Spaces:
Sleeping
Sleeping
Update src/ai/extractors/entity_extractor.py
Browse files
src/ai/extractors/entity_extractor.py
CHANGED
|
@@ -1,25 +1,641 @@
|
|
| 1 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
from langchain.chains import create_extraction_chain
|
| 3 |
-
from typing import Dict, Any
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
|
| 5 |
class EntityExtractor:
|
| 6 |
-
"""
|
|
|
|
|
|
|
|
|
|
| 7 |
|
| 8 |
def __init__(self, llm_service):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9 |
self.llm_service = llm_service
|
| 10 |
self.setup_extraction_chain()
|
| 11 |
|
| 12 |
def setup_extraction_chain(self):
|
| 13 |
-
"""
|
|
|
|
|
|
|
|
|
|
| 14 |
self.entity_schema = {
|
| 15 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 16 |
}
|
| 17 |
|
| 18 |
-
|
| 19 |
-
self.
|
| 20 |
-
|
| 21 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 22 |
|
| 23 |
async def extract_entities(self, text: str) -> Dict[str, Any]:
|
| 24 |
-
"""
|
| 25 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""
|
| 2 |
+
Advanced Entity Extractor Module
|
| 3 |
+
Handles hyper-contextual entity and relationship extraction from meeting transcripts
|
| 4 |
+
with multi-dimensional relationship mapping, temporal context tracking, and
|
| 5 |
+
deep domain-specific pattern recognition.
|
| 6 |
+
|
| 7 |
+
Key Features:
|
| 8 |
+
- Multi-level relationship mapping
|
| 9 |
+
- Temporal context preservation
|
| 10 |
+
- Cross-meeting context linking
|
| 11 |
+
- Emotional intelligence integration
|
| 12 |
+
- Strategic intent analysis
|
| 13 |
+
- Influence network mapping
|
| 14 |
+
- Business pattern recognition
|
| 15 |
+
"""
|
| 16 |
from langchain.chains import create_extraction_chain
|
| 17 |
+
from typing import Dict, Any, List, Optional
|
| 18 |
+
from datetime import datetime
|
| 19 |
+
import logging
|
| 20 |
+
|
| 21 |
+
logger = logging.getLogger(__name__)
|
| 22 |
|
| 23 |
class EntityExtractor:
|
| 24 |
+
"""
|
| 25 |
+
Extracts entities, relationships, and business context from meeting transcripts.
|
| 26 |
+
Provides detailed analysis of participants, actions, decisions, and opportunities.
|
| 27 |
+
"""
|
| 28 |
|
| 29 |
def __init__(self, llm_service):
|
| 30 |
+
"""
|
| 31 |
+
Initialize entity extractor with LLM service.
|
| 32 |
+
|
| 33 |
+
Args:
|
| 34 |
+
llm_service: Language model service for extraction
|
| 35 |
+
"""
|
| 36 |
self.llm_service = llm_service
|
| 37 |
self.setup_extraction_chain()
|
| 38 |
|
| 39 |
def setup_extraction_chain(self):
|
| 40 |
+
"""
|
| 41 |
+
Setup the extraction chain with comprehensive schema for business context.
|
| 42 |
+
Defines extractable entities, relationships, and metadata.
