Spaces:
Running
Running
File size: 10,763 Bytes
c2ea5ed |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 |
"""
Base Trace Parser Interface
Defines the common interface and data structures for platform-specific trace parsers.
Each parser extracts structured metadata that is guaranteed to be present in traces
from that specific platform.
"""
from abc import ABC, abstractmethod
from typing import Dict, List, Any, Optional, Union
from dataclasses import dataclass, field
from datetime import datetime
import json
import logging
logger = logging.getLogger(__name__)
@dataclass
class AgentInfo:
"""Information about an agent found in the trace"""
name: str
role: Optional[str] = None
agent_type: Optional[str] = None # e.g., "llm", "tool", "chain"
model: Optional[str] = None
parameters: Optional[Dict[str, Any]] = None
first_appearance_line: Optional[int] = None
@dataclass
class ToolInfo:
"""Information about a tool found in the trace"""
name: str
tool_type: Optional[str] = None # e.g., "function", "api", "external"
description: Optional[str] = None
parameters: Optional[Dict[str, Any]] = None
usage_count: int = 0
first_appearance_line: Optional[int] = None
@dataclass
class WorkflowInfo:
"""Information about the workflow structure"""
workflow_type: Optional[str] = None # e.g., "sequential", "parallel", "hierarchical"
total_steps: Optional[int] = None
project_name: Optional[str] = None
run_id: Optional[str] = None
start_time: Optional[datetime] = None
end_time: Optional[datetime] = None
duration_ms: Optional[float] = None
@dataclass
class DataFlowInfo:
"""Information about data flows and transformations"""
input_types: List[str] = field(default_factory=list)
output_types: List[str] = field(default_factory=list)
intermediate_data_types: List[str] = field(default_factory=list)
transformation_patterns: List[str] = field(default_factory=list)
@dataclass
class ParsedMetadata:
"""
Structured metadata extracted from a platform-specific trace.
This complements the multi-agent knowledge extractor by providing
reliable structural information that can guide the extraction process.
"""
# Source information
platform: str
trace_source: str
confidence: float # 0.0-1.0 confidence in parsing accuracy
# Core structural information
agents: List[AgentInfo] = field(default_factory=list)
tools: List[ToolInfo] = field(default_factory=list)
workflow: Optional[WorkflowInfo] = None
data_flow: Optional[DataFlowInfo] = None
# Platform-specific raw data (preserved for reference)
raw_platform_data: Optional[Dict[str, Any]] = None
# Extraction hints for the knowledge extractor
extraction_hints: Dict[str, Any] = field(default_factory=dict)
# Context document suggestions
suggested_context_types: List[str] = field(default_factory=list)
def to_dict(self) -> Dict[str, Any]:
"""Convert to dictionary for serialization"""
return {
'platform': self.platform,
'trace_source': self.trace_source,
'confidence': self.confidence,
'agents': [
{
'name': agent.name,
'role': agent.role,
'agent_type': agent.agent_type,
'model': agent.model,
'parameters': agent.parameters,
'first_appearance_line': agent.first_appearance_line
}
for agent in self.agents
],
'tools': [
{
'name': tool.name,
'tool_type': tool.tool_type,
'description': tool.description,
'parameters': tool.parameters,
'usage_count': tool.usage_count,
'first_appearance_line': tool.first_appearance_line
}
for tool in self.tools
],
'workflow': {
'workflow_type': self.workflow.workflow_type if self.workflow else None,
'total_steps': self.workflow.total_steps if self.workflow else None,
'project_name': self.workflow.project_name if self.workflow else None,
'run_id': self.workflow.run_id if self.workflow else None,
'start_time': self.workflow.start_time.isoformat() if self.workflow and self.workflow.start_time else None,
'end_time': self.workflow.end_time.isoformat() if self.workflow and self.workflow.end_time else None,
'duration_ms': self.workflow.duration_ms if self.workflow else None
} if self.workflow else None,
'data_flow': {
'input_types': self.data_flow.input_types if self.data_flow else [],
'output_types': self.data_flow.output_types if self.data_flow else [],
'intermediate_data_types': self.data_flow.intermediate_data_types if self.data_flow else [],
'transformation_patterns': self.data_flow.transformation_patterns if self.data_flow else []
} if self.data_flow else None,
'extraction_hints': self.extraction_hints,
'suggested_context_types': self.suggested_context_types
}
class BaseTraceParser(ABC):
"""
Abstract base class for platform-specific trace parsers.
