npm - create-genia-os - Versions diffs - 2.4.0 → 3.0.0 - Mend

create-genia-os 2.4.0 → 3.0.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Files changed (724) hide show

package/template/.aiox-core/utils/aiox-validator.js ADDED Viewed

@@ -0,0 +1,25 @@
+/**
+ * AIOX Validator - Re-export from canonical location
+ *
+ * @deprecated Use require('../infrastructure/scripts/aiox-validator') directly
+ * @module utils/aiox-validator
+ *
+ * This file re-exports from the canonical location in infrastructure/scripts/
+ * for backward compatibility with CI workflows.
+ *
+ * Migration note:
+ * - Canonical location: .aiox-core/infrastructure/scripts/aiox-validator.js
+ * - This file exists for backward compatibility with existing CI workflows
+ * - New code should import from infrastructure/scripts/aiox-validator directly
+ */
+'use strict';
+// Re-export from canonical location
+module.exports = require('../infrastructure/scripts/aiox-validator');
+// CLI Interface - delegate to canonical location
+if (require.main === module) {
+  // Pass through to the original script
+  require('../infrastructure/scripts/aiox-validator');
+}

package/template/.aiox-core/utils/filters/constants.js ADDED Viewed

@@ -0,0 +1,10 @@
+// =============================================================================
+// constants.js — Shared constants for filter modules
+// =============================================================================
+'use strict';
+/** Approximate characters per token (~4 chars for English text) */
+const CHARS_PER_TOKEN = 4;
+module.exports = { CHARS_PER_TOKEN };