|
| 43 |
+
"""
|
| 44 |
self.entity_schema = {
|
| 45 |
+
"properties": {
|
| 46 |
+
"extraction_metadata": {
|
| 47 |
+
"type": "object",
|
| 48 |
+
"properties": {
|
| 49 |
+
"timestamp": {"type": "string"},
|
| 50 |
+
"version": {"type": "string"},
|
| 51 |
+
"confidence_scores": {
|
| 52 |
+
"type": "object",
|
| 53 |
+
"properties": {
|
| 54 |
+
"entity_extraction": {"type": "number"},
|
| 55 |
+
"relationship_mapping": {"type": "number"},
|
| 56 |
+
"intent_detection": {"type": "number"},
|
| 57 |
+
"sentiment_analysis": {"type": "number"}
|
| 58 |
+
}
|
| 59 |
+
},
|
| 60 |
+
"context_sources": {
|
| 61 |
+
"type": "array",
|
| 62 |
+
"items": {
|
| 63 |
+
"type": "object",
|
| 64 |
+
"properties": {
|
| 65 |
+
"source_type": {"type": "string"},
|
| 66 |
+
"reference_id": {"type": "string"},
|
| 67 |
+
"relevance_score": {"type": "number"}
|
| 68 |
+
}
|
| 69 |
+
}
|
| 70 |
+
}
|
| 71 |
+
}
|
| 72 |
+
},
|
| 73 |
+
|
| 74 |
+
"historical_context": {
|
| 75 |
+
"type": "object",
|
| 76 |
+
"properties": {
|
| 77 |
+
"related_meetings": {
|
| 78 |
+
"type": "array",
|
| 79 |
+
"items": {
|
| 80 |
+
"type": "object",
|
| 81 |
+
"properties": {
|
| 82 |
+
"meeting_id": {"type": "string"},
|
| 83 |
+
"date": {"type": "string"},
|
| 84 |
+
"relevance_score": {"type": "number"},
|
| 85 |
+
"key_outcomes": {"type": "array", "items": {"type": "string"}},
|
| 86 |
+
"context_carryover": {
|
| 87 |
+
"type": "array",
|
| 88 |
+
"items": {
|
| 89 |
+
"type": "object",
|
| 90 |
+
"properties": {
|
| 91 |
+
"topic": {"type": "string"},
|
| 92 |
+
"status": {"type": "string"},
|
| 93 |
+
"evolution": {"type": "string"}
|
| 94 |
+
}
|
| 95 |
+
}
|
| 96 |
+
}
|
| 97 |
+
}
|
| 98 |
+
}
|
| 99 |
+
},
|
| 100 |
+
"relationship_history": {
|
| 101 |
+
"type": "array",
|
| 102 |
+
"items": {
|
| 103 |
+
"type": "object",
|
| 104 |
+
"properties": {
|
| 105 |
+
"entities": {
|
| 106 |
+
"type": "array",
|
| 107 |
+
"items": {"type": "string"}
|
| 108 |
+
},
|
| 109 |
+
"interaction_pattern": {"type": "string"},
|
| 110 |
+
"temporal_changes": {
|
| 111 |
+
"type": "array",
|
| 112 |
+
"items": {
|
| 113 |
+
"type": "object",
|
| 114 |
+
"properties": {
|
| 115 |
+
"timestamp": {"type": "string"},
|
| 116 |
+
"change_type": {"type": "string"},
|
| 117 |
+
"description": {"type": "string"}
|
| 118 |
+
}
|
| 119 |
+
}
|
| 120 |
+
}
|
| 121 |
+
}
|
| 122 |
+
}
|
| 123 |
+
}
|
| 124 |
+
}
|
| 125 |
+
},
|
| 126 |
+
|
| 127 |
+
"meeting_context": {
|
| 128 |
+
"type": "object",
|
| 129 |
+
"properties": {
|
| 130 |
+
"meeting_id": {"type": "string"},
|
| 131 |
+
"date_time": {"type": "string"},
|
| 132 |
+
"duration": {"type": "string"},
|
| 133 |
+
"type": {
|
| 134 |
+
"type": "string",
|
| 135 |
+
"enum": [
|
| 136 |
+
"sales_opportunity",
|
| 137 |
+
"project_review",
|
| 138 |
+
"status_update",
|
| 139 |
+
"contract_negotiation",
|
| 140 |
+
"strategic_planning",
|
| 141 |
+
"technical_discussion"
|
| 142 |
+
]
|
| 143 |
+
},
