adaptive-memory-multi-model-router 2.14.46 → 2.14.47

package/tmlpd-pi-extension/src/utils/compression.ts

@@ -1,325 +0,0 @@
-/**
- * TMLPD Context Compression Utilities
- * 
- * Strategies for reducing context window usage:
- * - Smart truncation
- * - Message summarization
- * - ISON-style encoding (inspired by Reddit ISON format)
- * - Context window management
- */
-
-export interface Message {
-  role: "system" | "user" | "assistant";
-  content: string;
-  name?: string;
-  tool_calls?: any[];
-  tool_call_id?: string;
-}
-
-export type CompressionStrategy = "smart" | "first" | "last" | "isentropy";
-
-/**
- * ISON (Intelligence-Sparse Object Notation)
- * 
- * A compression format that reduces token count by:
- * - Removing redundant whitespace
- * - Shortening common phrases
- * - Using abbreviations strategically
- * 
- * Example: "The quick brown fox jumps over the lazy dog"
- *      → "quick brown fox jumps lazy dog" (removes articles, repeated words)
- */
-
-const ISON_REPLACEMENTS: Array<[RegExp, string]> = [
-  // Common phrase abbreviations
-  [/\bthe\b/g, ""],
-  [/\ba\b/g, ""],
-  [/\ban\b/g, ""],
-  [/\bthat\b/g, "that"],
-  [/\bthis\b/g, "this"],
-  [/\bwith\b/g, "w/"],
-  [/\bwithout\b/g, "w/o"],
-  [/\band\b/g, "&"],
-  [/\bor\b/g, "|"],
-  [/\bfor\b/g, "4"],
-  [/\bto\b/g, "2"],
-  [/\binto\b/g, "2"],
-  [/\bfrom\b/g, "fr"],
-  [/\bplease\b/gi, ""],
-  [/\bthank you\b/gi, "thx"],
-  [/\byou are\b/gi, "u r"],
-  [/\byou can\b/gi, "u c"],
-  [/\bcan you\b/gi, "c?"],
-  [/\bhow do\b/gi, "how 2"],
-  [/\bwhat is\b/gi, "wat"],
-  [/\bwhat are\b/gi, "wat"],
-  
-  // Whitespace normalization
-  [/\s+/g, " "],
-  [/^\s+|\s+$/g, ""],
-  
-  // Remove repeated characters
-  [/(\w)\1{2,}/g, "$1$1"],
-  
-  // Shorten common technical terms
-  [/\binformation\b/gi, "info"],
-  [/\bprocessing\b/gi, "proc"],
-  [/\bdevelopment\b/gi, "dev"],
-  [/\bapplication\b/gi, "app"],
-  [/\bconfiguration\b/gi, "config"],
-  [/\brepresentation\b/gi, "repr"],
-  [/\bunderstanding\b/gi, "unders"],
-  [/\brecommendation\b/gi, "rec"],
-];
-
-const ISON_UNREPLACEMENTS: Array<[RegExp, string]> = [
-  [/w\//g, "with "],
-  [/w\/o/g, "without "],
-  [/&/g, " and "],
-  [/\b4\b/g, " for "],
-  [/\b2\b/g, " to "],
-  [/\bfr\b/g, "from "],
-  [/\bthx\b/gi, "thank you"],
-  [/\bu r\b/gi, "you are"],
-  [/\bu c\b/gi, "you can"],
-  [/\bc\?\b/g, "can you"],
-  [/how 2\b/gi, "how do"],
-  [/\bwat\b/g, "what is"],
-];
-
-export interface CompressionResult {
-  original_tokens: number;
-  compressed_tokens: number;
-  ratio: number;
-  compressed_text: string;
-}
-
-/**
- * ISON encode text for token reduction.
- */
-export function isonEncode(text: string): string {
-  let result = text;
-  
-  // Apply replacements
-  for (const [pattern, replacement] of ISON_REPLACEMENTS) {
-    result = result.replace(pattern, replacement);
-  }
-  
-  // Remove extra spaces and trim
-  result = result.replace(/\s+/g, " ").trim();
-  
-  return result;
-}
-
-/**
- * Decode ISON encoded text.
