llama-stack-mirror/llama_stack/providers/utils/telemetry/tracing.py
Xi Yan 3c72c034e6
[remove import *] clean up import *'s (#689)
# What does this PR do?

- as title, cleaning up `import *`'s
- upgrade tests to make them more robust to bad model outputs
- remove import *'s in llama_stack/apis/* (skip __init__ modules)
<img width="465" alt="image"
src="https://github.com/user-attachments/assets/d8339c13-3b40-4ba5-9c53-0d2329726ee2"
/>

- run `sh run_openapi_generator.sh`, no types gets affected

## Test Plan

### Providers Tests

**agents**
```
pytest -v -s llama_stack/providers/tests/agents/test_agents.py -m "together" --safety-shield meta-llama/Llama-Guard-3-8B --inference-model meta-llama/Llama-3.1-405B-Instruct-FP8
```

**inference**
```bash
# meta-reference
torchrun $CONDA_PREFIX/bin/pytest -v -s -k "meta_reference" --inference-model="meta-llama/Llama-3.1-8B-Instruct" ./llama_stack/providers/tests/inference/test_text_inference.py
torchrun $CONDA_PREFIX/bin/pytest -v -s -k "meta_reference" --inference-model="meta-llama/Llama-3.2-11B-Vision-Instruct" ./llama_stack/providers/tests/inference/test_vision_inference.py

# together
pytest -v -s -k "together" --inference-model="meta-llama/Llama-3.1-8B-Instruct" ./llama_stack/providers/tests/inference/test_text_inference.py
pytest -v -s -k "together" --inference-model="meta-llama/Llama-3.2-11B-Vision-Instruct" ./llama_stack/providers/tests/inference/test_vision_inference.py

pytest ./llama_stack/providers/tests/inference/test_prompt_adapter.py 
```

**safety**
```
pytest -v -s llama_stack/providers/tests/safety/test_safety.py -m together --safety-shield meta-llama/Llama-Guard-3-8B
```

**memory**
```
pytest -v -s llama_stack/providers/tests/memory/test_memory.py -m "sentence_transformers" --env EMBEDDING_DIMENSION=384
```

**scoring**
```
pytest -v -s -m llm_as_judge_scoring_together_inference llama_stack/providers/tests/scoring/test_scoring.py --judge-model meta-llama/Llama-3.2-3B-Instruct
pytest -v -s -m basic_scoring_together_inference llama_stack/providers/tests/scoring/test_scoring.py
pytest -v -s -m braintrust_scoring_together_inference llama_stack/providers/tests/scoring/test_scoring.py
```


**datasetio**
```
pytest -v -s -m localfs llama_stack/providers/tests/datasetio/test_datasetio.py
pytest -v -s -m huggingface llama_stack/providers/tests/datasetio/test_datasetio.py
```


**eval**
```
pytest -v -s -m meta_reference_eval_together_inference llama_stack/providers/tests/eval/test_eval.py
pytest -v -s -m meta_reference_eval_together_inference_huggingface_datasetio llama_stack/providers/tests/eval/test_eval.py
```

### Client-SDK Tests
```
LLAMA_STACK_BASE_URL=http://localhost:5000 pytest -v ./tests/client-sdk
```

### llama-stack-apps
```
PORT=5000
LOCALHOST=localhost

python -m examples.agents.hello $LOCALHOST $PORT
python -m examples.agents.inflation $LOCALHOST $PORT
python -m examples.agents.podcast_transcript $LOCALHOST $PORT
python -m examples.agents.rag_as_attachments $LOCALHOST $PORT
python -m examples.agents.rag_with_memory_bank $LOCALHOST $PORT
python -m examples.safety.llama_guard_demo_mm $LOCALHOST $PORT
python -m examples.agents.e2e_loop_with_custom_tools $LOCALHOST $PORT

# Vision model
python -m examples.interior_design_assistant.app
python -m examples.agent_store.app $LOCALHOST $PORT
```

### CLI
```
which llama
llama model prompt-format -m Llama3.2-11B-Vision-Instruct
llama model list
llama stack list-apis
llama stack list-providers inference

llama stack build --template ollama --image-type conda
```