package/template/.aiox-core/utils/filters/content-filter.js ADDED Viewed

@@ -0,0 +1,223 @@
+// =============================================================================
+// content-filter.js — Content-type filter for MCP tool responses
+// =============================================================================
+// Extracts main content from HTML/text, strips noise (nav, ads, boilerplate),
+// and truncates to max_tokens at a natural boundary.
+//
+// Usage (stdin):
+//   echo '{"content":"<html>...</html>"}' | node content-filter.js --max-tokens 2000
+//
+// Usage (file):
+//   node content-filter.js --input response.json --max-tokens 2000
+//
+// Usage (programmatic):
+//   const { filterContent } = require('./content-filter');
+//   const result = filterContent(inputData, { max_tokens: 2000, extract: ['title'] });
+// =============================================================================
+'use strict';
+const fs = require('fs');
+const { CHARS_PER_TOKEN } = require('./constants');
+/**
+ * Strip HTML tags and extract text content.
+ * Removes script, style, nav, footer, header, aside elements first.
+ * @param {string} html - Raw HTML string
+ * @returns {string} Clean text content
+ */
+function stripHtml(html) {
+  if (typeof html !== 'string') return String(html || '');
+  // Remove noise elements entirely (content + tags)
+  const noisePatterns = [
+    /<script[\s\S]*?<\/script>/gi,
+    /<style[\s\S]*?<\/style>/gi,
+    /<nav[\s\S]*?<\/nav>/gi,
+    /<footer[\s\S]*?<\/footer>/gi,
+    /<header[\s\S]*?<\/header>/gi,
+    /<aside[\s\S]*?<\/aside>/gi,
+    /<noscript[\s\S]*?<\/noscript>/gi,
+    /<!--[\s\S]*?-->/g,
+  ];
+  let clean = html;
+  for (const pattern of noisePatterns) {
+    clean = clean.replace(pattern, '');
+  }
+  // Replace block elements with newlines
+  clean = clean.replace(/<\/?(div|p|br|h[1-6]|li|tr|td|th|blockquote|pre|section|article)[^>]*>/gi, '\n');
+  // Remove remaining HTML tags
+  clean = clean.replace(/<[^>]+>/g, '');
+  // Decode common HTML entities
+  clean = clean
+    .replace(/&amp;/g, '&')
+    .replace(/&lt;/g, '<')
+    .replace(/&gt;/g, '>')
+    .replace(/&quot;/g, '"')
+    .replace(/&#39;/g, "'")
+    .replace(/&nbsp;/g, ' ');
+  // Normalize whitespace: collapse multiple blank lines
+  clean = clean.replace(/\n{3,}/g, '\n\n').trim();
+  return clean;
+}
+/**
+ * Truncate text to approximately max_tokens at a natural boundary.
+ * @param {string} text - Text to truncate
+ * @param {number} maxTokens - Approximate token limit
+ * @returns {string} Truncated text
+ */
+function truncateAtBoundary(text, maxTokens) {
+  const maxChars = maxTokens * CHARS_PER_TOKEN;
+  if (text.length <= maxChars) return text;
+  // Find last paragraph break before limit
+  const slice = text.slice(0, maxChars);
+  const lastParagraph = slice.lastIndexOf('\n\n');
+  if (lastParagraph > maxChars * 0.5) {
+    return slice.slice(0, lastParagraph).trim() + '\n\n[...truncated]';
+  }
+  // Fall back to last sentence
+  const lastSentence = slice.lastIndexOf('. ');
+  if (lastSentence > maxChars * 0.5) {
+    return slice.slice(0, lastSentence + 1).trim() + '\n\n[...truncated]';
+  }
+  // Hard truncate
+  return slice.trim() + '\n\n[...truncated]';
+}
+/**
+ * Extract specified fields from an object.
+ * @param {Object} data - Source object
+ * @param {string[]} fields - Fields to extract
+ * @returns {Object} Object with only extracted fields
+ */
+function extractFields(data, fields) {
+  if (!data || typeof data !== 'object' || !Array.isArray(fields)) return data;
+  const result = {};
+  for (const field of fields) {
+    if (field in data) {
+      result[field] = data[field];
+    }
+  }
+  return result;
+}
+/**
+ * Apply content filter to input data.
+ * @param {string|Object} input - Raw input (string or parsed JSON)
+ * @param {Object} options - Filter options
+ * @param {number} [options.max_tokens=3000] - Token limit
+ * @param {string[]} [options.extract] - Fields to extract before content filtering
+ * @returns {{ filtered: string, original_length: number, filtered_length: number, reduction_pct: number }}
+ */
+function filterContent(input, options = {}) {
+  const maxTokens = options.max_tokens || 3000;
+  const extractList = options.extract || null;
+  let raw = input;
+  // If input is an object, extract fields first or stringify
+  if (typeof raw === 'object' && raw !== null) {
+    if (extractList) {
+      if (Array.isArray(raw)) {
+        raw = raw.map((item) => extractFields(item, extractList));
+        raw = JSON.stringify(raw, null, 2);
+      } else {
+        raw = extractFields(raw, extractList);
+        raw = JSON.stringify(raw, null, 2);
+      }
+    } else {
+      raw = raw.content || raw.text || raw.body || JSON.stringify(raw, null, 2);
+    }
+  }
+  if (typeof raw !== 'string') raw = String(raw || '');
+  const originalLength = raw.length;
+  // Strip HTML if it looks like HTML
+  let content = raw;
+  if (content.includes('<') && content.includes('>')) {
+    content = stripHtml(content);
+  }
+  // Truncate to token limit
+  content = truncateAtBoundary(content, maxTokens);
+  const filteredLength = content.length;
+  const reductionPct = originalLength > 0
+    ? Math.round((1 - filteredLength / originalLength) * 100)
+    : 0;
+  return {
+    filtered: content,
+    original_length: originalLength,
+    filtered_length: filteredLength,
+    reduction_pct: reductionPct,
+  };
+}
+// ---------------------------------------------------------------------------
+// CLI entrypoint
+// ---------------------------------------------------------------------------
+function parseArgs(argv) {
+  const args = {};
+  for (let i = 2; i < argv.length; i++) {
+    if (argv[i] === '--max-tokens' && argv[i + 1]) {
+      args.max_tokens = parseInt(argv[i + 1], 10);
+      i++;
+    } else if (argv[i] === '--input' && argv[i + 1]) {
+      args.input = argv[i + 1];
+      i++;
+    } else if (argv[i] === '--extract' && argv[i + 1]) {
+      args.extract = argv[i + 1].split(',');
+      i++;
+    }
+  }
+  return args;
+}
+function main() {
+  const args = parseArgs(process.argv);
+  let inputData;
+  if (args.input) {
+    inputData = fs.readFileSync(args.input, 'utf8');
+  } else if (!process.stdin.isTTY) {
+    inputData = fs.readFileSync(0, 'utf8');
+  } else {
+    console.error('Usage: echo \'data\' | node content-filter.js --max-tokens 2000');
+    process.exit(1);
+  }
+  try {
+    inputData = JSON.parse(inputData);
+  } catch (_) {
+    // treat as raw text
+  }
+  const result = filterContent(inputData, {
+    max_tokens: args.max_tokens,
+    extract: args.extract,
+  });
+  console.log(JSON.stringify(result, null, 2));
+}
+// Run CLI if executed directly
+if (require.main === module) {
+  main();
+}
+module.exports = { filterContent, stripHtml, truncateAtBoundary, extractFields, CHARS_PER_TOKEN };