- */
-export function isonDecode(text: string): string {
-  let result = text;
-  
-  // Apply un-replacements
-  for (const [pattern, replacement] of ISON_UNREPLACEMENTS) {
-    result = result.replace(pattern, replacement);
-  }
-  
-  return result;
-}
-
-/**
- * Compress text using ISON encoding.
- */
-export function compressText(text: string): CompressionResult {
-  const original_tokens = estimateTokens(text);
-  const compressed = isonEncode(text);
-  const compressed_tokens = estimateTokens(compressed);
-  
-  return {
-    original_tokens,
-    compressed_tokens,
-    ratio: compressed_tokens / original_tokens,
-    compressed_text: compressed
-  };
-}
-
-/**
- * Estimate tokens (fallback if no model specified).
- */
-function estimateTokens(text: string): number {
-  const words = text.trim().split(/\s+/).filter(w => w.length > 0);
-  return Math.ceil(words.length * 1.3);
-}
-
-/**
- * Truncate messages to fit within token budget.
- * 
- * @param messages - Conversation messages
- * @param max_tokens - Maximum tokens allowed
- * @param strategy - "smart" (preserve system + recent), "first" (keep start), "last" (keep end)
- */
-export function truncateMessages(
-  messages: Message[],
-  max_tokens: number,
-  strategy: CompressionStrategy = "smart"
-): Message[] {
-  if (!messages || messages.length === 0) return [];
-  
-  // Calculate total tokens
-  const totalTokens = (msg: Message) => {
-    const content = typeof msg.content === 'string' ? msg.content : JSON.stringify(msg.content);
-    return estimateTokens(content) + 4; // +4 for role/format overhead
-  };
-  
-  let currentTokens = messages.reduce((sum, m) => sum + totalTokens(m), 0);
-  
-  if (currentTokens <= max_tokens) {
-    return messages; // Already fits
-  }
-  
-  const result: Message[] = [];
-  
-  if (strategy === "first") {
-    // Keep system (first) messages, truncate from middle
-    let keepTokens = 0;
-    for (const msg of messages) {
-      const msgTokens = totalTokens(msg);
-      if (keepTokens + msgTokens <= max_tokens) {
-        result.push(msg);
-        keepTokens += msgTokens;
-      } else if (msg.role === "system" && result.length === 0) {
-        // Always keep system message, possibly truncated
-        const systemContent = typeof msg.content === 'string' ? msg.content : "";
-        const truncated = truncateToTokenBudget(systemContent, max_tokens - 4);
-        result.push({ ...msg, content: truncated });
-        break;
-      } else {
-        break;
-      }
-    }
-  } else if (strategy === "last") {
-    // Keep only most recent messages
-    for (let i = messages.length - 1; i >= 0; i--) {
-      const msg = messages[i];
-      const msgTokens = totalTokens(msg);
-      
-      if (currentTokens - msgTokens <= max_tokens) {
-        result.unshift(msg);
-        currentTokens -= msgTokens;
-      } else if (msg.role === "user") {
-        // Try to keep a truncated user message
-        const truncated = truncateToTokenBudget(msg.content as string, max_tokens - currentTokens);
-        if (truncated.length > 20) {
-          result.unshift({ ...msg, content: truncated });
-        }
-        break;
-      }
-    }
-  } else if (strategy === "smart") {
-    // Keep system, compress middle, keep recent
