs - #2069,
#2075, and #2083. It turns out all 3 of those are needed to land a
working function calling Responses implementation. The web search
builtin tool was already working, but this wires in support for custom
function calling.
I ended up combining all three into one PR because they all had lots of
merge conflicts, both with each other but also with #1806 that just
landed. And, because landing any of them individually would have only
left a partially working implementation merged.
The new things added here are:
* Storing of input items from previous responses and restoring of those
input items when adding previous responses to the conversation state
* Handling of multiple input item messages roles, not just "user"
messages.
* Support for custom tools passed into the Responses API to enable
function calling outside of just the builtin websearch tool.
Closes#2074Closes#2080
## Test Plan
### Unit Tests
Several new unit tests were added, and they all pass. Ran via:
```
python -m pytest -s -v tests/unit/providers/agents/meta_reference/test_openai_responses.py
```
### Responses API Verification Tests
I ran our verification run.yaml against multiple providers to ensure we
were getting a decent pass rate. Specifically, I ensured the new custom
tool verification test passed across multiple providers and that the
multi-turn examples passed across at least some of the providers (some
providers struggle with the multi-turn workflows still).
Running the stack setup for verification testing:
```
llama stack run --image-type venv tests/verifications/openai-api-verification-run.yaml
```
Together, passing 100% as an example:
```
pytest -s -v 'tests/verifications/openai_api/test_responses.py' --provider=together-llama-stack
```
## Documentation
We will need to start documenting the OpenAI APIs, but for now the
Responses stuff is still rapidly evolving so delaying that.
---------
Signed-off-by: Derek Higgins <derekh@redhat.com>
Signed-off-by: Ben Browning <bbrownin@redhat.com>
Co-authored-by: Derek Higgins <derekh@redhat.com>
Co-authored-by: Ashwin Bharambe <ashwin.bharambe@gmail.com>
# What does this PR do?
The builtin implementation of code interpreter is not robust and has a
really weak sandboxing shell (the `bubblewrap` container). Given the
availability of better MCP code interpreter servers coming up, we should
use them instead of baking an implementation into the Stack and
expanding the vulnerability surface to the rest of the Stack.
This PR only does the removal. We will add examples with how to
integrate with MCPs in subsequent ones.
## Test Plan
Existing tests.
# What does this PR do?
The goal of this PR is code base modernization.
Schema reflection code needed a minor adjustment to handle UnionTypes
and collections.abc.AsyncIterator. (Both are preferred for latest Python
releases.)
Note to reviewers: almost all changes here are automatically generated
by pyupgrade. Some additional unused imports were cleaned up. The only
change worth of note can be found under `docs/openapi_generator` and
`llama_stack/strong_typing/schema.py` where reflection code was updated
to deal with "newer" types.
Signed-off-by: Ihar Hrachyshka <ihar.hrachyshka@gmail.com>
# What does this PR do?
Add support for the temperature to the responses API
## Test Plan
Manually tested simple case
unit tests added for simple case and tool calls
Signed-off-by: Derek Higgins <derekh@redhat.com>
# What does this PR do?
This provides an initial [OpenAI Responses
API](https://platform.openai.com/docs/api-reference/responses)
implementation. The API is not yet complete, and this is more a
proof-of-concept to show how we can store responses in our key-value
stores and use them to support the Responses API concepts like
`previous_response_id`.
## Test Plan
I've added a new
`tests/integration/openai_responses/test_openai_responses.py` as part of
a test-driven development for this new API. I'm only testing this
locally with the remote-vllm provider for now, but it should work with
any of our inference providers since the only API it requires out of the
inference provider is the `openai_chat_completion` endpoint.
```
VLLM_URL="http://localhost:8000/v1" \
INFERENCE_MODEL="meta-llama/Llama-3.2-3B-Instruct" \
llama stack build --template remote-vllm --image-type venv --run
```
```
LLAMA_STACK_CONFIG="http://localhost:8321" \
python -m pytest -v \
tests/integration/openai_responses/test_openai_responses.py \
--text-model "meta-llama/Llama-3.2-3B-Instruct"
```
---------
Signed-off-by: Ben Browning <bbrownin@redhat.com>
Co-authored-by: Ashwin Bharambe <ashwin.bharambe@gmail.com>
# What does this PR do?
Allow users to name an agent and use the name in telemetry instead of
relying on randomly generated agent_ids. This improves the developer
experience by making it easier to find specific agents in telemetry
logs.
