codymaster 4.8.0 → 5.2.0

package/scripts/viking-demo.ts

@@ -1,105 +0,0 @@
-import express from 'express';
-import { VikingBackend } from '../src/backends/viking-backend';
-import { DEFAULT_VIKING_CONFIG } from '../src/backends/viking-http-client';
-import chalk from 'chalk';
-
-/**
- * Viking Demo — Hamster-Powered Semantic Search
- * 
- * This script demonstrates the CodyMaster v4.6 integration with OpenViking.
- * It starts a mock OpenViking REST API and then uses CodyMaster's VikingBackend 
- * to perform a semantic search query.
- */
-
-async function runDemo() {
-  const PORT = 1933;
-  const app = express();
-  app.use(express.json());
-
-  console.log(chalk.bold.cyan('\n🐹 CodyMaster v4.6 — OpenViking Integration Demo\n'));
-
-  // ─── Phase 1: Mock OpenViking Server ───────────────────────────────────────
-  
-  console.log(chalk.dim('  [1/3] Starting Mock OpenViking Server...'));
-
-  // Mock /health endpoint
-  app.get('/health', (req, res) => res.json({ status: 'ok', version: '1.2.0' }));
-
-  // Mock /search endpoint (the core of Viking)
-  app.post('/search', (req, res) => {
-    const { query, limit, workspace } = req.body;
-    console.log(chalk.yellow(`        ☁️  Viking Server received search: "${query}" (limit: ${limit}, ws: ${workspace})`));
-    
-    const mockResults = [
-      {
-        uri: 'ov://demo-vibe/learnings/lear-882.json',
-        score: 0.98,
-        content: JSON.stringify({
-          what_failed: 'Async fetch timeout on slow networks',
-          why_failed: 'Default timeout too short for 3G/4G high-latency spikes',
-          how_to_prevent: 'Increase timeout to 30s + implement exponential backoff retry'
-        })
-      },
-      {
-        uri: 'ov://demo-vibe/learnings/lear-451.json',
-        score: 0.72,
-        content: JSON.stringify({
-          what_failed: 'Database connection leak in long-running loops',
-          why_failed: 'await missing in cleanup block',
-          how_to_prevent: 'Use try-finally with await client.close()'
-        })
-      }
-    ];
-    
-    res.json({ items: mockResults });
-  });
-
-  const server = app.listen(PORT);
-
-  // ─── Phase 2: CodyMaster Viking Backend ────────────────────────────────────
-  
-  console.log(chalk.dim('  [2/3] Initializing CodyMaster VikingBackend...'));
-  
-  const backend = new VikingBackend({
-    host: 'localhost',
-    port: PORT,
-    workspace: 'demo-vibe',
-    timeout: 5000
-  });
-
-  // ─── Phase 3: Perform Semantic Search ──────────────────────────────────────
-  
-  console.log(chalk.bold('  [3/3] Performing Semantic Search Query (Async Pipeline)...\n'));
-  
-  const query = 'network latency issues';
-  console.log(chalk.white(`        Query: "${query}"`));
-  
-  // Use the native ASYNC extra searchAll instead of the SYNC wrapper
-  // This bypasses the blockUntil spin-loop for a clean demo.
-  const results = await backend.searchAll(query, 5);
-
-  if (results.length > 0) {
-    console.log(chalk.green(`\n    ✅ FOUND ${results.length} RELEVANT MEMORIES VIA VIKING:\n`));
-    
-    results.forEach((r, idx) => {
-      const content = JSON.parse(r.content || '{}');
-      console.log(chalk.bold(`    ${idx + 1}. ${chalk.cyan(content.what_failed)}`));
-      console.log(chalk.dim(`       Why: ${content.why_failed}`));
-      console.log(chalk.dim(`       Fix: ${chalk.white(content.how_to_prevent)}`));
-      console.log(chalk.italic.magenta(`       Vector Similarity: ${Math.round((r.score || 0) * 100)}%\n`));
-    });
-  } else {
-    console.log(chalk.red('    ❌ No results found.'));
-  }
-
-  // ─── Cleanup ───────────────────────────────────────────────────────────────
-  
-  console.log(chalk.dim('  Demo complete. Shutting down...'));
-  server.close();
-  process.exit(0);
-}
-
-runDemo().catch(err => {
-  console.error(chalk.red('\n  🛑 Demo Failed:'), err);
-  process.exit(1);
-});

package/skills/cm-content-factory/landing/docs/content/openviking.md

@@ -1,33 +0,0 @@
-# OpenViking Semantic Memory
-
-**OpenViking** is CodyMaster's high-performance memory backend. It replaces chaotic keyword-based grep searches and blind RAG chunking with a highly structured Semantic Vector Engine.
-
-## The Problem with Standard AI Search
-
-Typical AI coding agents use raw text searches or simple AST parsing limits to find context. This causes **Code Amnesia**:
-- The agent forgets standard component library practices.
-- The agent invents (hallucinates) missing variables.
-- Scaling to million-line codebases overflows the token limit, crashing the session.
-
-## The OpenViking Solution
-
-OpenViking introduces **L0/L1/L2 Progressive Summarization** coupled with SQLite FTS5 and Vector Embeddings.
-
-### How it works:
-
-1. **L0 (Skeleton Index)**: A fast map of the entire directory structure, capturing file names, exports, and imports. 
-2. **L1 (Symbol Index)**: Extracts function signatures, class interfaces, and variable definitions without the bulky implementation details.
-3. **L2 (Full Context)**: The heavy vector embeddings (using local or API-driven embedding models) that store actual logic and edge cases.
-
-When the agent attempts to refactor a component, OpenViking performs a semantic lookup on L1 and L2 databases, pulling *only* the exact functions and files impacted by the change.
-
-## Native Integration
-
-OpenViking works autonomously in the background. When CodyMaster initializes in a project, it creates a `.viking` directory (ignored via `.gitignore`) mapped entirely to a rapid local SQLite store.
-
-```bash
-# Force an on-demand re-index of the OpenViking graph
-cody-master viking index --force
-```
-
-By ensuring agents have a complete layout of the system at all times, OpenViking single-handedly eliminates structural regressions in mature projects.