Each parser is responsible for extracting structured metadata that is
guaranteed to be present in traces from that specific platform.
"""
def __init__(self):
self.logger = logging.getLogger(f"{__name__}.{self.__class__.__name__}")
@property
@abstractmethod
def platform_name(self) -> str:
"""Return the name of the platform this parser handles"""
pass
@property
@abstractmethod
def supported_trace_types(self) -> List[str]:
"""Return list of trace types this parser can handle"""
pass
@abstractmethod
def can_parse(self, trace_content: str, trace_metadata: Optional[Dict[str, Any]] = None) -> bool:
"""
Determine if this parser can handle the given trace.
Args:
trace_content: Raw trace content
trace_metadata: Optional metadata from database
Returns:
True if this parser can handle the trace
"""
pass
@abstractmethod
def parse_trace(self, trace_content: str, trace_metadata: Optional[Dict[str, Any]] = None) -> ParsedMetadata:
"""
Parse the trace and extract structured metadata.
Args:
trace_content: Raw trace content
trace_metadata: Optional metadata from database
Returns:
ParsedMetadata object with extracted information
"""
pass
def _safe_json_parse(self, content: str) -> Optional[Dict[str, Any]]:
"""Safely parse JSON content, returning None if invalid"""
try:
return json.loads(content)
except (json.JSONDecodeError, TypeError):
return None
def _extract_timestamps(self, data: Dict[str, Any]) -> tuple[Optional[datetime], Optional[datetime]]:
"""Extract start and end timestamps from platform data"""
start_time = None
end_time = None
# Common timestamp field names
start_fields = ['start_time', 'startTime', 'created_at', 'createdAt', 'timestamp']
end_fields = ['end_time', 'endTime', 'finished_at', 'finishedAt', 'completed_at']
for field in start_fields:
if field in data and data[field]:
try:
if isinstance(data[field], str):
start_time = datetime.fromisoformat(data[field].replace('Z', '+00:00'))
elif isinstance(data[field], (int, float)):
start_time = datetime.fromtimestamp(data[field])
break
except (ValueError, TypeError):
continue
for field in end_fields:
if field in data and data[field]:
try:
if isinstance(data[field], str):
end_time = datetime.fromisoformat(data[field].replace('Z', '+00:00'))
elif isinstance(data[field], (int, float)):
end_time = datetime.fromtimestamp(data[field])
break
except (ValueError, TypeError):
continue
return start_time, end_time
def _calculate_duration(self, start_time: Optional[datetime], end_time: Optional[datetime]) -> Optional[float]:
"""Calculate duration in milliseconds between start and end times"""
if start_time and end_time:
try:
delta = end_time - start_time
return delta.total_seconds() * 1000
except (TypeError, ValueError):
return None
return None
def generate_extraction_hints(self, parsed_metadata: ParsedMetadata) -> Dict[str, Any]:
"""
Generate hints for the multi-agent knowledge extractor based on parsed metadata.
This method can be overridden by specific parsers to provide platform-specific hints.
"""
hints = {}
# Agent-related hints
if parsed_metadata.agents:
hints['expected_agent_count'] = len(parsed_metadata.agents)
hints['agent_types'] = list(set(agent.agent_type for agent in parsed_metadata.agents if agent.agent_type))
hints['agent_names'] = [agent.name for agent in parsed_metadata.agents]
# Tool-related hints
if parsed_metadata.tools:
hints['expected_tool_count'] = len(parsed_metadata.tools)
hints['tool_types'] = list(set(tool.tool_type for tool in parsed_metadata.tools if tool.tool_type))
hints['tool_names'] = [tool.name for tool in parsed_metadata.tools]
# Workflow hints
if parsed_metadata.workflow:
if parsed_metadata.workflow.workflow_type:
hints['workflow_pattern'] = parsed_metadata.workflow.workflow_type
if parsed_metadata.workflow.total_steps:
hints['expected_task_count'] = parsed_metadata.workflow.total_steps
# Data flow hints
if parsed_metadata.data_flow:
hints['input_types'] = parsed_metadata.data_flow.input_types
hints['output_types'] = parsed_metadata.data_flow.output_types
hints['transformation_patterns'] = parsed_metadata.data_flow.transformation_patterns
return hints |