Closes#1832
## Test Plan
- Added tests to verify the agent name is properly stored and retrieved
- Ran `uv run -- pytest -v
tests/integration/telemetry/test_telemetry.py::test_agent_name_filtering`
from the root of the project and made sure the tests pass
- Ran `uv run -- pytest -v
tests/integration/telemetry/test_telemetry.py::test_agent_query_spans`
to verify existing code without agent names still works correctly
## Use Example
```
agent = Agent(
llama_stack_client,
model=text_model_id,
name="CustomerSupportAgent", # New parameter
instructions="You are a helpful customer support assistant"
)
session_id = agent.create_session(f"test-session-{uuid4()}")
```
## Implementation Notes
- Agent names are optional string parameters with no additional
validation
- Names are not required to be unique - multiple agents can have the
same name
- The agent_id remains the unique identifier for an agent
---------
Co-authored-by: raghotham <raghotham@gmail.com>
# What does this PR do?
Move around bits. This makes the copies from llama-models _much_ easier
to maintain and ensures we don't entangle meta-reference specific
tidbits into llama-models code even by accident.
Also, kills the meta-reference-quantized-gpu distro and rolls
quantization deps into meta-reference-gpu.
## Test Plan
```
LLAMA_MODELS_DEBUG=1 \
with-proxy llama stack run meta-reference-gpu \
--env INFERENCE_MODEL=meta-llama/Llama-4-Scout-17B-16E-Instruct \
--env INFERENCE_CHECKPOINT_DIR=<DIR> \
--env MODEL_PARALLEL_SIZE=4 \
--env QUANTIZATION_TYPE=fp8_mixed
```
Start a server with and without quantization. Point integration tests to
it using:
```
pytest -s -v tests/integration/inference/test_text_inference.py \
--stack-config http://localhost:8321 --text-model meta-llama/Llama-4-Scout-17B-16E-Instruct
```
# What does this PR do?
**What**
Instead of adhoc creating a vectordb and chunking when documents ae sent
as an attachment to agent turn, we directly pass raw text from document
into messages to model for user context, and let model perform
summarization directly.
This removes the magic behaviour, and yields better performance than
existing approach.
**Improved Performance**
- RAG lifecycle notebook
- Model: 0.3 factuality score
- (+ websearch) Agent: 0.44 factuality score
- (+ vector db) Agent: 0.3 factuality score
- (+ raw context) Agent: 0.6 factuality score
Closes https://github.com/meta-llama/llama-stack/issues/1478
[//]: # (If resolving an issue, uncomment and update the line below)
[//]: # (Closes #[issue-number])
## Test Plan
- [NEW] added section in RAG lifecycle notebook shows better performance
<img width="840" alt="image"
src="https://github.com/user-attachments/assets/a0c4e816-809a-41c0-9124-89825983e3f5"
/>
[//]: # (## Documentation)
# Summary:
Includes fixes to get test_agents working with openAI model, e.g. tool
parsing and message conversion
# Test Plan:
```
LLAMA_STACK_CONFIG=dev pytest -s -v tests/integration/agents/test_agents.py --safety-shield meta-llama/Llama-Guard-3-8B --text-model openai/gpt-4o-mini
```
---
[//]: # (BEGIN SAPLING FOOTER)
Stack created with [Sapling](https://sapling-scm.com). Best reviewed
with
[ReviewStack](https://reviewstack.dev/meta-llama/llama-stack/pull/1550).
* #1556
* __->__ #1550
# What does this PR do?
Updated all instances of datetime.now() to use timezone.utc for
consistency in handling time across different systems. This ensures that
timestamps are always in Coordinated Universal Time (UTC), avoiding
issues with time zone discrepancies and promoting uniformity in
time-related data.
Signed-off-by: Sébastien Han <seb@redhat.com>
Summary:
This is not used anywhere.
closes#1421
Test Plan:
LLAMA_STACK_CONFIG=fireworks pytest -s -v
tests/integration/agents/test_agents.py --safety-shield
meta-llama/Llama-Guard-3-8B --text-model
meta-llama/Llama-3.1-8B-Instruct --record-responses
Summary:
Refactoring only.
Centralize logic to preprocess toolgroup to one place.