|
| 144 |
+
"format": {
|
| 145 |
+
"type": "string",
|
| 146 |
+
"enum": ["in_person", "virtual", "hybrid"]
|
| 147 |
+
},
|
| 148 |
+
"location": {"type": "string"},
|
| 149 |
+
"objectives": {"type": "array", "items": {"type": "string"}},
|
| 150 |
+
"previous_meeting_reference": {"type": "string"}
|
| 151 |
+
}
|
| 152 |
+
},
|
| 153 |
+
|
| 154 |
+
"participants": {
|
| 155 |
+
"type": "array",
|
| 156 |
+
"items": {
|
| 157 |
+
"type": "object",
|
| 158 |
+
"properties": {
|
| 159 |
+
"name": {"type": "string"},
|
| 160 |
+
"role": {"type": "string"},
|
| 161 |
+
"organization": {"type": "string"},
|
| 162 |
+
"department": {"type": "string"},
|
| 163 |
+
"title": {"type": "string"},
|
| 164 |
+
"attendance": {
|
| 165 |
+
"type": "string",
|
| 166 |
+
"enum": ["full", "partial", "absent"]
|
| 167 |
+
},
|
| 168 |
+
"engagement_level": {
|
| 169 |
+
"type": "string",
|
| 170 |
+
"enum": ["high", "medium", "low"]
|
| 171 |
+
},
|
| 172 |
+
"key_contributions": {
|
| 173 |
+
"type": "array",
|
| 174 |
+
"items": {"type": "string"}
|
| 175 |
+
},
|
| 176 |
+
"follow_up_required": {"type": "boolean"}
|
| 177 |
+
},
|
| 178 |
+
"required": ["name"]
|
| 179 |
+
}
|
| 180 |
+
},
|
| 181 |
+
|
| 182 |
+
"business_context": {
|
| 183 |
+
"type": "object",
|
| 184 |
+
"properties": {
|
| 185 |
+
"market_dynamics": {
|
| 186 |
+
"type": "object",
|
| 187 |
+
"properties": {
|
| 188 |
+
"industry_trends": {
|
| 189 |
+
"type": "array",
|
| 190 |
+
"items": {
|
| 191 |
+
"type": "object",
|
| 192 |
+
"properties": {
|
| 193 |
+
"trend": {"type": "string"},
|
| 194 |
+
"impact": {"type": "string"},
|
| 195 |
+
"relevance": {"type": "string"}
|
| 196 |
+
}
|
| 197 |
+
}
|
| 198 |
+
},
|
| 199 |
+
"competitive_landscape": {
|
| 200 |
+
"type": "object",
|
| 201 |
+
"properties": {
|
| 202 |
+
"direct_competitors": {
|
| 203 |
+
"type": "array",
|
| 204 |
+
"items": {
|
| 205 |
+
"type": "object",
|
| 206 |
+
"properties": {
|
| 207 |
+
"name": {"type": "string"},
|
| 208 |
+
"strengths": {"type": "array", "items": {"type": "string"}},
|
| 209 |
+
"weaknesses": {"type": "array", "items": {"type": "string"}},
|
| 210 |
+
"threat_level": {"type": "string"}
|
| 211 |
+
}
|
| 212 |
+
}
|
| 213 |
+
},
|
| 214 |
+
"market_position": {
|
| 215 |
+
"type": "object",
|
| 216 |
+
"properties": {
|
| 217 |
+
"current_position": {"type": "string"},
|
| 218 |
+
"desired_position": {"type": "string"},
|
| 219 |
+
"gaps": {"type": "array", "items": {"type": "string"}}
|
| 220 |
+
}
|
| 221 |
+
}
|
| 222 |
+
}
|
| 223 |
+
}
|
| 224 |
+
}
|
| 225 |
+
},
|
| 226 |
+
"strategic_alignment": {
|
| 227 |
+
"type": "object",
|
| 228 |
+
"properties": {
|
| 229 |
+
"company_objectives": {
|
| 230 |
+
"type": "array",
|
| 231 |
+
"items": {
|
| 232 |
+
"type": "object",
|
| 233 |
+
"properties": {
|
| 234 |
+
"objective": {"type": "string"},
|
| 235 |
+
"alignment_level": {"type": "string"},
|
| 236 |
+
"contribution": {"type": "string"}
|
| 237 |