package/template/.aiox-core/utils/filters/field-filter.js ADDED Viewed

@@ -0,0 +1,126 @@
+// =============================================================================
+// field-filter.js — Field-type filter for MCP tool responses (array data)
+// =============================================================================
+// Projects specified columns from arrays of objects and limits row count.
+// Analogous to SQL: SELECT field1, field2 FROM data LIMIT max_rows
+//
+// Usage (stdin):
+//   echo '[{"a":1,"b":2,"c":3}]' | node field-filter.js --fields a,b --max-rows 10
+//
+// Usage (programmatic):
+//   const { filterFields } = require('./field-filter');
+//   const result = filterFields(data, { fields: ['a','b'], max_rows: 10 });
+// =============================================================================
+'use strict';
+const fs = require('fs');
+/**
+ * Apply field filter to array data.
+ * @param {Object[]} input - Array of objects
+ * @param {Object} options - Filter options
+ * @param {string[]} options.fields - Columns to project
+ * @param {number} [options.max_rows] - Maximum rows to include
+ * @returns {{ filtered: Object[], original_length: number, filtered_length: number, reduction_pct: number, rows_original: number, rows_filtered: number }}
+ */
+function filterFields(input, options = {}) {
+  const fields = options.fields || [];
+  const maxRows = options.max_rows || Infinity;
+  // Ensure input is an array
+  const data = Array.isArray(input) ? input : [input];
+  const originalSerialized = JSON.stringify(data, null, 2);
+  const originalLength = originalSerialized.length;
+  const rowsOriginal = data.length;
+  // Limit rows
+  const sliced = data.slice(0, maxRows);
+  // Project fields
+  let filtered;
+  if (fields.length > 0) {
+    filtered = sliced.map((item) => {
+      if (!item || typeof item !== 'object') return item;
+      const projected = {};
+      for (const field of fields) {
+        if (field in item) {
+          projected[field] = item[field];
+        }
+      }
+      return projected;
+    });
+  } else {
+    filtered = sliced;
+  }
+  const filteredSerialized = JSON.stringify(filtered, null, 2);
+  const filteredLength = filteredSerialized.length;
+  const reductionPct = originalLength > 0
+    ? Math.round((1 - filteredLength / originalLength) * 100)
+    : 0;
+  return {
+    filtered,
+    original_length: originalLength,
+    filtered_length: filteredLength,
+    reduction_pct: reductionPct,
+    rows_original: rowsOriginal,
+    rows_filtered: filtered.length,
+  };
+}
+// ---------------------------------------------------------------------------
+// CLI entrypoint
+// ---------------------------------------------------------------------------
+function parseArgs(argv) {
+  const args = {};
+  for (let i = 2; i < argv.length; i++) {
+    if (argv[i] === '--fields' && argv[i + 1]) {
+      args.fields = argv[i + 1].split(',');
+      i++;
+    } else if (argv[i] === '--max-rows' && argv[i + 1]) {
+      args.max_rows = parseInt(argv[i + 1], 10);
+      i++;
+    } else if (argv[i] === '--input' && argv[i + 1]) {
+      args.input = argv[i + 1];
+      i++;
+    }
+  }
+  return args;
+}
+function main() {
+  const args = parseArgs(process.argv);
+  let inputData;
+  if (args.input) {
+    inputData = fs.readFileSync(args.input, 'utf8');
+  } else if (!process.stdin.isTTY) {
+    inputData = fs.readFileSync(0, 'utf8');
+  } else {
+    console.error('Usage: echo \'[{"a":1}]\' | node field-filter.js --fields a --max-rows 10');
+    process.exit(1);
+  }
+  try {
+    inputData = JSON.parse(inputData);
+  } catch (err) {
+    console.error('Error: Input must be valid JSON.', err.message);
+    process.exit(1);
+  }
+  const result = filterFields(inputData, {
+    fields: args.fields,
+    max_rows: args.max_rows,
+  });
+  console.log(JSON.stringify(result, null, 2));
+}
+if (require.main === module) {
+  main();
+}
+module.exports = { filterFields };