-    const systemMessages: Message[] = [];
-    const middleMessages: Message[] = [];
-    const recentMessages: Message[] = [];
-    
-    for (const msg of messages) {
-      if (msg.role === "system") {
-        systemMessages.push(msg);
-      } else if (messages.indexOf(msg) >= messages.length - 3) {
-        recentMessages.push(msg);
-      } else {
-        middleMessages.push(msg);
-      }
-    }
-    
-    // Start with system
-    for (const msg of systemMessages) {
-      const msgTokens = totalTokens(msg);
-      if (currentTokens <= max_tokens) {
-        result.push(msg);
-        currentTokens -= msgTokens;
-      } else {
-        // Truncate system message
-        const truncated = truncateToTokenBudget(msg.content as string, max_tokens - currentTokens - 10);
-        result.push({ ...msg, content: truncated });
-        currentTokens = max_tokens;
-        break;
-      }
-    }
-    
-    // Add compressed middle
-    if (currentTokens > max_tokens / 2 && middleMessages.length > 0) {
-      // Compress middle messages into a summary
-      const middleContent = middleMessages
-        .map(m => m.content)
-        .join("\n");
-      const summaryTokenBudget = Math.min(
-        max_tokens / 4,
-        max_tokens - currentTokens
-      );
-      const summary = truncateToTokenBudget(
-        `[Previous ${middleMessages.length} messages]: ${middleContent}`,
-        summaryTokenBudget
-      );
-      result.push({ role: "assistant", content: summary });
-      currentTokens -= estimateTokens(summary);
-    }
-    
-    // Add recent messages if room
-    for (const msg of recentMessages) {
-      const msgTokens = totalTokens(msg);
-      if (currentTokens + msgTokens <= max_tokens) {
-        result.push(msg);
-        currentTokens += msgTokens;
-      }
-    }
-  }
-  
-  return result;
-}
-
-/**
- * Truncate a single string to fit within token budget.
- */
-export function truncateToTokenBudget(text: string, max_tokens: number): string {
-  const words = text.split(/\s+/);
-  let current = 0;
-  const targetWords: string[] = [];
-  
-  for (const word of words) {
-    const wordTokens = estimateTokens(word);
-    if (current + wordTokens <= max_tokens) {
-      targetWords.push(word);
-      current += wordTokens;
-    } else {
-      break;
-    }
-  }
-  
-  let result = targetWords.join(" ");
-  
-  // If we truncated, add ellipsis
-  if (result.length < text.length) {
-    result += "...";
-  }
-  
-  return result;
-}
-
-/**
- * Calculate compression ratio for context.
- */
-export function calculateCompressionRatio(messages: Message[], max_tokens: number): number {
-  const totalTokens = messages.reduce((sum, m) => {
-    const content = typeof m.content === 'string' ? m.content : JSON.stringify(m.content);
-    return sum + estimateTokens(content) + 4;
-  }, 0);
-  
-  return Math.min(1, max_tokens / totalTokens);
-}
-
-export default {
-  isonEncode,
-  isonDecode,
-  compressText,
-  truncateMessages,
-  truncateToTokenBudget,
-  calculateCompressionRatio
-};

package/tmlpd-pi-extension/src/utils/reliability.ts

@@ -1,221 +0,0 @@
-/**
- * TMLPD Reliability Engine
- * 
- * Circuit breaker, retry with jitter, and enhanced cooldown logic.
- * Designed to handle flaky API calls gracefully.
