chore: initial snapshot for gitea/github upload
This commit is contained in:
Your Name
@@ -0,0 +1,10 @@
|
||||
"""
|
||||
Opik payload builder namespace.
|
||||
|
||||
Public API:
|
||||
build_opik_payload - Main function to create Opik trace and span payloads
|
||||
"""
|
||||
|
||||
from .api import build_opik_payload
|
||||
|
||||
__all__ = ["build_opik_payload"]
|
||||
@@ -0,0 +1,121 @@
|
||||
"""Public API for Opik payload building."""
|
||||
|
||||
from datetime import datetime
|
||||
from typing import Any, Dict, Optional, Tuple
|
||||
|
||||
from litellm.integrations.opik import utils
|
||||
|
||||
from . import extractors, payload_builders, types
|
||||
|
||||
|
||||
def build_opik_payload(
|
||||
kwargs: Dict[str, Any],
|
||||
response_obj: Dict[str, Any],
|
||||
start_time: datetime,
|
||||
end_time: datetime,
|
||||
project_name: str,
|
||||
) -> Tuple[Optional[types.TracePayload], types.SpanPayload]:
|
||||
"""
|
||||
Build Opik trace and span payloads from LiteLLM completion data.
|
||||
|
||||
This is the main public API for creating Opik payloads. It:
|
||||
1. Extracts all necessary data from LiteLLM kwargs and response
|
||||
2. Decides whether to create a new trace or attach to existing
|
||||
3. Builds trace payload (if new trace)
|
||||
4. Builds span payload (always)
|
||||
|
||||
Args:
|
||||
kwargs: LiteLLM kwargs containing request metadata and logging data
|
||||
response_obj: LiteLLM response object containing model response
|
||||
start_time: Request start time
|
||||
end_time: Request end time
|
||||
project_name: Default Opik project name
|
||||
|
||||
Returns:
|
||||
Tuple of (optional trace payload, span payload)
|
||||
- First element is TracePayload if creating a new trace, None if attaching to existing
|
||||
- Second element is always SpanPayload
|
||||
"""
|
||||
standard_logging_object = kwargs["standard_logging_object"]
|
||||
|
||||
# Extract litellm params and metadata
|
||||
litellm_params = kwargs.get("litellm_params", {}) or {}
|
||||
litellm_metadata = litellm_params.get("metadata", {}) or {}
|
||||
standard_logging_metadata = standard_logging_object.get("metadata", {}) or {}
|
||||
|
||||
# Extract and merge Opik metadata
|
||||
opik_metadata = extractors.extract_opik_metadata(
|
||||
litellm_metadata, standard_logging_metadata
|
||||
)
|
||||
|
||||
# Extract project name
|
||||
current_project_name = opik_metadata.get("project_name", project_name)
|
||||
|
||||
# Extract trace identifiers
|
||||
current_span_data = opik_metadata.get("current_span_data")
|
||||
trace_id, parent_span_id = extractors.extract_span_identifiers(current_span_data)
|
||||
|
||||
# Extract tags and thread_id
|
||||
tags = extractors.extract_tags(opik_metadata, kwargs.get("custom_llm_provider"))
|
||||
thread_id = opik_metadata.get("thread_id")
|
||||
|
||||
# Apply proxy header overrides
|
||||
proxy_request = litellm_params.get("proxy_server_request", {}) or {}
|
||||
proxy_headers = proxy_request.get("headers", {}) or {}
|
||||
current_project_name, tags, thread_id = extractors.apply_proxy_header_overrides(
|
||||
