Spaces:
Sleeping
Sleeping
File size: 17,781 Bytes
5e1a30c |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 |
"""
Rich Response Assembler Implementation.
This module provides comprehensive response assembly with detailed metadata,
source information, and enhanced formatting for production use.
Features:
- Comprehensive metadata collection
- Citation analysis and validation
- Source document summaries
- Quality metrics and confidence scoring
- Detailed assembly diagnostics
"""
import logging
from typing import Dict, Any, List, Optional
from pathlib import Path
import sys
# Add project paths for imports
project_root = Path(__file__).parent.parent.parent.parent.parent
sys.path.append(str(project_root))
from ..base import ContextSelection, QueryAnalysis
from .base_assembler import BaseResponseAssembler
from src.core.interfaces import Answer, Document
logger = logging.getLogger(__name__)
class RichAssembler(BaseResponseAssembler):
"""
Rich response assembler with comprehensive metadata and formatting.
This assembler creates Answer objects with detailed metadata, source
summaries, citation analysis, and quality metrics. It's designed for
production use where comprehensive information is needed.
Configuration Options:
- include_source_summaries: Include document summaries (default: True)
- include_citation_analysis: Analyze citations in answer (default: True)
- include_quality_metrics: Include quality assessment (default: True)
- include_debug_info: Include assembly diagnostics (default: False)
- citation_format: Citation format style ("inline", "numbered", "document")
"""
def __init__(self, config: Optional[Dict[str, Any]] = None):
"""
Initialize rich assembler with configuration.
Args:
config: Configuration dictionary
"""
# Initialize attributes first before calling super().__init__
config_dict = config or {}
self._include_source_summaries = config_dict.get('include_source_summaries', True)
self._include_citation_analysis = config_dict.get('include_citation_analysis', True)
self._include_quality_metrics = config_dict.get('include_quality_metrics', True)
self._include_debug_info = config_dict.get('include_debug_info', False)
self._citation_format = config_dict.get('citation_format', 'inline')
super().__init__(config)
logger.debug(f"Initialized RichAssembler with citation_format={self._citation_format}")
def _assemble_answer(
self,
query: str,
answer_text: str,
context: ContextSelection,
confidence: float,
query_analysis: Optional[QueryAnalysis] = None,
generation_metadata: Optional[Dict[str, Any]] = None
) -> Answer:
"""
Assemble comprehensive Answer object with rich metadata.
Args:
query: Validated query string
answer_text: Validated answer text
context: Context selection
confidence: Validated confidence score
query_analysis: Optional query analysis
generation_metadata: Optional generation metadata
Returns:
Answer object with comprehensive metadata
"""
# Format answer text
formatted_text = self._format_answer_text(answer_text)
# Create sources list
sources = self._create_sources_list(context)
# Create base metadata
metadata = self._create_base_metadata(query, context, query_analysis, generation_metadata)
# Add rich metadata
if self._include_source_summaries:
metadata['source_summaries'] = self._create_source_summaries(context.selected_documents)
if self._include_citation_analysis:
citation_analysis = self._analyze_citations(formatted_text, context.selected_documents)
metadata['citation_analysis'] = citation_analysis
if self._include_quality_metrics:
quality_metrics = self._calculate_quality_metrics(
formatted_text, context, confidence, query_analysis
)
metadata['quality_metrics'] = quality_metrics
if self._include_debug_info:
debug_info = self._create_debug_info(context, generation_metadata)
metadata['debug_info'] = debug_info
# Add assembly-specific metadata
metadata.update({
'assembler_version': '1.0',
'assembly_features': self._get_enabled_features(),
'answer_length': len(formatted_text),
'word_count': len(formatted_text.split()),
'source_count': len(sources)
})
return Answer(
text=formatted_text,
sources=sources,
confidence=confidence,
metadata=metadata
)
def _create_source_summaries(self, documents: List[Document]) -> List[Dict[str, Any]]:
"""
Create summaries for source documents.