Test Plan:
LLAMA_STACK_CONFIG=fireworks pytest -s -v
tests/api/agents/test_agents.py --safety-shield
meta-llama/Llama-Guard-3-8B
---
[//]: # (BEGIN SAPLING FOOTER)
Stack created with [Sapling](https://sapling-scm.com). Best reviewed
with
[ReviewStack](https://reviewstack.dev/meta-llama/llama-stack/pull/1381).
* #1384
* __->__ #1381
# What does this PR do?
- recent merge https://github.com/meta-llama/llama-stack/pull/1410
introduce error
```
ValueError: Provider meta-reference (Api.agents) does not implement the following methods:
[('list_agent_sessions', 'not_actually_implemented'), ('list_agents', 'not_actually_implemented')]
```
[//]: # (If resolving an issue, uncomment and update the line below)
[//]: # (Closes #[issue-number])
## Test Plan
```
llama stack run
```
```
LLAMA_STACK_CONFIG=fireworks pytest -v tests/integration/agents/test_agents.py --text-model meta-llama/Llama-3.3-70B-Instruct
```
1379530386
[//]: # (## Documentation)
# What does this PR do?
It's a dict that may contain different types, as per
resolver:instantiate_provider implementation. (AFAIU it also never
contains ProviderSpecs, but *instances* of provider implementations.)
[//]: # (If resolving an issue, uncomment and update the line below)
[//]: # (Closes #[issue-number])
## Test Plan
mypy passing if enabled checks for these modules. (See #1543)
[//]: # (## Documentation)
Signed-off-by: Ihar Hrachyshka <ihar.hrachyshka@gmail.com>
# What does this PR do?
This PR has two fixes needed for correct trace context propagation
across asycnio boundary
Fix 1: Start using context vars to store the global trace context.
This is needed since we cannot use the same trace context across
coroutines since the state is shared. each coroutine
should have its own trace context so that each of it can start storing
its state correctly.
Fix 2: Start a new span for each new coroutines started for running
shields to keep the span tree clean
## Test Plan
### Integration tests with server
LLAMA_STACK_DISABLE_VERSION_CHECK=true llama stack run
~/.llama/distributions/together/together-run.yaml
LLAMA_STACK_CONFIG=http://localhost:8321 pytest -s --safety-shield
meta-llama/Llama-Guard-3-8B --text-model
meta-llama/Llama-3.1-8B-Instruct
server logs:
https://gist.github.com/dineshyv/51ac5d9864ed031d0d89ce77352821fe
test logs:
https://gist.github.com/dineshyv/e66acc1c4648a42f1854600609c467f3
### Integration tests with library client
LLAMA_STACK_CONFIG=fireworks pytest -s --safety-shield
meta-llama/Llama-Guard-3-8B --text-model
meta-llama/Llama-3.1-8B-Instruct
logs: https://gist.github.com/dineshyv/ca160696a0b167223378673fb1dcefb8
### Apps test with server:
```
LLAMA_STACK_DISABLE_VERSION_CHECK=true llama stack run ~/.llama/distributions/together/together-run.yaml
python -m examples.agents.e2e_loop_with_client_tools localhost 8321
```
server logs:
https://gist.github.com/dineshyv/1717a572d8f7c14279c36123b79c5797
app logs:
https://gist.github.com/dineshyv/44167e9f57806a0ba3b710c32aec02f8
Summary:
| File
"/Users/erichuang/projects/llama-stack/llama_stack/distribution/server/server.py",
line 213, in sse_generator
| logger.exception(f"Error in sse_generator: {e}")
| File
"/opt/homebrew/Caskroom/miniconda/base/envs/myenv/lib/python3.10/logging/__init__.py",
line 1864, in exception
| self.log(ERROR, msg, *args, exc_info=exc_info, **kwargs)
| File
"/opt/homebrew/Caskroom/miniconda/base/envs/myenv/lib/python3.10/logging/__init__.py",
line 1879, in log
| self.logger.log(level, msg, *args, **kwargs)
| File
"/opt/homebrew/Caskroom/miniconda/base/envs/myenv/lib/python3.10/logging/__init__.py",
line 1547, in log
| self._log(level, msg, args, **kwargs)
| File
"/opt/homebrew/Caskroom/miniconda/base/envs/myenv/lib/python3.10/logging/__init__.py",
line 1624, in _log
| self.handle(record)
| File
"/opt/homebrew/Caskroom/miniconda/base/envs/myenv/lib/python3.10/logging/__init__.py",
line 1634, in handle
