Add AI NLP and router modules for advanced natural language processing

- Introduced comprehensive NLP processing modules for intent classification, entity extraction, and context analysis
- Created AI router with rate-limited endpoints for command interpretation and execution
- Added prompt templates for different AI models with configurable system and user prompts
- Implemented robust type definitions for AI-related interfaces and schemas
- Enhanced security and error handling in AI processing pipeline
This commit is contained in:
jango-blockchained
2025-01-30 10:05:57 +01:00
parent 3115c41e87
commit c64dc4334b
13 changed files with 1274 additions and 37 deletions

132
src/ai/nlp/processor.ts Normal file
View File

@@ -0,0 +1,132 @@
import { AIIntent, AIContext, AIConfidence, AIError } from '../types/index.js';
import { EntityExtractor } from './entity-extractor.js';
import { IntentClassifier } from './intent-classifier.js';
import { ContextAnalyzer } from './context-analyzer.js';
export class NLPProcessor {
private entityExtractor: EntityExtractor;
private intentClassifier: IntentClassifier;
private contextAnalyzer: ContextAnalyzer;
constructor() {
this.entityExtractor = new EntityExtractor();
this.intentClassifier = new IntentClassifier();
this.contextAnalyzer = new ContextAnalyzer();
}
async processCommand(
input: string,
context: AIContext
): Promise<{
intent: AIIntent;
confidence: AIConfidence;
error?: AIError;
}> {
try {
// Extract entities from the input
const entities = await this.entityExtractor.extract(input);
// Classify the intent
const intent = await this.intentClassifier.classify(input, entities);
// Analyze context relevance
const contextRelevance = await this.contextAnalyzer.analyze(intent, context);
// Calculate confidence scores
const confidence: AIConfidence = {
overall: (intent.confidence + entities.confidence + contextRelevance.confidence) / 3,
intent: intent.confidence,
entities: entities.confidence,
context: contextRelevance.confidence
};
// Create structured intent
const structuredIntent: AIIntent = {
action: intent.action,
target: entities.primary_target,
parameters: {
...entities.parameters,
...intent.parameters,
context_parameters: contextRelevance.relevant_params
},
raw_input: input
};
return {
intent: structuredIntent,
confidence
};
} catch (error: unknown) {
const errorMessage = error instanceof Error ? error.message : 'Unknown error occurred';
return {
intent: {
action: 'error',
target: 'system',
parameters: {},
raw_input: input
},
confidence: {
overall: 0,
intent: 0,
entities: 0,
context: 0
},
error: {
code: 'NLP_PROCESSING_ERROR',
message: errorMessage,
suggestion: 'Please try rephrasing your command',
recovery_options: [
'Use simpler language',
'Break down the command into smaller parts',
'Specify the target device explicitly'
],
context
}
};
}
}
async validateIntent(
intent: AIIntent,
confidence: AIConfidence,
threshold = 0.7
): Promise<boolean> {
return (
confidence.overall >= threshold &&
confidence.intent >= threshold &&
confidence.entities >= threshold &&
confidence.context >= threshold
);
}
async suggestCorrections(
input: string,
error: AIError
): Promise<string[]> {
// Implement correction suggestions based on the error
const suggestions: string[] = [];
if (error.code === 'ENTITY_NOT_FOUND') {
suggestions.push(
'Try specifying the device name more clearly',
'Use the exact device name from your Home Assistant setup'
);
}
if (error.code === 'AMBIGUOUS_INTENT') {
suggestions.push(
'Please specify what you want to do with the device',
'Use action words like "turn on", "set", "adjust"'
);
}
if (error.code === 'CONTEXT_MISMATCH') {
suggestions.push(
'Specify the location if referring to a device',
'Clarify which device you mean in the current context'
);
}
return suggestions;
}
}