current_project_name, tags, thread_id, proxy_headers
|
||||
)
|
||||
|
||||
# Build shared metadata
|
||||
metadata = extractors.extract_and_build_metadata(
|
||||
opik_metadata=opik_metadata,
|
||||
standard_logging_metadata=standard_logging_metadata,
|
||||
standard_logging_object=standard_logging_object,
|
||||
litellm_kwargs=kwargs,
|
||||
)
|
||||
|
||||
# Get input/output data
|
||||
input_data = standard_logging_object.get("messages", {})
|
||||
output_data = standard_logging_object.get("response", {})
|
||||
|
||||
# Decide whether to create a new trace or attach to existing
|
||||
trace_payload: Optional[types.TracePayload] = None
|
||||
if trace_id is None:
|
||||
trace_id = utils.create_uuid7()
|
||||
trace_payload = payload_builders.build_trace_payload(
|
||||
project_name=current_project_name,
|
||||
trace_id=trace_id,
|
||||
response_obj=response_obj,
|
||||
start_time=start_time,
|
||||
end_time=end_time,
|
||||
input_data=input_data,
|
||||
output_data=output_data,
|
||||
metadata=metadata,
|
||||
tags=tags,
|
||||
thread_id=thread_id,
|
||||
)
|
||||
|
||||
# Always create a span
|
||||
usage = utils.create_usage_object(response_obj["usage"])
|
||||
|
||||
# Extract provider and cost
|
||||
provider = extractors.normalize_provider_name(kwargs.get("custom_llm_provider"))
|
||||
cost = kwargs.get("response_cost")
|
||||
|
||||
span_payload = payload_builders.build_span_payload(
|
||||
project_name=current_project_name,
|
||||
trace_id=trace_id,
|
||||
parent_span_id=parent_span_id,
|
||||
response_obj=response_obj,
|
||||
start_time=start_time,
|
||||
end_time=end_time,
|
||||
input_data=input_data,
|
||||
output_data=output_data,
|
||||
metadata=metadata,
|
||||
tags=tags,
|
||||
usage=usage,
|
||||
provider=provider,
|
||||
cost=cost,
|
||||
)
|
||||
|
||||
return trace_payload, span_payload
|
||||
@@ -0,0 +1,221 @@
|
||||
"""Data extraction functions for Opik payload building."""
|
||||
|
||||
import json
|
||||
from typing import Any, Dict, List, Optional, Tuple
|
||||
|
||||
from litellm import _logging
|
||||
|
||||
|
||||
def normalize_provider_name(provider: Optional[str]) -> Optional[str]:
|
||||
"""
|
||||
Normalize LiteLLM provider names to standardized string names.
|
||||
|
||||
Args:
|
||||
provider: LiteLLM internal provider name
|
||||
|
||||
Returns:
|
||||
Normalized provider name or the original if no mapping exists
|
||||
"""
|
||||
if provider is None:
|
||||
return None
|
||||
|
||||
# Provider mapping to names used in Opik
|
||||
provider_mapping = {
|
||||
"openai": "openai",
|
||||
"vertex_ai-language-models": "google_vertexai",
|
||||
"gemini": "google_ai",
|
||||
"anthropic": "anthropic",
|
||||
"vertex_ai-anthropic_models": "anthropic_vertexai",
|
||||
"bedrock": "bedrock",
|
||||
"bedrock_converse": "bedrock",
|
||||
"groq": "groq",
|
||||
}
|
||||
|
||||
return provider_mapping.get(provider, provider)
|
||||
|
||||
|
||||
def extract_opik_metadata(
|
||||
litellm_metadata: Dict[str, Any],
|
||||
standard_logging_metadata: Dict[str, Any],
|
||||
) -> Dict[str, Any]:
|
||||
"""
|
||||
Extract and merge Opik metadata from request and requester.