| self.callHandlers(record)
| File
"/opt/homebrew/Caskroom/miniconda/base/envs/myenv/lib/python3.10/logging/__init__.py",
line 1696, in callHandlers
| hdlr.handle(record)
| File
"/opt/homebrew/Caskroom/miniconda/base/envs/myenv/lib/python3.10/logging/__init__.py",
line 968, in handle
| self.emit(record)
| File
"/opt/homebrew/Caskroom/miniconda/base/envs/myenv/lib/python3.10/site-packages/rich/logging.py",
line 167, in emit
| message_renderable = self.render_message(record, message)
| File
"/opt/homebrew/Caskroom/miniconda/base/envs/myenv/lib/python3.10/site-packages/rich/logging.py",
line 193, in render_message
| message_text = Text.from_markup(message) if use_markup else
Text(message)
| File
"/opt/homebrew/Caskroom/miniconda/base/envs/myenv/lib/python3.10/site-packages/rich/text.py",
line 287, in from_markup
| rendered_text = render(text, style, emoji=emoji,
emoji_variant=emoji_variant)
| File
"/opt/homebrew/Caskroom/miniconda/base/envs/myenv/lib/python3.10/site-packages/rich/markup.py",
line 167, in render
| raise MarkupError(
| rich.errors.MarkupError: closing tag '[/INST]' at position 105 doesn't
match any open tag
Test Plan:
reran failing rag_with_vector_db example
Summary:
error:
llama_stack/providers/inline/agents/meta_reference/agent_instance.py:1032:
in execute_tool_call_maybe
logger.info(f"tool call {name} completed with result: {result}")
/opt/homebrew/Caskroom/miniconda/base/envs/myenv/lib/python3.10/logging/__init__.py:1841:
in info
self.log(INFO, msg, *args, **kwargs)
/opt/homebrew/Caskroom/miniconda/base/envs/myenv/lib/python3.10/logging/__init__.py:1879:
in log
self.logger.log(level, msg, *args, **kwargs)
/opt/homebrew/Caskroom/miniconda/base/envs/myenv/lib/python3.10/logging/__init__.py:1547:
in log
self._log(level, msg, args, **kwargs)
/opt/homebrew/Caskroom/miniconda/base/envs/myenv/lib/python3.10/logging/__init__.py:1624:
in _log
self.handle(record)
/opt/homebrew/Caskroom/miniconda/base/envs/myenv/lib/python3.10/logging/__init__.py:1634:
in handle
self.callHandlers(record)
/opt/homebrew/Caskroom/miniconda/base/envs/myenv/lib/python3.10/logging/__init__.py:1696:
in callHandlers
hdlr.handle(record)
/opt/homebrew/Caskroom/miniconda/base/envs/myenv/lib/python3.10/logging/__init__.py:968:
in handle
self.emit(record)
/opt/homebrew/Caskroom/miniconda/base/envs/myenv/lib/python3.10/site-packages/rich/logging.py:167:
in emit
message_renderable = self.render_message(record, message)
/opt/homebrew/Caskroom/miniconda/base/envs/myenv/lib/python3.10/site-packages/rich/logging.py:193:
in render_message
message_text = Text.from_markup(message) if use_markup else
Text(message)
/opt/homebrew/Caskroom/miniconda/base/envs/myenv/lib/python3.10/site-packages/rich/text.py:287:
in from_markup
rendered_text = render(text, style, emoji=emoji,
emoji_variant=emoji_variant)
/opt/homebrew/Caskroom/miniconda/base/envs/myenv/lib/python3.10/site-packages/rich/markup.py:167:
in render
raise MarkupError(
E rich.errors.MarkupError: closing tag '[/INST]' at position 3274
doesn't match any open tag
Test Plan:
# What does this PR do?
This commit introduces a new logging system that allows loggers to be
assigned
a category while retaining the logger name based on the file name. The
log
format includes both the logger name and the category, producing output
like:
```
INFO 2025-03-03 21:44:11,323 llama_stack.distribution.stack:103 [core]: Tool_groups: builtin::websearch served by
tavily-search
```
Key features include:
- Category-based logging: Loggers can be assigned a category (e.g.,
"core", "server") when programming. The logger can be loaded like
this: `logger = get_logger(name=__name__, category="server")`
- Environment variable control: Log levels can be configured
per-category using the
`LLAMA_STACK_LOGGING` environment variable. For example:
`LLAMA_STACK_LOGGING="server=DEBUG;core=debug"` enables DEBUG level for
the "server"
and "core" categories.