+
}
|
| 238 |
+
}
|
| 239 |
+
},
|
| 240 |
+
"value_proposition": {
|
| 241 |
+
"type": "object",
|
| 242 |
+
"properties": {
|
| 243 |
+
"key_value_drivers": {"type": "array", "items": {"type": "string"}},
|
| 244 |
+
"differentiators": {"type": "array", "items": {"type": "string"}},
|
| 245 |
+
"client_benefits": {"type": "array", "items": {"type": "string"}}
|
| 246 |
+
}
|
| 247 |
+
}
|
| 248 |
+
}
|
| 249 |
+
},
|
| 250 |
+
"account_status": {
|
| 251 |
+
"type": "string",
|
| 252 |
+
"enum": [
|
| 253 |
+
"active_client",
|
| 254 |
+
"prospect",
|
| 255 |
+
"partner",
|
| 256 |
+
"competitor",
|
| 257 |
+
"other"
|
| 258 |
+
]
|
| 259 |
+
},
|
| 260 |
+
"project_phase": {"type": "string"},
|
| 261 |
+
"business_units_involved": {
|
| 262 |
+
"type": "array",
|
| 263 |
+
"items": {"type": "string"}
|
| 264 |
+
},
|
| 265 |
+
"current_contracts": {
|
| 266 |
+
"type": "array",
|
| 267 |
+
"items": {
|
| 268 |
+
"type": "object",
|
| 269 |
+
"properties": {
|
| 270 |
+
"name": {"type": "string"},
|
| 271 |
+
"status": {"type": "string"},
|
| 272 |
+
"value": {"type": "string"}
|
| 273 |
+
}
|
| 274 |
+
}
|
| 275 |
+
},
|
| 276 |
+
"strategic_importance": {
|
| 277 |
+
"type": "string",
|
| 278 |
+
"enum": ["high", "medium", "low"]
|
| 279 |
+
}
|
| 280 |
+
}
|
| 281 |
+
},
|
| 282 |
+
|
| 283 |
+
"discussion_topics": {
|
| 284 |
+
"type": "array",
|
| 285 |
+
"items": {
|
| 286 |
+
"type": "object",
|
| 287 |
+
"properties": {
|
| 288 |
+
"topic": {"type": "string"},
|
| 289 |
+
"duration": {"type": "string"},
|
| 290 |
+
"priority": {
|
| 291 |
+
"type": "string",
|
| 292 |
+
"enum": ["high", "medium", "low"]
|
| 293 |
+
},
|
| 294 |
+
"key_points": {
|
| 295 |
+
"type": "array",
|
| 296 |
+
"items": {"type": "string"}
|
| 297 |
+
},
|
| 298 |
+
"stakeholders_involved": {
|
| 299 |
+
"type": "array",
|
| 300 |
+
"items": {"type": "string"}
|
| 301 |
+
},
|
| 302 |
+
"outcomes": {
|
| 303 |
+
"type": "array",
|
| 304 |
+
"items": {"type": "string"}
|
| 305 |
+
},
|
| 306 |
+
"sentiment": {
|
| 307 |
+
"type": "string",
|
| 308 |
+
"enum": ["positive", "neutral", "negative", "mixed"]
|
| 309 |
+
}
|
| 310 |
+
}
|
| 311 |
+
}
|
| 312 |
+
},
|
| 313 |
+
|
| 314 |
+
"action_items": {
|
| 315 |
+
"type": "array",
|
| 316 |
+
"items": {
|
| 317 |
+
"type": "object",
|
| 318 |
+
"properties": {
|
| 319 |
+
"description": {"type": "string"},
|
| 320 |
+
"type": {
|
| 321 |
+
"type": "string",
|
| 322 |
+
"enum": [
|
| 323 |
+
"task",
|
| 324 |
+
"decision_needed",
|
| 325 |
+
"information_request",
|
| 326 |
+
"approval_required",
|
| 327 |
+
"follow_up"
|
| 328 |
+
]
|
| 329 |
+
},
|
| 330 |
+
"owner": {"type": "string"},
|
| 331 |
+
"assignees": {
|
| 332 |
+
"type": "array",
|
| 333 |
+
"items": {"type": "string"}
|
| 334 |
+
},
|
| 335 |
+
"due_date": {"type": "string"},
|
| 336 |
+
"priority": {
|
| 337 |
+
"type": "string",
|
| 338 |
+
"enum": ["high", "medium", "low"]
|
| 339 |
+
},
|
| 340 |
+
"status": {
|