- */
-
-export interface RetryConfig {
-  max_attempts: number;
-  base_delay_ms: number;
-  max_delay_ms: number;
-  jitter: number; // 0-1, percentage of delay to randomize
-  retryable_status_codes: number[];
-}
-
-export interface CircuitBreakerConfig {
-  failure_threshold: number;
-  recovery_timeout_ms: number;
-  half_open_max_calls: number;
-}
-
-export interface CircuitState {
-  status: "closed" | "open" | "half_open";
-  failure_count: number;
-  last_failure_time: number | null;
-  last_success_time: number | null;
-  consecutive_successes: number;
-}
-
-export const DEFAULT_RETRY_CONFIG: RetryConfig = {
-  max_attempts: 3,
-  base_delay_ms: 500,
-  max_delay_ms: 30000,
-  jitter: 0.3,
-  retryable_status_codes: [408, 429, 500, 502, 503, 504],
-};
-
-export const DEFAULT_CIRCUIT_BREAKER_CONFIG: CircuitBreakerConfig = {
-  failure_threshold: 5,
-  recovery_timeout_ms: 60000,
-  half_open_max_calls: 3,
-};
-
-/**
- * Calculate delay with exponential backoff and jitter
- */
-export function calculateRetryDelay(
-  attempt: number,
-  config: RetryConfig = DEFAULT_RETRY_CONFIG
-): number {
-  // Exponential backoff
-  const exponential_delay = config.base_delay_ms * Math.pow(2, attempt - 1);
-  
-  // Cap at max delay
-  const capped_delay = Math.min(exponential_delay, config.max_delay_ms);
-  
-  // Add jitter
-  const jitter_range = capped_delay * config.jitter;
-  const jitter = (Math.random() * 2 - 1) * jitter_range;
-  
-  return Math.round(capped_delay + jitter);
-}
-
-/**
- * Check if status code is retryable
- */
-export function isRetryableStatus(statusCode: number | null, config: RetryConfig = DEFAULT_RETRY_CONFIG): boolean {
-  if (statusCode === null) return true; // Network errors are retryable
-  return config.retryable_status_codes.includes(statusCode);
-}
-
-/**
- * Circuit Breaker implementation
- */
-export class CircuitBreaker {
-  private config: CircuitBreakerConfig;
-  private state: CircuitState;
-  private half_open_calls = 0;
-
-  constructor(config: Partial<CircuitBreakerConfig> = {}) {
-    this.config = { ...DEFAULT_CIRCUIT_BREAKER_CONFIG, ...config };
-    this.state = {
-      status: "closed",
-      failure_count: 0,
-      last_failure_time: null,
-      last_success_time: null,
-      consecutive_successes: 0,
-    };
-  }
-
-  /**
-   * Check if circuit allows requests
-   */
-  canExecute(): boolean {
-    if (this.state.status === "closed") return true;
-
-    if (this.state.status === "open") {
-      // Check if recovery timeout has passed
-      if (this.state.last_failure_time && Date.now() - this.state.last_failure_time >= this.config.recovery_timeout_ms) {
-        this.state.status = "half_open";
-        this.half_open_calls = 0;
-        return true;
-      }
-      return false;
-    }
-
-    // half_open
-    return this.half_open_calls < this.config.half_open_max_calls;
-  }
-
-  /**
-   * Record a successful execution
-   */
-  recordSuccess(): void {
-    this.state.last_success_time = Date.now();
-    this.half_open_calls++;
-
-    if (this.state.status === "half_open") {
-      this.state.consecutive_successes++;
-      if (this.state.consecutive_successes >= this.config.half_open_max_calls) {
-        // Circuit recovered
-        this.state.status = "closed";
-        this.state.failure_count = 0;
-        this.state.consecutive_successes = 0;
-      }
-    } else {
-      this.state.failure_count = 0;
-    }
-  }
-
-  /**
-   * Record a failed execution
-   */
-  recordFailure(): void {
-    this.state.last_failure_time = Date.now();
-    this.state.failure_count++;
-    this.state.consecutive_successes = 0;
-
-    if (this.state.status === "half_open") {
-      // Trip circuit back open
-      this.state.status = "open";
-    } else if (this.state.failure_count >= this.config.failure_threshold) {