- `LLAMA_STACK_LOGGING="all=debug"` sets DEBUG level globally for all
categories and
third-party libraries.
This provides fine-grained control over logging levels while maintaining
a clean and
informative log format.
The formatter uses the rich library which provides nice colors better
stack traces like so:
```
ERROR 2025-03-03 21:49:37,124 asyncio:1758 [uncategorized]: unhandled exception during asyncio.run() shutdown
task: <Task finished name='Task-16' coro=<handle_signal.<locals>.shutdown() done, defined at
/Users/leseb/Documents/AI/llama-stack/llama_stack/distribution/server/server.py:146>
exception=UnboundLocalError("local variable 'loop' referenced before assignment")>
╭────────────────────────────────────── Traceback (most recent call last) ───────────────────────────────────────╮
│ /Users/leseb/Documents/AI/llama-stack/llama_stack/distribution/server/server.py:178 in shutdown │
│ │
│ 175 │ │ except asyncio.CancelledError: │
│ 176 │ │ │ pass │
│ 177 │ │ finally: │
│ ❱ 178 │ │ │ loop.stop() │
│ 179 │ │
│ 180 │ loop = asyncio.get_running_loop() │
│ 181 │ loop.create_task(shutdown()) │
╰────────────────────────────────────────────────────────────────────────────────────────────────────────────────╯
UnboundLocalError: local variable 'loop' referenced before assignment
```
Co-authored-by: Ashwin Bharambe <@ashwinb>
Signed-off-by: Sébastien Han <seb@redhat.com>
[//]: # (If resolving an issue, uncomment and update the line below)
[//]: # (Closes #[issue-number])
## Test Plan
```
python -m llama_stack.distribution.server.server --yaml-config ./llama_stack/templates/ollama/run.yaml
INFO 2025-03-03 21:55:35,918 __main__:365 [server]: Using config file: llama_stack/templates/ollama/run.yaml
INFO 2025-03-03 21:55:35,925 __main__:378 [server]: Run configuration:
INFO 2025-03-03 21:55:35,928 __main__:380 [server]: apis:
- agents
```
[//]: # (## Documentation)
---------
Signed-off-by: Sébastien Han <seb@redhat.com>
Co-authored-by: Ashwin Bharambe <ashwin.bharambe@gmail.com>
You now run the integration tests with these options:
```bash
Custom options:
--stack-config=STACK_CONFIG
a 'pointer' to the stack. this can be either be:
(a) a template name like `fireworks`, or
(b) a path to a run.yaml file, or
(c) an adhoc config spec, e.g.
`inference=fireworks,safety=llama-guard,agents=meta-
reference`
--env=ENV Set environment variables, e.g. --env KEY=value
--text-model=TEXT_MODEL
comma-separated list of text models. Fixture name:
text_model_id
--vision-model=VISION_MODEL
comma-separated list of vision models. Fixture name:
vision_model_id
--embedding-model=EMBEDDING_MODEL
comma-separated list of embedding models. Fixture name:
embedding_model_id
--safety-shield=SAFETY_SHIELD
comma-separated list of safety shields. Fixture name:
shield_id
--judge-model=JUDGE_MODEL
comma-separated list of judge models. Fixture name:
judge_model_id
--embedding-dimension=EMBEDDING_DIMENSION
Output dimensionality of the embedding model to use for
testing. Default: 384
--record-responses Record new API responses instead of using cached ones.
--report=REPORT Path where the test report should be written, e.g.
--report=/path/to/report.md
```
Importantly, if you don't specify any of the models (text-model,
vision-model, etc.) the relevant tests will get **skipped!**
This will make running tests somewhat more annoying since all options
will need to be specified. We will make this easier by adding some easy
wrapper yaml configs.
## Test Plan
Example:
```bash
ashwin@ashwin-mbp ~/local/llama-stack/tests/integration (unify_tests) $
LLAMA_STACK_CONFIG=fireworks pytest -s -v inference/test_text_inference.py \
--text-model meta-llama/Llama-3.2-3B-Instruct
```
# Summary:
Client side change in
https://github.com/meta-llama/llama-stack-client-python/pull/180
Changes the resume_turn API to accept `ToolResponse` instead of
`ToolResponseMessage`:
1. `ToolResponse` contains `metadata`
2. `ToolResponseMessage` is a concept for model inputs. Here we are just
submitting the outputs of tool execution.