### Distributions Tests
**ollama**
```
llama stack build --template ollama --image-type conda
ollama run llama3.2:1b-instruct-fp16
llama stack run ./llama_stack/templates/ollama/run.yaml --env INFERENCE_MODEL=meta-llama/Llama-3.2-1B-Instruct
```

**fireworks**
```
llama stack build --template fireworks --image-type conda
llama stack run ./llama_stack/templates/fireworks/run.yaml
```

**together**
```
llama stack build --template together --image-type conda
llama stack run ./llama_stack/templates/together/run.yaml
```

**tgi**
```
llama stack run ./llama_stack/templates/tgi/run.yaml --env TGI_URL=http://0.0.0.0:5009 --env INFERENCE_MODEL=meta-llama/Llama-3.1-8B-Instruct
```

## Sources

Please link relevant resources if necessary.


## Before submitting

- [ ] This PR fixes a typo or improves the docs (you can dismiss the
other checks if that's the case).
- [ ] Ran pre-commit to handle lint / formatting issues.
- [ ] Read the [contributor
guideline](https://github.com/meta-llama/llama-stack/blob/main/CONTRIBUTING.md),
      Pull Request section?
- [ ] Updated relevant documentation.
- [ ] Wrote necessary unit or integration tests.
2024-12-27 15:45:44 -08:00