package/template/.aiox-core/utils/filters/index.js ADDED Viewed

@@ -0,0 +1,180 @@
+// =============================================================================
+// index.js — Filter engine entry point
+// =============================================================================
+// Reads tool-registry.yaml, looks up filter config for a given tool,
+// and dispatches to the appropriate filter (content, schema, field).
+//
+// Usage:
+//   node .aiox-core/utils/filters/index.js --tool exa --input response.json
+//   echo '{"data":"..."}' | node .aiox-core/utils/filters/index.js --tool exa
+//
+// Programmatic:
+//   const { applyFilter, loadFilterConfig } = require('./index');
+//   const result = applyFilter('exa', inputData);
+// =============================================================================
+'use strict';
+const fs = require('fs');
+const path = require('path');
+const { filterContent } = require('./content-filter');
+const { filterSchema } = require('./schema-filter');
+const { filterFields } = require('./field-filter');
+// Registry path relative to project root
+const REGISTRY_PATH = path.join(__dirname, '..', '..', 'data', 'tool-registry.yaml');
+/**
+ * Load filter configuration for a tool from tool-registry.yaml.
+ * @param {string} toolName - Tool name (case-sensitive, matches registry key)
+ * @param {string} [registryPath] - Override registry path (for testing)
+ * @returns {Object|null} Filter config or null if not found
+ */
+function loadFilterConfig(toolName, registryPath) {
+  const regPath = registryPath || REGISTRY_PATH;
+  let yaml;
+  try {
+    yaml = require('js-yaml');
+  } catch (_) {
+    // js-yaml not available — try yaml package
+    try {
+      yaml = require('yaml');
+      const content = fs.readFileSync(regPath, 'utf8');
+      const registry = yaml.parse(content);
+      const tool = registry.tools && registry.tools[toolName];
+      return (tool && tool.filter) || null;
+    } catch (__) {
+      return null;
+    }
+  }
+  try {
+    const content = fs.readFileSync(regPath, 'utf8');
+    const registry = yaml.load(content);
+    const tool = registry.tools && registry.tools[toolName];
+    return (tool && tool.filter) || null;
+  } catch (_) {
+    return null;
+  }
+}
+/**
+ * Apply the appropriate filter to input data based on tool's registry config.
+ * @param {string} toolName - Tool name
+ * @param {*} input - Input data (string, object, or array)
+ * @param {Object} [overrideConfig] - Override filter config (skip registry lookup)
+ * @param {string} [registryPath] - Override registry path
+ * @returns {{ filtered: *, original_length: number, filtered_length: number, reduction_pct: number, filter_type: string }}
+ */
+function applyFilter(toolName, input, overrideConfig, registryPath) {
+  const config = overrideConfig || loadFilterConfig(toolName, registryPath);
+  if (!config) {
+    // No filter config — pass through
+    const serialized = typeof input === 'string' ? input : JSON.stringify(input, null, 2);
+    return {
+      filtered: input,
+      original_length: serialized.length,
+      filtered_length: serialized.length,
+      reduction_pct: 0,
+      filter_type: 'none',
+    };
+  }
+  const filterType = config.type;
+  switch (filterType) {
+    case 'content':
+      return {
+        ...filterContent(input, {
+          max_tokens: config.max_tokens,
+          extract: config.extract,
+        }),
+        filter_type: 'content',
+      };
+    case 'schema':
+      return {
+        ...filterSchema(input, {
+          fields: config.fields,
+          max_tokens: config.max_tokens,
+        }),
+        filter_type: 'schema',
+      };
+    case 'field':
+      return {
+        ...filterFields(input, {
+          fields: config.fields,
+          max_rows: config.max_rows,
+        }),
+        filter_type: 'field',
+      };
+    default:
+      // Unknown filter type — pass through
+      const serialized = typeof input === 'string' ? input : JSON.stringify(input, null, 2);
+      return {
+        filtered: input,
+        original_length: serialized.length,
+        filtered_length: serialized.length,
+        reduction_pct: 0,
+        filter_type: 'unknown',
+      };
+  }
+}
+// ---------------------------------------------------------------------------
+// CLI entrypoint
+// ---------------------------------------------------------------------------
+function parseArgs(argv) {
+  const args = {};
+  for (let i = 2; i < argv.length; i++) {
+    if (argv[i] === '--tool' && argv[i + 1]) {
+      args.tool = argv[i + 1];
+      i++;
+    } else if (argv[i] === '--input' && argv[i + 1]) {
+      args.input = argv[i + 1];
+      i++;
+    } else if (argv[i] === '--registry' && argv[i + 1]) {
+      args.registry = argv[i + 1];
+      i++;
+    }
+  }
+  return args;
+}
+function main() {
+  const args = parseArgs(process.argv);
+  if (!args.tool) {
+    console.error('Usage: node index.js --tool <tool-name> [--input <file>] [--registry <path>]');
+    process.exit(1);
+  }
+  let inputData;
+  if (args.input) {
+    inputData = fs.readFileSync(args.input, 'utf8');
+  } else if (!process.stdin.isTTY) {
+    inputData = fs.readFileSync(0, 'utf8');
+  } else {
+    console.error('Error: Provide input via --input file or stdin pipe.');
+    process.exit(1);
+  }
+  try {
+    inputData = JSON.parse(inputData);
+  } catch (_) {
+    // treat as raw text
+  }
+  const result = applyFilter(args.tool, inputData, null, args.registry);
+  console.log(JSON.stringify(result, null, 2));
+}
+if (require.main === module) {
+  main();
+}
+module.exports = { applyFilter, loadFilterConfig };