Spaces:
Running
Running
File size: 12,190 Bytes
c2ea5ed 7bc750c c2ea5ed 7bc750c c2ea5ed 7bc750c |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 |
import logging
import asyncio
import uuid
import time
import traceback
from typing import Dict, Any
from sqlalchemy.orm import Session
from backend.database import models
from backend.database.utils import get_trace, save_knowledge_graph, update_trace_status, get_context_documents_from_trace
from agentgraph.extraction.graph_processing import SlidingWindowMonitor
from agentgraph.input.text_processing import ChunkingService
from backend.services.task_service import update_task_status
logger = logging.getLogger(__name__)
class PipelineError(Exception):
"""Exception raised for errors in the pipeline processing."""
pass
async def process_trace_task(trace_id: str, session: Session, task_id: str = None, splitter_type: str = "agent_semantic", force_regenerate: bool = False, method_name: str = "production", model: str = "gpt-5-mini", chunking_config = None) -> Dict[str, Any]:
"""
Process a single trace and return the merged knowledge graph with proper timeout handling.
"""
start_time = time.time()
max_processing_time = 3600 # 1 hour maximum processing time
try:
if task_id:
update_task_status(task_id, "PROCESSING", "Initializing trace processing", 5)
trace = get_trace(session, trace_id)
if not trace:
raise PipelineError(f"Trace with ID {trace_id} not found")
if not force_regenerate:
existing_kgs = session.query(models.KnowledgeGraph).filter(
models.KnowledgeGraph.trace_id == trace_id,
models.KnowledgeGraph.status == "created"
).all()
if existing_kgs and task_id:
update_task_status(task_id, "COMPLETED", f"Found {len(existing_kgs)} existing knowledge graphs. Use force_regenerate=True to create new graphs.", 100)
return {
"message": f"Trace {trace_id} already processed",
"existing_graphs": len(existing_kgs),
"status": "already_processed",
"suggestion": "Use force_regenerate=True to bypass this check and generate new graphs"
}
else:
if task_id:
update_task_status(task_id, "PROCESSING", "Force regenerate enabled - creating new knowledge graphs", 5)
trace_content = trace.content
if not trace_content:
raise PipelineError(f"Trace {trace_id} has no content")
context_documents = get_context_documents_from_trace(session, trace_id)
context_count = len(context_documents) if context_documents else 0
logger.info(f"Retrieved {context_count} context documents for trace {trace_id}")
def check_timeout():
if time.time() - start_time > max_processing_time:
raise PipelineError(f"Processing timeout exceeded {max_processing_time} seconds")
check_timeout()
if task_id:
context_msg = f" with {context_count} context documents" if context_count > 0 else ""
update_task_status(task_id, "PROCESSING", f"Splitting trace content ({len(trace_content)} characters){context_msg}", 10)
logger.info(f"Splitting trace content with {splitter_type} splitter")
chunking_service = ChunkingService(
default_batch_size=3,
default_model="gpt-5-mini"
)
check_timeout()
# Extract chunking parameters from config
chunk_kwargs = {}
if chunking_config:
# Map min_chunk_size and max_chunk_size to the ChunkingService parameters
# For AgentAwareSemanticSplitter, max_chunk_size maps to window_size
if chunking_config.max_chunk_size:
chunk_kwargs["window_size"] = chunking_config.max_chunk_size
# Pass min_chunk_size directly
if chunking_config.min_chunk_size:
chunk_kwargs["min_chunk_size"] = chunking_config.min_chunk_size
# Calculate overlap_size based on window_size (5% default)
if chunking_config.max_chunk_size:
chunk_kwargs["overlap_size"] = max(1000, int(chunking_config.max_chunk_size * 0.05))
logger.info(f"Using custom chunking config: {chunking_config}")
logger.info(f"Mapped to chunk_kwargs: {chunk_kwargs}")
chunks = chunking_service.chunk_trace_content(
content=trace_content,
splitter_type=splitter_type,
**chunk_kwargs
)
logger.info(f"Split content into {len(chunks)} chunks")
if not chunks:
raise PipelineError("No chunks were created from the trace content")
if task_id:
update_task_status(task_id, "PROCESSING", f"Processing {len(chunks)} chunks", 20)
processing_run_id = str(uuid.uuid4())[:8]
monitor = SlidingWindowMonitor(
batch_size=3,
parallel_processing=True,
model=model,
source_trace_id=trace_id,
processing_run_id=processing_run_id,
method_name=method_name,
context_documents=context_documents,
trace_content=trace_content,
trace_metadata=trace.trace_metadata
)
if task_id:
def progress_callback(stage, step, total_steps, message=None):
check_timeout()
base_progress = 30
stage_progress = 50
raw_progress = base_progress + int((step / total_steps) * stage_progress)