Args:
documents: Source documents
Returns:
List of document summaries
"""
summaries = []
for i, doc in enumerate(documents):
# Get source and chunk_id from metadata or attributes
source = getattr(doc, 'source', None) or doc.metadata.get('source', 'unknown')
chunk_id = getattr(doc, 'chunk_id', None) or doc.metadata.get('chunk_id', 'unknown')
summary = {
'index': i,
'source': source,
'chunk_id': chunk_id,
'content_length': len(doc.content),
'word_count': len(doc.content.split()),
'preview': doc.content[:200] + "..." if len(doc.content) > 200 else doc.content
}
# Add document metadata if available
if doc.metadata:
# Extract useful metadata fields
metadata_fields = ['page', 'title', 'section', 'quality_score']
for field in metadata_fields:
if field in doc.metadata:
summary[field] = doc.metadata[field]
# Add relevance score if available
if hasattr(doc, 'score'):
summary['relevance_score'] = doc.score
summaries.append(summary)
return summaries
def _analyze_citations(self, answer_text: str, documents: List[Document]) -> Dict[str, Any]:
"""
Analyze citations in the answer text.
Args:
answer_text: Answer text to analyze
documents: Source documents used
Returns:
Citation analysis results
"""
citations_found = self._extract_citations_from_text(answer_text)
analysis = {
'citations_found': citations_found,
'citation_count': len(citations_found),
'has_citations': len(citations_found) > 0,
'citation_density': len(citations_found) / max(1, len(answer_text.split())) * 100, # Citations per 100 words
}
# Validate citations against available sources
validation_results = self._validate_citations(citations_found, documents)
analysis['validation'] = validation_results
# Analyze citation patterns
pattern_analysis = self._analyze_citation_patterns(citations_found)
analysis['patterns'] = pattern_analysis
return analysis
def _validate_citations(self, citations: List[str], documents: List[Document]) -> Dict[str, Any]:
"""
Validate that citations reference available documents.
Args:
citations: List of citation strings
documents: Available source documents
Returns:
Citation validation results
"""
validation = {
'valid_citations': [],
'invalid_citations': [],
'validation_rate': 0.0
}
# Create mapping of available document references
available_refs = set()
for i, doc in enumerate(documents):
# Common reference formats
available_refs.add(f"[Document {i+1}]")
available_refs.add(f"[{i+1}]")
# Get chunk_id from attribute or metadata
chunk_id = getattr(doc, 'chunk_id', None) or doc.metadata.get('chunk_id', None)
if chunk_id:
available_refs.add(f"[{chunk_id}]")
# Validate each citation
for citation in citations:
if citation in available_refs:
validation['valid_citations'].append(citation)
else:
validation['invalid_citations'].append(citation)
# Calculate validation rate
total_citations = len(citations)
if total_citations > 0:
validation['validation_rate'] = len(validation['valid_citations']) / total_citations
return validation
def _analyze_citation_patterns(self, citations: List[str]) -> Dict[str, Any]:
"""
Analyze patterns in citation usage.
Args:
citations: List of citation strings
Returns:
Pattern analysis results
"""
import re
patterns = {
'document_format': 0, # [Document N]
'simple_format': 0, # [N]
'chunk_format': 0, # [chunk_N]
'page_format': 0 # [Document N, Page N]
}
for citation in citations:
if re.match(r'\[Document \d+, Page \d+\]', citation):
patterns['page_format'] += 1
elif re.match(r'\[Document \d+\]', citation):
patterns['document_format'] += 1
elif re.match(r'\[chunk_\d+\]', citation):
patterns['chunk_format'] += 1
elif re.match(r'\[\d+\]', citation):
patterns['simple_format'] += 1
# Determine dominant pattern
dominant_pattern = max(patterns.items(), key=lambda x: x[1])[0] if citations else 'none'
return {
'format_counts': patterns,
'dominant_format': dominant_pattern,
'format_consistency': max(patterns.values()) / max(1, len(citations))
}
def _calculate_quality_metrics(
self,
answer_text: str,
context: ContextSelection,
confidence: float,
query_analysis: Optional[QueryAnalysis] = None
) -> Dict[str, Any]:
"""
Calculate quality metrics for the assembled answer.
Args:
answer_text: Generated answer text
context: Context selection used
confidence: Answer confidence score
query_analysis: Optional query analysis
Returns:
Quality metrics dictionary
"""
metrics = {}
# Text quality metrics
metrics['answer_length'] = len(answer_text)
metrics['word_count'] = len(answer_text.split())
metrics['sentence_count'] = answer_text.count('.') + answer_text.count('!') + answer_text.count('?')