| 341 |
+
"type": "string",
|
| 342 |
+
"enum": [
|
| 343 |
+
"not_started",
|
| 344 |
+
"in_progress",
|
| 345 |
+
"blocked",
|
| 346 |
+
"completed"
|
| 347 |
+
]
|
| 348 |
+
},
|
| 349 |
+
"dependencies": {
|
| 350 |
+
"type": "array",
|
| 351 |
+
"items": {"type": "string"}
|
| 352 |
+
},
|
| 353 |
+
"related_topic": {"type": "string"},
|
| 354 |
+
"notes": {"type": "string"}
|
| 355 |
+
}
|
| 356 |
+
}
|
| 357 |
+
},
|
| 358 |
+
|
| 359 |
+
"decisions": {
|
| 360 |
+
"type": "array",
|
| 361 |
+
"items": {
|
| 362 |
+
"type": "object",
|
| 363 |
+
"properties": {
|
| 364 |
+
"topic": {"type": "string"},
|
| 365 |
+
"decision": {"type": "string"},
|
| 366 |
+
"rationale": {"type": "string"},
|
| 367 |
+
"impact_level": {
|
| 368 |
+
"type": "string",
|
| 369 |
+
"enum": ["high", "medium", "low"]
|
| 370 |
+
},
|
| 371 |
+
"decision_makers": {
|
| 372 |
+
"type": "array",
|
| 373 |
+
"items": {"type": "string"}
|
| 374 |
+
},
|
| 375 |
+
"stakeholders_affected": {
|
| 376 |
+
"type": "array",
|
| 377 |
+
"items": {"type": "string"}
|
| 378 |
+
},
|
| 379 |
+
"implementation_timeline": {"type": "string"},
|
| 380 |
+
"dependencies": {
|
| 381 |
+
"type": "array",
|
| 382 |
+
"items": {"type": "string"}
|
| 383 |
+
},
|
| 384 |
+
"risks_identified": {
|
| 385 |
+
"type": "array",
|
| 386 |
+
"items": {"type": "string"}
|
| 387 |
+
}
|
| 388 |
+
}
|
| 389 |
+
}
|
| 390 |
+
},
|
| 391 |
+
|
| 392 |
+
"opportunities": {
|
| 393 |
+
"type": "array",
|
| 394 |
+
"items": {
|
| 395 |
+
"type": "object",
|
| 396 |
+
"properties": {
|
| 397 |
+
"name": {"type": "string"},
|
| 398 |
+
"type": {
|
| 399 |
+
"type": "string",
|
| 400 |
+
"enum": [
|
| 401 |
+
"new_business",
|
| 402 |
+
"expansion",
|
| 403 |
+
"renewal",
|
| 404 |
+
"upsell",
|
| 405 |
+
"cross_sell"
|
| 406 |
+
]
|
| 407 |
+
},
|
| 408 |
+
"estimated_value": {"type": "string"},
|
| 409 |
+
"probability": {"type": "string"},
|
| 410 |
+
"timeline": {"type": "string"},
|
| 411 |
+
"key_stakeholders": {
|
| 412 |
+
"type": "array",
|
| 413 |
+
"items": {"type": "string"}
|
| 414 |
+
},
|
| 415 |
+
"decision_makers": {
|
| 416 |
+
"type": "array",
|
| 417 |
+
"items": {"type": "string"}
|
| 418 |
+
},
|
| 419 |
+
"next_steps": {
|
| 420 |
+
"type": "array",
|
| 421 |
+
"items": {"type": "string"}
|
| 422 |
+
},
|
| 423 |
+
"competitive_situation": {
|
| 424 |
+
"type": "object",
|
| 425 |
+
"properties": {
|
| 426 |
+
"competitors": {
|
| 427 |
+
"type": "array",
|
| 428 |
+
"items": {"type": "string"}
|
| 429 |
+
},
|
| 430 |
+
"our_strengths": {
|
| 431 |
+
"type": "array",
|
| 432 |
+
"items": {"type": "string"}
|
| 433 |
+
},
|
| 434 |
+
"our_weaknesses": {
|
| 435 |
+
"type": "array",
|
| 436 |
+
"items": {"type": "string"}
|
| 437 |
+
}
|
| 438 |
+
}
|
| 439 |
+
}
|
| 440 |
+
}
|
| 441 |
+
}
|
| 442 |
+
},
|
| 443 |
+
|
| 444 |
+
"risks_and_issues": {
|
| 445 |
+
"type": "array",
|
| 446 |
+
"items": {
|
| 447 |
+
"type": "object",
|
| 448 |
+