|
||||
|
||||
Args:
|
||||
litellm_metadata: Metadata from litellm_params
|
||||
standard_logging_metadata: Metadata from standard_logging_object
|
||||
|
||||
Returns:
|
||||
Merged Opik metadata dictionary
|
||||
"""
|
||||
opik_meta = litellm_metadata.get("opik", {}).copy()
|
||||
|
||||
requester_metadata = standard_logging_metadata.get("requester_metadata", {}) or {}
|
||||
requester_opik = requester_metadata.get("opik", {}) or {}
|
||||
opik_meta.update(requester_opik)
|
||||
|
||||
_logging.verbose_logger.debug(
|
||||
f"litellm_opik_metadata - {json.dumps(opik_meta, default=str)}"
|
||||
)
|
||||
|
||||
return opik_meta
|
||||
|
||||
|
||||
def extract_span_identifiers(
|
||||
current_span_data: Any,
|
||||
) -> Tuple[Optional[str], Optional[str]]:
|
||||
"""
|
||||
Extract trace_id and parent_span_id from current_span_data.
|
||||
|
||||
Args:
|
||||
current_span_data: Either dict with trace_id/id keys or Opik object
|
||||
|
||||
Returns:
|
||||
Tuple of (trace_id, parent_span_id), both optional
|
||||
"""
|
||||
if current_span_data is None:
|
||||
return None, None
|
||||
|
||||
if isinstance(current_span_data, dict):
|
||||
return (current_span_data.get("trace_id"), current_span_data.get("id"))
|
||||
|
||||
try:
|
||||
return current_span_data.trace_id, current_span_data.id
|
||||
except AttributeError:
|
||||
_logging.verbose_logger.warning(
|
||||
f"Unexpected current_span_data format: {type(current_span_data)}"
|
||||
)
|
||||
return None, None
|
||||
|
||||
|
||||
def extract_tags(
|
||||
opik_metadata: Dict[str, Any],
|
||||
custom_llm_provider: Optional[str],
|
||||
) -> List[str]:
|
||||
"""
|
||||
Extract and build list of tags.
|
||||
|
||||
Args:
|
||||
opik_metadata: Opik metadata dictionary
|
||||
custom_llm_provider: LLM provider name to add as tag
|
||||
|
||||
Returns:
|
||||
List of tags
|
||||
"""
|
||||
tags = list(opik_metadata.get("tags", []))
|
||||
|
||||
if custom_llm_provider:
|
||||
tags.append(custom_llm_provider)
|
||||
|
||||
return tags
|
||||
|
||||
|
||||
def apply_proxy_header_overrides(
|
||||
project_name: str,
|
||||
tags: List[str],
|
||||
thread_id: Optional[str],
|
||||
proxy_headers: Dict[str, Any],
|
||||
) -> Tuple[str, List[str], Optional[str]]:
|
||||
"""
|
||||
Apply overrides from proxy request headers (opik_* prefix).
|
||||
|
||||
Args:
|
||||
project_name: Current project name
|
||||
tags: Current tags list
|
||||
thread_id: Current thread ID
|
||||
proxy_headers: HTTP headers from proxy request
|
||||
|
||||
Returns:
|
||||
Tuple of (project_name, tags, thread_id) with overrides applied
|
||||
"""
|
||||
for key, value in proxy_headers.items():
|
||||
if not key.startswith("opik_") or not value:
|
||||
continue
|
||||
|
||||
param_key = key.replace("opik_", "", 1)
|
||||
|
||||
if param_key == "project_name":
|
||||
project_name = value
|
||||
elif param_key == "thread_id":
|
||||
thread_id = value
|
||||
elif param_key == "tags":
|
||||
try:
|
||||
parsed_tags = json.loads(value)
|
||||
if isinstance(parsed_tags, list):
|
||||
tags.extend(parsed_tags)
|
||||
except (json.JSONDecodeError, TypeError):
|
||||
_logging.verbose_logger.warning(
|
||||
f"Failed to parse tags from header: {value}"
|
||||
)
|
||||
|
||||
return project_name, tags, thread_id
|
||||
|
||||
|
||||
def extract_and_build_metadata(
|
||||
opik_metadata: Dict[str, Any],
|
||||
standard_logging_metadata: Dict[str, Any],
|
||||
standard_logging_object: Dict[str, Any],
|
||||
litellm_kwargs: Dict[str, Any],
|
||||
) -> Dict[str, Any]:
|
||||
"""
|
||||
Build the complete metadata dictionary from all available sources.