# Calculate average sentence length
if metrics['sentence_count'] > 0:
metrics['avg_sentence_length'] = metrics['word_count'] / metrics['sentence_count']
else:
metrics['avg_sentence_length'] = 0.0
# Content quality indicators
metrics['has_technical_content'] = any(
term in answer_text.lower()
for term in ['implementation', 'algorithm', 'protocol', 'configuration', 'api']
)
metrics['has_examples'] = any(
phrase in answer_text.lower()
for phrase in ['example', 'for instance', 'such as', 'like']
)
metrics['has_explanations'] = any(
phrase in answer_text.lower()
for phrase in ['because', 'since', 'due to', 'this means', 'in other words']
)
# Source utilization metrics
metrics['sources_used'] = len(context.selected_documents)
metrics['token_efficiency'] = context.total_tokens / max(1, len(answer_text))
if hasattr(context, 'relevance_score') and context.relevance_score is not None:
metrics['source_relevance'] = context.relevance_score
if hasattr(context, 'diversity_score') and context.diversity_score is not None:
metrics['source_diversity'] = context.diversity_score
# Overall quality score (0.0 - 1.0)
quality_components = [
confidence, # LLM confidence
min(1.0, metrics['word_count'] / 50), # Length adequacy (up to 50 words)
1.0 if metrics['has_technical_content'] else 0.5, # Technical content
1.0 if metrics['has_explanations'] else 0.7, # Explanatory content
min(1.0, metrics['sources_used'] / 3) # Source utilization (up to 3 sources)
]
metrics['overall_quality'] = sum(quality_components) / len(quality_components)
return metrics
def _create_debug_info(
self,
context: ContextSelection,
generation_metadata: Optional[Dict[str, Any]] = None
) -> Dict[str, Any]:
"""
Create debug information for troubleshooting.
Args:
context: Context selection
generation_metadata: Generation metadata
Returns:
Debug information dictionary
"""
debug_info = {
'context_metadata': context.metadata,
'selection_strategy': context.selection_strategy,
'total_tokens': context.total_tokens
}
if generation_metadata:
debug_info['generation_metadata'] = generation_metadata
# Add assembler configuration
debug_info['assembler_config'] = {
'include_source_summaries': self._include_source_summaries,
'include_citation_analysis': self._include_citation_analysis,
'include_quality_metrics': self._include_quality_metrics,
'citation_format': self._citation_format
}
return debug_info
def _get_enabled_features(self) -> List[str]:
"""
Get list of enabled rich assembler features.
Returns:
List of enabled feature names
"""
features = []
if self._include_source_summaries:
features.append('source_summaries')
if self._include_citation_analysis:
features.append('citation_analysis')
if self._include_quality_metrics:
features.append('quality_metrics')
if self._include_debug_info:
features.append('debug_info')
if self._format_citations:
features.append('citation_formatting')
return features
def get_supported_formats(self) -> List[str]:
"""
Return list of formats this rich assembler supports.
Returns:
List of format names
"""
base_formats = super().get_supported_formats()
rich_formats = [
'rich',
'comprehensive',
'detailed',
'production'
]
return base_formats + rich_formats
def configure(self, config: Dict[str, Any]) -> None:
"""
Configure the rich assembler with provided settings.
Args:
config: Configuration dictionary
"""
super().configure(config)
# Update rich assembler specific configuration
self._include_source_summaries = config.get('include_source_summaries', self._include_source_summaries)
self._include_citation_analysis = config.get('include_citation_analysis', self._include_citation_analysis)
self._include_quality_metrics = config.get('include_quality_metrics', self._include_quality_metrics)
self._include_debug_info = config.get('include_debug_info', self._include_debug_info)
# Validate citation format
valid_formats = ['inline', 'numbered', 'document']
new_format = config.get('citation_format', self._citation_format)
if new_format in valid_formats:
self._citation_format = new_format
else:
logger.warning(f"Invalid citation_format {new_format}, keeping {self._citation_format}")
def _format_answer_text(self, answer_text: str) -> str:
"""
Format answer text with rich formatting options.
Args:
answer_text: Raw answer text
Returns:
Formatted answer text
"""
# Base formatting
formatted = super()._format_answer_text(answer_text)
# Additional rich formatting
if self._format_citations:
formatted = self._apply_citation_formatting(formatted)
return formatted
def _apply_citation_formatting(self, text: str) -> str:
"""
Apply citation formatting based on configuration.
Args:
text: Text with citations
Returns:
Text with formatted citations
"""
# This is a placeholder for citation formatting
# Can be extended based on specific formatting requirements
return text |