"properties": {
|
| 449 |
+
"description": {"type": "string"},
|
| 450 |
+
"type": {
|
| 451 |
+
"type": "string",
|
| 452 |
+
"enum": [
|
| 453 |
+
"technical",
|
| 454 |
+
"commercial",
|
| 455 |
+
"operational",
|
| 456 |
+
"strategic",
|
| 457 |
+
"compliance"
|
| 458 |
+
]
|
| 459 |
+
},
|
| 460 |
+
"severity": {
|
| 461 |
+
"type": "string",
|
| 462 |
+
"enum": ["high", "medium", "low"]
|
| 463 |
+
},
|
| 464 |
+
"probability": {
|
| 465 |
+
"type": "string",
|
| 466 |
+
"enum": ["high", "medium", "low"]
|
| 467 |
+
},
|
| 468 |
+
"raised_by": {"type": "string"},
|
| 469 |
+
"impact_areas": {
|
| 470 |
+
"type": "array",
|
| 471 |
+
"items": {"type": "string"}
|
| 472 |
+
},
|
| 473 |
+
"mitigation_plan": {"type": "string"},
|
| 474 |
+
"owner": {"type": "string"},
|
| 475 |
+
"status": {
|
| 476 |
+
"type": "string",
|
| 477 |
+
"enum": [
|
| 478 |
+
"identified",
|
| 479 |
+
"being_mitigated",
|
| 480 |
+
"mitigated",
|
| 481 |
+
"accepted"
|
| 482 |
+
]
|
| 483 |
+
}
|
| 484 |
+
}
|
| 485 |
+
}
|
| 486 |
+
},
|
| 487 |
+
|
| 488 |
+
"interaction_dynamics": {
|
| 489 |
+
"type": "object",
|
| 490 |
+
"properties": {
|
| 491 |
+
"communication_patterns": {
|
| 492 |
+
"type": "array",
|
| 493 |
+
"items": {
|
| 494 |
+
"type": "object",
|
| 495 |
+
"properties": {
|
| 496 |
+
"participants": {"type": "array", "items": {"type": "string"}},
|
| 497 |
+
"pattern_type": {"type": "string"},
|
| 498 |
+
"frequency": {"type": "string"},
|
| 499 |
+
"effectiveness": {"type": "string"},
|
| 500 |
+
"power_dynamics": {
|
| 501 |
+
"type": "array",
|
| 502 |
+
"items": {
|
| 503 |
+
"type": "object",
|
| 504 |
+
"properties": {
|
| 505 |
+
"actor": {"type": "string"},
|
| 506 |
+
"influence_type": {"type": "string"},
|
| 507 |
+
"impact": {"type": "string"}
|
| 508 |
+
}
|
| 509 |
+
}
|
| 510 |
+
}
|
| 511 |
+
}
|
| 512 |
+
}
|
| 513 |
+
},
|
| 514 |
+
"emotional_intelligence": {
|
| 515 |
+
"type": "object",
|
| 516 |
+
"properties": {
|
| 517 |
+
"group_dynamics": {
|
| 518 |
+
"type": "array",
|
| 519 |
+
"items": {
|
| 520 |
+
"type": "object",
|
| 521 |
+
"properties": {
|
| 522 |
+
"dynamic_type": {"type": "string"},
|
| 523 |
+
"intensity": {"type": "string"},
|
| 524 |
+
"impact": {"type": "string"}
|
| 525 |
+
}
|
| 526 |
+
}
|
| 527 |
+
},
|
| 528 |
+
"emotional_triggers": {
|
| 529 |
+
"type": "array",
|
| 530 |
+
"items": {
|
| 531 |
+
"type": "object",
|
| 532 |
+
"properties": {
|
| 533 |
+
"trigger": {"type": "string"},
|
| 534 |
+
"response": {"type": "string"},
|
| 535 |
+
"participants_affected": {"type": "array", "items": {"type": "string"}}
|
| 536 |
+
}
|
| 537 |
+
}
|
| 538 |
+
}
|
| 539 |
+
}
|
| 540 |
+
}
|
| 541 |
+
}
|
| 542 |
+
},
|
| 543 |
+
|
| 544 |
+
"relationship_insights": {
|
| 545 |
+
"type": "array",
|
| 546 |
+
"items": {
|
| 547 |
+
"type": "object",
|
| 548 |
+
"properties": {
|
| 549 |
+
"stakeholder": {"type": "string"},
|
| 550 |
+
"influence_level": {
|
| 551 |
+
"type": "string",