# Test Plan:
Ran integration tests with newly added test using client tool with
metadata
LLAMA_STACK_CONFIG=fireworks pytest -s -v
tests/integration/agents/test_agents.py --safety-shield
meta-llama/Llama-Guard-3-8B --record-responses
# What does this PR do?
The agent API allows to query multiple DBs using the `vector_db_ids`
argument of the `rag` tool:
```py
toolgroups=[
{
"name": "builtin::rag",
"args": {"vector_db_ids": [vector_db_id]},
}
],
```
This means that multiple DBs can be used to compose an aggregated
context by executing the query on each of them.
When documents are passed to the next agent turn, there is no explicit
way to configure the vector DB where the embeddings will be ingested. In
such cases, we can assume that:
- if any `vector_db_ids` is given, we use the first one (it probably
makes sense to assume that it's the only one in the list, otherwise we
should loop on all the given DBs to have a consistent ingestion)
- if no `vector_db_ids` is given, we can use the current logic to
generate a default DB using the default provider. If multiple providers
are defined, the API will fail as expected: the user has to provide
details on where to ingest the documents.
(Closes#1270)
## Test Plan
The issue description details how to replicate the problem.
[//]: # (## Documentation)
---------
Signed-off-by: Daniele Martinoli <dmartino@redhat.com>
Summary:
Test Plan:
added new test
LLAMA_STACK_CONFIG=fireworks pytest -s -v
tests/api/agents/test_agents.py --safety-shield
meta-llama/Llama-Guard-3-8B
# What does this PR do?
- Deprecate allow_turn_resume flag as this is used for staying backward
compat.
- Closes https://github.com/meta-llama/llama-stack/issues/1363
[//]: # (If resolving an issue, uncomment and update the line below)
[//]: # (Closes #[issue-number])
## Test Plan
```
LLAMA_STACK_CONFIG=fireworks pytest -v tests/api/agents/test_agents.py --inference-model "meta-llama/Llama-3.3-70B-Instruct" --record-responses
```
<img width="1054" alt="image"
src="https://github.com/user-attachments/assets/d31de2d4-0953-41e1-a71a-7e1579fa351a"
/>
[//]: # (## Documentation)
Summary:
1. The `tools` parameter we construct to pass the inference API is
non-deterministic. As a result, our recordable mocks is flaky as the
ordering change sometimes. This PR makes it so that `tools` ordering is
deterministic and aligned with the order user specified.
2. In recordable mock key generation, client tool's parameter type was
'str' and now is 'string' for some reason. I didn't dig into exactly
why, but just regenerated the fixtures.
Test Plan:
Regenerate mocks:
```
LLAMA_STACK_CONFIG=fireworks pytest -s -v tests/client-sdk/agents/test_agents.py --safety-shield meta-llama/Llama-Guard-3-8B --record-responses
```
Rerun tests without --record-responses:
```
LLAMA_STACK_CONFIG=fireworks pytest -s -v tests/client-sdk/agents/test_agents.py --safety-shield meta-llama/Llama-Guard-3-8B
```
# What does this PR do?
We currently use `max_infer_iters` in 2 different ways
1/ Server: track number of times
2/ Client side: track number of times we send `resume_turn` request
This PR gets rid of the need of (2) and makes server track total number
of times we perform inference within a Turn
**NOTE**
The PR will assume StopReason is set to
- end_of_message: turn is not finished, we could be waiting for client
tool call responses
- end_of_turn: if the entire turn is finished and there's no more things
to be done.
[//]: # (If resolving an issue, uncomment and update the line below)
[//]: # (Closes #[issue-number])
## Test Plan
```
LLAMA_STACK_BASE_URL=http://localhost:8321 pytest -v tests/client-sdk/agents/test_agents.py::test_custom_tool_infinite_loop --inference-model "meta-llama/Llama-3.3-70B-Instruct"
```
[//]: # (## Documentation)
Original telemetry outputs for agent turns look like this.