|
||||
|
||||
This combines:
|
||||
- Opik-specific metadata (tags, etc.)
|
||||
- Standard logging metadata
|
||||
- Fields from standard_logging_object (model info, status, etc.)
|
||||
- Cost information from litellm_kwargs (calculated after completion)
|
||||
|
||||
Args:
|
||||
opik_metadata: Opik-specific metadata from request
|
||||
standard_logging_metadata: Standard logging metadata
|
||||
standard_logging_object: Full standard logging object with call details
|
||||
litellm_kwargs: Original LiteLLM kwargs (includes response_cost)
|
||||
|
||||
Returns:
|
||||
Complete metadata dictionary for trace/span
|
||||
"""
|
||||
# Start with opik metadata (excluding current_span_data which is used for trace linking)
|
||||
metadata = {k: v for k, v in opik_metadata.items() if k != "current_span_data"}
|
||||
metadata["created_from"] = "litellm"
|
||||
|
||||
# Merge with standard logging metadata
|
||||
metadata.update(standard_logging_metadata)
|
||||
|
||||
# Add fields from standard_logging_object
|
||||
# These come from the LiteLLM logging infrastructure
|
||||
field_mappings = {
|
||||
"call_type": "type",
|
||||
"status": "status",
|
||||
"model": "model",
|
||||
"model_id": "model_id",
|
||||
"model_group": "model_group",
|
||||
"api_base": "api_base",
|
||||
"cache_hit": "cache_hit",
|
||||
"saved_cache_cost": "saved_cache_cost",
|
||||
"error_str": "error_str",
|
||||
"model_parameters": "model_parameters",
|
||||
"hidden_params": "hidden_params",
|
||||
"model_map_information": "model_map_information",
|
||||
}
|
||||
|
||||
for source_key, dest_key in field_mappings.items():
|
||||
if source_key in standard_logging_object:
|
||||
metadata[dest_key] = standard_logging_object[source_key]
|
||||
|
||||
# Add cost information
|
||||
# response_cost is calculated by LiteLLM after completion and added to kwargs
|
||||
# See: litellm/litellm_core_utils/llm_response_utils/response_metadata.py
|
||||
if "response_cost" in litellm_kwargs:
|
||||
metadata["cost"] = {
|
||||
"total_tokens": litellm_kwargs["response_cost"],
|
||||
"currency": "USD",
|
||||
}
|
||||
|
||||
# Add debug info if cost calculation failed
|
||||
if "response_cost_failure_debug_info" in litellm_kwargs:
|
||||
metadata["response_cost_failure_debug_info"] = litellm_kwargs[
|
||||
"response_cost_failure_debug_info"
|
||||
]
|
||||
|
||||
return metadata
|
||||
@@ -0,0 +1,89 @@
|
||||
"""Payload builders for Opik traces and spans."""
|
||||
|
||||
from datetime import datetime, timezone
|
||||
from typing import Any, Dict, List, Optional
|
||||
|
||||
from litellm import _logging
|
||||
from litellm.integrations.opik import utils
|
||||
|
||||
from . import types
|
||||
|
||||
|
||||
def build_trace_payload(
|
||||
project_name: str,
|
||||
trace_id: str,
|
||||
response_obj: Dict[str, Any],
|
||||
start_time: datetime,
|
||||
end_time: datetime,
|
||||
input_data: Any,
|
||||
output_data: Any,
|
||||
metadata: Dict[str, Any],
|
||||
tags: List[str],
|
||||
thread_id: Optional[str],
|
||||
) -> types.TracePayload:
|
||||
"""Build a complete trace payload."""