280 lines
8 KiB
Python

# Copyright (c) Meta Platforms, Inc. and affiliates.
# All rights reserved.
#
# This source code is licensed under the terms described in the LICENSE file in
# the root directory of this source tree.
import asyncio
import base64
import logging
import queue
import threading
import uuid
from datetime import datetime
from functools import wraps
from typing import Any, Callable, Dict, List, Optional
from llama_stack.apis.telemetry import (
LogSeverity,
Span,
SpanEndPayload,
SpanStartPayload,
SpanStatus,
StructuredLogEvent,
Telemetry,
UnstructuredLogEvent,
)
from llama_stack.providers.utils.telemetry.trace_protocol import serialize_value
log = logging.getLogger(__name__)
def generate_short_uuid(len: int = 8):
full_uuid = uuid.uuid4()
uuid_bytes = full_uuid.bytes
encoded = base64.urlsafe_b64encode(uuid_bytes)
return encoded.rstrip(b"=").decode("ascii")[:len]
CURRENT_TRACE_CONTEXT = None
BACKGROUND_LOGGER = None
class BackgroundLogger:
def __init__(self, api: Telemetry, capacity: int = 1000):
self.api = api
self.log_queue = queue.Queue(maxsize=capacity)
self.worker_thread = threading.Thread(target=self._process_logs, daemon=True)
self.worker_thread.start()
def log_event(self, event):
try:
self.log_queue.put_nowait(event)
except queue.Full:
log.error("Log queue is full, dropping event")
def _process_logs(self):
while True:
try:
event = self.log_queue.get()
# figure out how to use a thread's native loop
asyncio.run(self.api.log_event(event))
except Exception:
import traceback
traceback.print_exc()
print("Error processing log event")
finally:
self.log_queue.task_done()
def __del__(self):
self.log_queue.join()
class TraceContext:
spans: List[Span] = []
def __init__(self, logger: BackgroundLogger, trace_id: str):
self.logger = logger
self.trace_id = trace_id
def push_span(self, name: str, attributes: Dict[str, Any] = None) -> Span:
current_span = self.get_current_span()
span = Span(
span_id=generate_short_uuid(),
trace_id=self.trace_id,
name=name,
start_time=datetime.now(),
parent_span_id=current_span.span_id if current_span else None,
attributes=attributes,
)
self.logger.log_event(
StructuredLogEvent(
trace_id=span.trace_id,
span_id=span.span_id,
timestamp=span.start_time,
attributes=span.attributes,
payload=SpanStartPayload(
name=span.name,
parent_span_id=span.parent_span_id,
),
)
)
self.spans.append(span)
return span
def pop_span(self, status: SpanStatus = SpanStatus.OK):
span = self.spans.pop()
if span is not None:
self.logger.log_event(
StructuredLogEvent(
trace_id=span.trace_id,
span_id=span.span_id,
timestamp=span.start_time,
attributes=span.attributes,
payload=SpanEndPayload(
status=status,
),
)
)
def get_current_span(self):
return self.spans[-1] if self.spans else None
def setup_logger(api: Telemetry, level: int = logging.INFO):
global BACKGROUND_LOGGER
BACKGROUND_LOGGER = BackgroundLogger(api)
logger = logging.getLogger()
logger.setLevel(level)
logger.addHandler(TelemetryHandler())
async def start_trace(name: str, attributes: Dict[str, Any] = None) -> TraceContext:
global CURRENT_TRACE_CONTEXT, BACKGROUND_LOGGER
if BACKGROUND_LOGGER is None:
log.info("No Telemetry implementation set. Skipping trace initialization...")
return
trace_id = generate_short_uuid(16)
context = TraceContext(BACKGROUND_LOGGER, trace_id)
context.push_span(name, {"__root__": True, **(attributes or {})})
CURRENT_TRACE_CONTEXT = context
return context
async def end_trace(status: SpanStatus = SpanStatus.OK):
global CURRENT_TRACE_CONTEXT
context = CURRENT_TRACE_CONTEXT
if context is None:
return
context.pop_span(status)
CURRENT_TRACE_CONTEXT = None
def severity(levelname: str) -> LogSeverity:
if levelname == "DEBUG":
return LogSeverity.DEBUG
elif levelname == "INFO":
return LogSeverity.INFO
elif levelname == "WARNING":
return LogSeverity.WARN
elif levelname == "ERROR":
return LogSeverity.ERROR
elif levelname == "CRITICAL":
return LogSeverity.CRITICAL
else:
raise ValueError(f"Unknown log level: {levelname}")
# TODO: ideally, the actual emitting should be done inside a separate daemon
# process completely isolated from the server
class TelemetryHandler(logging.Handler):
def emit(self, record: logging.LogRecord):
# horrendous hack to avoid logging from asyncio and getting into an infinite loop
if record.module in ("asyncio", "selector_events"):
return
global CURRENT_TRACE_CONTEXT, BACKGROUND_LOGGER
if BACKGROUND_LOGGER is None:
raise RuntimeError("Telemetry API not initialized")
context = CURRENT_TRACE_CONTEXT
if context is None:
return
span = context.get_current_span()
if span is None:
return
BACKGROUND_LOGGER.log_event(
UnstructuredLogEvent(
trace_id=span.trace_id,
span_id=span.span_id,
timestamp=datetime.now(),
message=self.format(record),
severity=severity(record.levelname),
)
)
def close(self):
pass
class SpanContextManager:
def __init__(self, name: str, attributes: Dict[str, Any] = None):
self.name = name
self.attributes = attributes
self.span = None
def __enter__(self):
global CURRENT_TRACE_CONTEXT
context = CURRENT_TRACE_CONTEXT
if context:
self.span = context.push_span(self.name, self.attributes)
return self
def __exit__(self, exc_type, exc_value, traceback):
global CURRENT_TRACE_CONTEXT
context = CURRENT_TRACE_CONTEXT
if context:
context.pop_span()
def set_attribute(self, key: str, value: Any):
if self.span:
if self.span.attributes is None:
self.span.attributes = {}
self.span.attributes[key] = serialize_value(value)
async def __aenter__(self):
global CURRENT_TRACE_CONTEXT
context = CURRENT_TRACE_CONTEXT
if context:
self.span = context.push_span(self.name, self.attributes)
return self
async def __aexit__(self, exc_type, exc_value, traceback):
global CURRENT_TRACE_CONTEXT
context = CURRENT_TRACE_CONTEXT
if context:
context.pop_span()
def __call__(self, func: Callable):
@wraps(func)
def sync_wrapper(*args, **kwargs):
with self:
return func(*args, **kwargs)
@wraps(func)
async def async_wrapper(*args, **kwargs):
async with self:
return await func(*args, **kwargs)
@wraps(func)
def wrapper(*args, **kwargs):
if asyncio.iscoroutinefunction(func):
return async_wrapper(*args, **kwargs)
else:
return sync_wrapper(*args, **kwargs)
return wrapper
def span(name: str, attributes: Dict[str, Any] = None):
return SpanContextManager(name, attributes)
def get_current_span() -> Optional[Span]:
global CURRENT_TRACE_CONTEXT
context = CURRENT_TRACE_CONTEXT
if context:
return context.get_current_span()
return None