Note: how output was a `str(message)` making it difficult to read them
back for downstream tasks ( eg. building eval datasets )
```
{
│ │ 'input': [
│ │ │ '{"role":"system","content":"You are a helpful assistant. Use search tool to answer the questions. "}',
│ │ │ '{"role":"user","content":"Which teams played in the NBA western conference finals of 2024","context":null}'
│ │ ],
│ │ 'output': "content: tool_calls: [ToolCall(call_id='8b7294ec-a83f-4798-ad8f-6bed662f08b6', tool_name=<BuiltinTool.brave_search: 'brave_search'>, arguments={'query': 'NBA Western Conference Finals 2024 teams'})]"
│ },
```
Updated the outputs to be structured .
## Test
```python
import uuid
from llama_stack_client.lib.agents.agent import Agent
from llama_stack_client.lib.agents.event_logger import EventLogger
from llama_stack_client.types.agent_create_params import AgentConfig
model_id = "meta-llama/Llama-3.1-8B-Instruct"
agent_config = AgentConfig(
model=model_id,
instructions="You are a helpful assistant who will use the web search tools to help with answering questions.\nOnly provide final answer in short without writing full sentences. Use web search",
toolgroups=["builtin::websearch"],
enable_session_persistence=True,
)
agent = Agent(client, agent_config)
session_id = agent.create_session(uuid.uuid4().hex)
response = agent.create_turn(
messages=[
{
"role": "user",
"content": "latest news about llama stack",
}
],
session_id=session_id,
stream=False,
)
pprint(response)
```
Output:
```
Turn(
│ input_messages=[UserMessage(content='latest news about llama stack', role='user', context=None)],
│ output_message=CompletionMessage(
│ │ content="The latest news about Llama Stack is that Meta has released Llama 3.2, which includes small and medium-sized vision LLMs (11B and 90B) and lightweight, text-only models (1B and 3B) that fit onto select edge and mobile devices. Additionally, Llama Stack distributions have been released to simplify the way developers work with Llama models in different environments. However, a critical vulnerability has been discovered in Meta's Llama-Stack, which puts AI applications at risk.",
│ │ role='assistant',
│ │ stop_reason='end_of_turn',
│ │ tool_calls=[]
│ ),
│ session_id='77379546-4598-485a-b4f4-84e5da28c513',
│ started_at=datetime.datetime(2025, 2, 27, 11, 2, 43, 915243, tzinfo=TzInfo(-08:00)),
│ steps=[
│ │ InferenceStep(
│ │ │ api_model_response=CompletionMessage(
│ │ │ │ content='',
│ │ │ │ role='assistant',
│ │ │ │ stop_reason='end_of_turn',
│ │ │ │ tool_calls=[
│ │ │ │ │ ToolCall(
│ │ │ │ │ │ arguments={'query': 'latest news llama stack'},
│ │ │ │ │ │ call_id='84c0fa10-e24a-4f91-a9ff-415a9ec0bb0b',
│ │ │ │ │ │ tool_name='brave_search'
│ │ │ │ │ )
│ │ │ │ ]
│ │ │ ),
│ │ │ step_id='81c16bd3-eb00-4721-8edc-f386e07391a3',
│ │ │ step_type='inference',
│ │ │ turn_id='2c6b5273-4b16-404f-bed2-c0025fd63b45',
│ │ │ completed_at=datetime.datetime(2025, 2, 27, 11, 2, 44, 637149, tzinfo=TzInfo(-08:00)),
│ │ │ started_at=datetime.datetime(2025, 2, 27, 11, 2, 43, 915831, tzinfo=TzInfo(-08:00))
│ │ ),
│ │ ToolExecutionStep(
│ │ │ step_id='4782d609-a62e-45f5-8d2a-25a43db46288',
│ │ │ step_type='tool_execution',
│ │ │ tool_calls=[
│ │ │ │ ToolCall(
│ │ │ │ │ arguments={'query': 'latest news llama stack'},
│ │ │ │ │ call_id='84c0fa10-e24a-4f91-a9ff-415a9ec0bb0b',
│ │ │ │ │ tool_name='brave_search'
│ │ │ │ )
│ │ │ ],
│ │ │ tool_responses=[
│ │ │ │ ToolResponse(
│ │ │ │ │ call_id='84c0fa10-e24a-4f91-a9ff-415a9ec0bb0b',
│ │ │ │ │ content='{"query": "latest news llama stack", "top_k": [{"title": "Llama 3.2: Revol. ....... Hacker News.", "score": 0.6186197, "raw_content": null}]}',
│ │ │ │ │ tool_name='brave_search',
│ │ │ │ │ metadata=None
│ │ │ │ )
│ │ │ ],
│ │ │ turn_id='2c6b5273-4b16-404f-bed2-c0025fd63b45',