-      this.state.status = "open";
-    }
-  }
-
-  /**
-   * Get current circuit state
-   */
-  getState(): CircuitState {
-    return { ...this.state };
-  }
-
-  /**
-   * Force reset circuit
-   */
-  reset(): void {
-    this.state = {
-      status: "closed",
-      failure_count: 0,
-      last_failure_time: null,
-      last_success_time: null,
-      consecutive_successes: 0,
-    };
-    this.half_open_calls = 0;
-  }
-}
-
-/**
- * Enhanced retry wrapper with circuit breaker integration
- */
-export async function withRetry<T>(
-  fn: () => Promise<T>,
-  config: Partial<RetryConfig> = {},
-  circuitBreaker?: CircuitBreaker
-): Promise<{ result: T | null; error: Error | null; attempts: number; circuit_tripped: boolean }> {
-  const retryConfig = { ...DEFAULT_RETRY_CONFIG, ...config };
-  let lastError: Error | null = null;
-  let attempts = 0;
-  let circuit_tripped = false;
-
-  for (let i = 0; i < retryConfig.max_attempts; i++) {
-    attempts++;
-
-    try {
-      // Check circuit breaker before attempt
-      if (circuitBreaker && !circuitBreaker.canExecute()) {
-        circuit_tripped = true;
-        throw new Error("Circuit breaker is open");
-      }
-
-      const result = await fn();
-      
-      if (circuitBreaker) {
-        circuitBreaker.recordSuccess();
-      }
-      
-      return { result, error: null, attempts, circuit_tripped };
-    } catch (error) {
-      lastError = error instanceof Error ? error : new Error(String(error));
-
-      // Check if should retry
-      const statusCode = (error as any).statusCode || (error as any).response?.statusCode || null;
-      if (!isRetryableStatus(statusCode, retryConfig)) {
-        return { result: null, error: lastError, attempts, circuit_tripped };
-      }
-
-      if (circuitBreaker) {
-        circuitBreaker.recordFailure();
-      }
-
-      // Don't wait after last attempt
-      if (i < retryConfig.max_attempts - 1) {
-        const delay = calculateRetryDelay(i + 1, retryConfig);
-        await new Promise((resolve) => setTimeout(resolve, delay));
-      }
-    }
-  }
-
-  return { result: null, error: lastError, attempts, circuit_tripped };
-}

package/tmlpd-pi-extension/src/utils/tokenUtils.ts

@@ -1,145 +0,0 @@
-/**
- * TMLPD Token Utilities
- * 
- * Token counting, cost estimation, and context management.
- */
-
-export interface TokenCost {
-  input_per_1k: number;
-  output_per_1k: number;
-}
-
-// Current provider rates (2024-2025)
-export const MODEL_COSTS: Record<string, TokenCost> = {
-  // OpenAI
-  "gpt-4o": { input_per_1k: 2.50, output_per_1k: 10.00 },
-  "gpt-4o-mini": { input_per_1k: 0.15, output_per_1k: 0.60 },
-  "gpt-4-turbo": { input_per_1k: 10.00, output_per_1k: 30.00 },
-  "gpt-3.5-turbo": { input_per_1k: 0.50, output_per_1k: 1.50 },
-  
-  // Anthropic
-  "claude-3.5-sonnet": { input_per_1k: 3.00, output_per_1k: 15.00 },
-  "claude-3-opus": { input_per_1k: 15.00, output_per_1k: 75.00 },
-  "claude-3-haiku": { input_per_1k: 0.25, output_per_1k: 1.25 },
-  
-  // Google
-  "gemini-2.0-flash": { input_per_1k: 0.00, output_per_1k: 0.00 },  // Free
-  "gemini-1.5-pro": { input_per_1k: 1.25, output_per_1k: 5.00 },
-  "gemini-1.5-flash": { input_per_1k: 0.075, output_per_1k: 0.30 },
-  
-  // Groq
-  "groq/llama-3.3-70b": { input_per_1k: 0.59, output_per_1k: 0.79 },
-  "groq/llama-3.1-8b": { input_per_1k: 0.05, output_per_1k: 0.08 },
-  
-  // Cerebras
-  "cerebras/llama-3.3-70b": { input_per_1k: 0.60, output_per_1k: 0.60 },
-  
-  // Mistral
-  "mistral-large": { input_per_1k: 2.00, output_per_1k: 6.00 },
-  "mistral-small": { input_per_1k: 0.20, output_per_1k: 0.60 },
-};
-
-/**
- * Count tokens in text (approximate for English).
- * Based on ~1.3 tokens per word for typical English text.