|
| 552 |
+
"enum": ["high", "medium", "low"]
|
| 553 |
+
},
|
| 554 |
+
"sentiment": {
|
| 555 |
+
"type": "string",
|
| 556 |
+
"enum": ["positive", "neutral", "negative", "mixed"]
|
| 557 |
+
},
|
| 558 |
+
"key_interests": {
|
| 559 |
+
"type": "array",
|
| 560 |
+
"items": {"type": "string"}
|
| 561 |
+
},
|
| 562 |
+
"concerns": {
|
| 563 |
+
"type": "array",
|
| 564 |
+
"items": {"type": "string"}
|
| 565 |
+
},
|
| 566 |
+
"relationships": {
|
| 567 |
+
"type": "array",
|
| 568 |
+
"items": {
|
| 569 |
+
"type": "object",
|
| 570 |
+
"properties": {
|
| 571 |
+
"with": {"type": "string"},
|
| 572 |
+
"nature": {"type": "string"},
|
| 573 |
+
"strength": {
|
| 574 |
+
"type": "string",
|
| 575 |
+
"enum": ["strong", "moderate", "weak"]
|
| 576 |
+
}
|
| 577 |
+
}
|
| 578 |
+
}
|
| 579 |
+
}
|
| 580 |
+
}
|
| 581 |
+
}
|
| 582 |
+
}
|
| 583 |
+
}
|
| 584 |
}
|
| 585 |
|
| 586 |
+
try:
|
| 587 |
+
self.extraction_chain = create_extraction_chain(
|
| 588 |
+
self.entity_schema,
|
| 589 |
+
self.llm_service.extraction_model
|
| 590 |
+
)
|
| 591 |
+
logger.info("Entity extraction chain initialized successfully")
|
| 592 |
+
except Exception as e:
|
| 593 |
+
logger.error(f"Failed to initialize extraction chain: {str(e)}")
|
| 594 |
+
raise
|
| 595 |
|
| 596 |
async def extract_entities(self, text: str) -> Dict[str, Any]:
|
| 597 |
+
"""
|
| 598 |
+
Extract comprehensive entities and insights from text.
|
| 599 |
+
|
| 600 |
+
Args:
|
| 601 |
+
text: Meeting transcript or text to analyze
|
| 602 |
+
|
| 603 |
+
Returns:
|
| 604 |
+
Dictionary containing extracted entities, relationships, and business context
|
| 605 |
+
|
| 606 |
+
Raises:
|
| 607 |
+
Exception: If extraction fails
|
| 608 |
+
"""
|
| 609 |
+
try:
|
| 610 |
+
extracted_data = await self.extraction_chain.arun(text)
|
| 611 |
+
|
| 612 |
+
# Add extraction timestamp
|
| 613 |
+
extracted_data['metadata'] = {
|
| 614 |
+
'extraction_timestamp': datetime.now().isoformat(),
|
| 615 |
+
'schema_version': '2.0'
|
| 616 |
+
}
|
| 617 |
+
|
| 618 |
+
logger.info("Entity extraction completed successfully")
|
| 619 |
+
return extracted_data
|
| 620 |
+
|
| 621 |
+
except Exception as e:
|
| 622 |
+
logger.error(f"Entity extraction failed: {str(e)}")
|
| 623 |
+
raise Exception(f"Failed to extract entities: {str(e)}")
|
| 624 |
+
|
| 625 |
+
def validate_extraction(self, extracted_data: Dict[str, Any]) -> bool:
|
| 626 |
+
"""
|
| 627 |
+
Validate extracted data against schema requirements.
|
| 628 |
+
|
| 629 |
+
Args:
|
| 630 |
+
extracted_data: Dictionary of extracted information
|
| 631 |
+
|
| 632 |
+
Returns:
|
| 633 |
+
Boolean indicating validation status
|
| 634 |
+
"""
|
| 635 |
+
try:
|
| 636 |
+
# Implement validation logic here
|
| 637 |
+
# This could check for required fields, data consistency, etc.
|
| 638 |
+
return True
|
| 639 |
+
except Exception as e:
|
| 640 |
+
logger.error(f"Validation failed: {str(e)}")
|
| 641 |
+
return False
|