|
||||
trace_name = response_obj.get("object", "unknown type")
|
||||
|
||||
return types.TracePayload(
|
||||
project_name=project_name,
|
||||
id=trace_id,
|
||||
name=trace_name,
|
||||
start_time=(
|
||||
start_time.astimezone(timezone.utc).isoformat().replace("+00:00", "Z")
|
||||
),
|
||||
end_time=end_time.astimezone(timezone.utc).isoformat().replace("+00:00", "Z"),
|
||||
input=input_data,
|
||||
output=output_data,
|
||||
metadata=metadata,
|
||||
tags=tags,
|
||||
thread_id=thread_id,
|
||||
)
|
||||
|
||||
|
||||
def build_span_payload(
|
||||
project_name: str,
|
||||
trace_id: str,
|
||||
parent_span_id: Optional[str],
|
||||
response_obj: Dict[str, Any],
|
||||
start_time: datetime,
|
||||
end_time: datetime,
|
||||
input_data: Any,
|
||||
output_data: Any,
|
||||
metadata: Dict[str, Any],
|
||||
tags: List[str],
|
||||
usage: Dict[str, int],
|
||||
provider: Optional[str] = None,
|
||||
cost: Optional[float] = None,
|
||||
) -> types.SpanPayload:
|
||||
"""Build a complete span payload."""
|
||||
span_id = utils.create_uuid7()
|
||||
|
||||
model = response_obj.get("model", "unknown-model")
|
||||
obj_type = response_obj.get("object", "unknown-object")
|
||||
created = response_obj.get("created", 0)
|
||||
span_name = f"{model}_{obj_type}_{created}"
|
||||
|
||||
_logging.verbose_logger.debug(
|
||||
f"OpikLogger creating span with id {span_id} for trace {trace_id}"
|
||||
)
|
||||
|
||||
return types.SpanPayload(
|
||||
id=span_id,
|
||||
project_name=project_name,
|
||||
trace_id=trace_id,
|
||||
parent_span_id=parent_span_id,
|
||||
name=span_name,
|
||||
type="llm",
|
||||
model=model,
|
||||
start_time=(
|
||||
start_time.astimezone(timezone.utc).isoformat().replace("+00:00", "Z")
|
||||
),
|
||||
end_time=end_time.astimezone(timezone.utc).isoformat().replace("+00:00", "Z"),
|
||||
input=input_data,
|
||||
output=output_data,
|
||||
metadata=metadata,
|
||||
tags=tags,
|
||||
usage=usage,
|
||||
provider=provider,
|
||||
total_cost=cost,
|
||||
)
|
||||
@@ -0,0 +1,46 @@
|
||||
"""Type definitions for Opik payload building."""
|
||||
|
||||
from dataclasses import dataclass
|
||||
from typing import Any, Dict, List, Literal, Optional, Tuple, Union
|
||||
|
||||
|
||||
@dataclass
|
||||
class TracePayload:
|
||||
"""Opik trace payload structure"""
|
||||
|
||||
project_name: str
|
||||
id: str
|
||||
name: str
|
||||
start_time: str
|
||||
end_time: str
|
||||
input: Any
|
||||
output: Any
|
||||
metadata: Dict[str, Any]
|
||||
tags: List[str]
|
||||
thread_id: Optional[str] = None
|
||||
|
||||
|
||||
@dataclass
|
||||
class SpanPayload:
|
||||
"""Opik span payload structure"""
|
||||
|
||||
id: str
|
||||
project_name: str
|
||||
trace_id: str
|
||||
name: str
|
||||
type: Literal["llm"]
|
||||
model: str
|
||||
start_time: str
|
||||
end_time: str
|
||||
input: Any
|
||||
output: Any
|
||||
metadata: Dict[str, Any]
|
||||
tags: List[str]
|
||||
usage: Dict[str, int]
|
||||
parent_span_id: Optional[str] = None
|
||||
provider: Optional[str] = None
|
||||
total_cost: Optional[float] = None
|
||||
|
||||
|
||||
PayloadItem = Union[TracePayload, SpanPayload]
|
||||
TraceSpanPayloadTuple = Tuple[Optional[TracePayload], SpanPayload]
|
||||
Reference in New Issue
Block a user