- */
-export function countTokens(text: string, model: string = "gpt-4o"): number {
-  if (!text || text.length === 0) return 0;
-  
-  // Use model-specific approximation if available
-  // Otherwise use generic word-based estimate
-  const words = text.trim().split(/\s+/).length;
-  
-  // Fine-tune based on model family
-  if (model.includes("claude")) {
-    // Anthropic models: ~1.5 tokens per word
-    return Math.ceil(words * 1.5);
-  } else if (model.includes("gemini")) {
-    // Google: ~1.2 tokens per word (SentencePiece)
-    return Math.ceil(words * 1.2);
-  } else if (model.includes("llama")) {
-    // Llama: ~1.4 tokens per word (BPE)
-    return Math.ceil(words * 1.4);
-  }
-  
-  // Default: ~1.3 tokens per word (GPT-4 average)
-  return Math.ceil(words * 1.3);
-}
-
-/**
- * Estimate cost for a prompt/completion pair.
- */
-export function estimateCost(
-  prompt_tokens: number,
-  completion_tokens: number,
-  model: string
-): number {
-  const costs = MODEL_COSTS[model] || MODEL_COSTS["gpt-4o"];
-  
-  const input_cost = (prompt_tokens / 1000) * costs.input_per_1k;
-  const output_cost = (completion_tokens / 1000) * costs.output_per_1k;
-  
-  return input_cost + output_cost;
-}
-
-/**
- * Estimate cost from raw text (approximates both prompt and completion).
- */
-export function estimateCostFromText(
-  prompt: string,
-  completion: string,
-  model: string
-): number {
-  const prompt_tokens = countTokens(prompt, model);
-  // Completion typically has higher token density
-  const completion_tokens = Math.ceil(countTokens(completion, model) * 1.2);
-  
-  return estimateCost(prompt_tokens, completion_tokens, model);
-}
-
-/**
- * Get cost info for a model.
- */
-export function getModelCost(model: string): TokenCost {
-  return MODEL_COSTS[model] || MODEL_COSTS["gpt-4o"];
-}
-
-/**
- * List all supported models with their costs.
- */
-export function listModelsByCost(): Array<{ model: string; input: number; output: number }> {
-  return Object.entries(MODEL_COSTS)
-    .map(([model, cost]) => ({
-      model,
-      input: cost.input_per_1k,
-      output: cost.output_per_1k
-    }))
-    .sort((a, b) => (a.input + a.output) - (b.input + b.output));
-}
-
-/**
- * Find cheapest models for a given task.
- */
-export function findCheapestModels(task: "fast" | "quality" | "balanced" | "coding", count: number = 3): string[] {
-  const sorted = listModelsByCost();
-  
-  // Different profiles for different needs
-  const profiles = {
-    fast: sorted.filter(m => m.output < 1.0).slice(0, count).map(m => m.model),
-    quality: sorted.filter(m => m.output > 10).slice(0, count).map(m => m.model),
-    balanced: sorted.slice(0, count * 2).slice(count, count * 2).map(m => m.model),
-    coding: sorted.filter(m => m.model.includes("codex") || m.model.includes("claude") || m.model.includes("llama")).slice(0, count).map(m => m.model)
-  };
-  
-  return profiles[task] || profiles.balanced;
-}
-
-export default {
-  countTokens,
-  estimateCost,
-  estimateCostFromText,
-  getModelCost,
-  listModelsByCost,
-  findCheapestModels,
-  MODEL_COSTS
-};

package/tmlpd-pi-extension/tsconfig.json

@@ -1,18 +0,0 @@
-{
-  "compilerOptions": {
-    "target": "ES2022",
-    "module": "commonjs",
-    "lib": ["ES2022"],
-    "outDir": "./dist",
-    "rootDir": "./src",
-    "strict": true,
-    "esModuleInterop": true,
-    "skipLibCheck": true,
-    "forceConsistentCasingInFileNames": true,
-    "declaration": true,
-    "declarationMap": true,
-    "sourceMap": true
-  },
-  "include": ["src/**/*"],
-  "exclude": ["node_modules", "dist"]
-}