overall_progress = min(raw_progress, 80)
status_message = f"{stage}: {step}/{total_steps}"
if message:
status_message += f" - {message}"
update_task_status(task_id, "PROCESSING", status_message, overall_progress)
else:
def progress_callback(stage, step, total_steps, message=None):
check_timeout()
pass
check_timeout()
output_identifier = f"trace_{trace_id}_{processing_run_id}"
if task_id:
update_task_status(task_id, "PROCESSING", "Starting knowledge graph processing", 30)
try:
final_kg_data = await asyncio.wait_for(
monitor.process_trace(chunks, output_identifier, progress_callback=progress_callback),
timeout=max_processing_time - (time.time() - start_time)
)
except asyncio.TimeoutError:
raise PipelineError(f"Trace processing timed out after {max_processing_time} seconds")
check_timeout()
if not isinstance(final_kg_data, dict):
raise PipelineError("Invalid result format from trace processing - expected dictionary")
if 'final_kg' not in final_kg_data:
raise PipelineError("Invalid result format - missing 'final_kg' key in response")
final_kg_info = final_kg_data['final_kg']
if 'graph_data' not in final_kg_info:
raise PipelineError("Invalid result format - missing 'graph_data' key in final_kg")
actual_graph_data = final_kg_info['graph_data']
try:
from agentgraph.reconstruction.content_reference_resolver import ContentReferenceResolver
resolver = ContentReferenceResolver()
actual_graph_data = resolver.resolve_knowledge_graph_content(
actual_graph_data,
trace_content,
{"window_index": 0}
)
except Exception as _resolver_err:
logger.warning(f"ContentReferenceResolver failed on merged KG: {_resolver_err}")
if not actual_graph_data.get('entities') and not actual_graph_data.get('relations'):
raise PipelineError("Knowledge graph processing completed but no entities or relations were found")
if task_id:
update_task_status(task_id, "PROCESSING", "Saving knowledge graph to database", 85)
window_count = final_kg_data.get('window_count', len(chunks))
kg_result = save_knowledge_graph(
session,
output_identifier,
actual_graph_data,
trace_id=trace_id,
window_total=window_count,
processing_run_id=processing_run_id
)
kg_id = kg_result.id
window_kgs_saved = 0
if 'window_kgs' in final_kg_data and final_kg_data['window_kgs']:
if task_id:
update_task_status(task_id, "PROCESSING", "Saving window knowledge graphs to database", 90)
logger.info(f"Saving {len(final_kg_data['window_kgs'])} window knowledge graphs")
for window_kg_data in final_kg_data['window_kgs']:
try:
window_filename = window_kg_data.get('filename')
window_graph_data = window_kg_data.get('graph_data')
window_index = window_kg_data.get('window_index')
window_total = window_kg_data.get('window_total')
window_start_char = window_kg_data.get('window_start_char')
window_end_char = window_kg_data.get('window_end_char')
window_processing_run_id = window_kg_data.get('processing_run_id', processing_run_id)
window_trace_id = window_kg_data.get('trace_id', trace_id)
window_kg_result = save_knowledge_graph(
session,
window_filename,
window_graph_data,
trace_id=window_trace_id,
window_index=window_index,
window_total=window_total,
window_start_char=window_start_char,
window_end_char=window_end_char,
processing_run_id=window_processing_run_id
)
window_kgs_saved += 1
logger.debug(f"Saved window KG {window_index}: {window_filename} (ID: {window_kg_result.id})")
except Exception as window_error:
logger.error(f"Failed to save window knowledge graph {window_kg_data.get('filename', 'unknown')}: {window_error}")
logger.info(f"Successfully saved {window_kgs_saved} window knowledge graphs")
update_trace_status(session, trace_id, "processed")
processing_duration = time.time() - start_time
if task_id:
update_task_status(
task_id,
"COMPLETED",
f"Processing completed in {processing_duration:.1f}s. Knowledge graph saved with ID: {kg_id}",
100
)
logger.info(f"Successfully processed trace {trace_id} in {processing_duration:.1f}s")
return {
"trace_id": trace_id,
"knowledge_graph_id": kg_id,
"processing_duration": processing_duration,
"chunks_processed": len(chunks),
"window_graphs_saved": window_kgs_saved,
"entities_count": len(actual_graph_data.get('entities', [])),
"relations_count": len(actual_graph_data.get('relations', [])),
"status": "completed"
}
except Exception as e:
error_message = f"Error processing trace {trace_id}: {str(e)}"
logger.error(error_message)
logger.error(traceback.format_exc())
if task_id:
update_task_status(task_id, "FAILED", error_message)
try:
update_trace_status(session, trace_id, "failed")
except Exception as update_error:
logger.error(f"Failed to update trace status: {update_error}")
raise PipelineError(error_message)
|