│ │ │ completed_at=datetime.datetime(2025, 2, 27, 11, 2, 46, 272176, tzinfo=TzInfo(-08:00)),
│ │ │ started_at=datetime.datetime(2025, 2, 27, 11, 2, 44, 640743, tzinfo=TzInfo(-08:00))
│ │ ),
│ │ InferenceStep(
│ │ │ api_model_response=CompletionMessage(
│ │ │ │ content="The latest news about Llama Stack is that Meta has released Llama 3.2, which includes small and medium-sized vision LLMs (11B and 90B) and lightweight, text-only models (1B and 3B) that fit onto select edge and mobile devices. Additionally, Llama Stack distributions have been released to simplify the way developers work with Llama models in different environments. However, a critical vulnerability has been discovered in Meta's Llama-Stack, which puts AI applications at risk.",
│ │ │ │ role='assistant',
│ │ │ │ stop_reason='end_of_turn',
│ │ │ │ tool_calls=[]
│ │ │ ),
│ │ │ step_id='37994419-5da3-4e84-a010-8d9b85366262',
│ │ │ step_type='inference',
│ │ │ turn_id='2c6b5273-4b16-404f-bed2-c0025fd63b45',
│ │ │ completed_at=datetime.datetime(2025, 2, 27, 11, 2, 48, 961275, tzinfo=TzInfo(-08:00)),
│ │ │ started_at=datetime.datetime(2025, 2, 27, 11, 2, 46, 273168, tzinfo=TzInfo(-08:00))
│ │ )
│ ],
│ turn_id='2c6b5273-4b16-404f-bed2-c0025fd63b45',
│ completed_at=datetime.datetime(2025, 2, 27, 11, 2, 48, 962318, tzinfo=TzInfo(-08:00)),
│ output_attachments=[]
)
```
## Check for Telemetry
```python
agent_logs = []
for span in client.telemetry.query_spans(
attribute_filters=[
{"key": "session_id", "op": "eq", "value": session_id},
],
attributes_to_return=['input', 'output'],
):
agent_logs.append(span.attributes)
pprint(json.loads(agent_logs[-1]['output']))
```
```
{
│ 'content': "The latest news about Llama Stack is that Meta has released Llama 3.2, which includes small and medium-sized vision LLMs (11B and 90B) and lightweight, text-only models (1B and 3B) that fit onto select edge and mobile devices. Additionally, Llama Stack distributions have been released to simplify the way developers work with Llama models in different environments. However, a critical vulnerability has been discovered in Meta's Llama-Stack, which puts AI applications at risk.",
│ 'tool_calls': []
}
```
# Summary:
Right now we would include toolgroup args when we encode messages with
tool_calls, which is confusing the model since they not in the function
description (see test plan for example).
# Test Plan:
Add a print statement before raw prompt is sent to providers (no good
way to test this currently)
Before:
```
cated in the same neighborhood?<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n[knowledge_search(query="Laleli Mosque and Esma Sultan Mansion same neighborhood", vector_db_ids=["829a68735d744dc3830409dcc782964a"])]<|eot_id|><|start_header_id|>ipython<|end_header_id|>\n\nknowledge_search tool found 5 chunks:\nBEGIN of
```
Note the extra `vector_db_ids`
After
```
>user<|end_header_id|>\n\nAre the Laleli Mosque and Esma Sultan Mansion located in the same neighborhood?<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n[knowledge_search(query="Laleli Mosque and Esma Sultan Mansion same neighborhood")]<|eot_id|><|start_header_id|>ipython<|end_header_id|>\n\nknowledge_search tool found
```
Summary:
Lets the model decide which tool it needs to call to respond to a query.
Test Plan:
```
LLAMA_STACK_CONFIG=fireworks pytest -s -v tests/client-sdk/ --safety-shield meta-llama/Llama-Guard-3-8B
```
Also evaluated on a small benchmark with 20 questions from HotpotQA.
With this PR and some prompting, the performance is 77% recall compared
to 50% currently.
---
[//]: # (BEGIN SAPLING FOOTER)
Stack created with [Sapling](https://sapling-scm.com). Best reviewed
with
[ReviewStack](https://reviewstack.dev/meta-llama/llama-stack/pull/1015).
* #1268
* #1239
* __->__ #1015
