snow-flow 8.42.12 → 9.0.1

package/dist/mcp/servicenow-mcp-unified/tools/machine-learning-MIGRATED/ml_train_change_risk.js

@@ -1,262 +0,0 @@
-"use strict";
-/**
- * ml_train_change_risk - Train neural networks LOCALLY for change risk prediction
- *
- * ⚠️ IMPORTANT: Trains models LOCALLY on your machine, NOT in ServiceNow.
- * Alternative to ServiceNow Predictive Intelligence (PI) for dev/testing.
- *
- * Trains a model to predict change risk based on historical change data including:
- * - Change description and category
- * - Approval count and test/backout plans
- * - Historical success/failure rates
- * - Assignment group patterns
- *
- * How it works:
- * 1. Fetches change data from ServiceNow via OAuth2 API
- * 2. Trains TensorFlow.js model locally (Node.js)
- * 3. Saves to .snow-flow/ml-models/
- * 4. NOT importable to ServiceNow PI
- */
-var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
-    if (k2 === undefined) k2 = k;
-    var desc = Object.getOwnPropertyDescriptor(m, k);
-    if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
-      desc = { enumerable: true, get: function() { return m[k]; } };
-    }
-    Object.defineProperty(o, k2, desc);
-}) : (function(o, m, k, k2) {
-    if (k2 === undefined) k2 = k;
-    o[k2] = m[k];
-}));
-var __setModuleDefault = (this && this.__setModuleDefault) || (Object.create ? (function(o, v) {
-    Object.defineProperty(o, "default", { enumerable: true, value: v });
-}) : function(o, v) {
-    o["default"] = v;
-});
-var __importStar = (this && this.__importStar) || (function () {
-    var ownKeys = function(o) {
-        ownKeys = Object.getOwnPropertyNames || function (o) {
-            var ar = [];
-            for (var k in o) if (Object.prototype.hasOwnProperty.call(o, k)) ar[ar.length] = k;
-            return ar;
-        };
-        return ownKeys(o);
-    };
-    return function (mod) {
-        if (mod && mod.__esModule) return mod;
-        var result = {};
-        if (mod != null) for (var k = ownKeys(mod), i = 0; i < k.length; i++) if (k[i] !== "default") __createBinding(result, mod, k[i]);
-        __setModuleDefault(result, mod);
-        return result;
-    };
-})();
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.version = exports.toolDefinition = void 0;
-exports.execute = execute;
-const auth_js_1 = require("../../shared/auth.js");
-const error_handler_js_1 = require("../../shared/error-handler.js");
-const tf = __importStar(require("@tensorflow/tfjs"));
-exports.toolDefinition = {
-    name: 'ml_train_change_risk',
-    description: '⚠️ LOCAL ML TRAINING: Trains neural networks on your machine using ServiceNow change data fetched via API. NOT in ServiceNow. Alternative to PI license for dev/testing.',
-    // Metadata for tool discovery (not sent to LLM)
-    category: 'ml-analytics',
-    subcategory: 'machine-learning',
-    use_cases: ['training', 'risk-prediction', 'local-ml'],
-    complexity: 'advanced',
-    frequency: 'low',
-    inputSchema: {
-        type: 'object',
-        properties: {
-            sample_size: {
-                type: 'number',
-                description: 'Number of change requests to use for training',
-                default: 500
-            },
-            include_failed_changes: {
-                type: 'boolean',
-                description: 'Include failed changes in training data',
-                default: true
-            },
-            epochs: {
-                type: 'number',
-                description: 'Training epochs',
-                default: 50
-            },
-            validation_split: {
-                type: 'number',
-                description: 'Validation data percentage',
-                default: 0.2
-            }
-        }
-    }
-};
-async function execute(args, context) {
-    const { sample_size = 500, include_failed_changes = true, epochs = 50, validation_split = 0.2 } = args;
-    try {
-        const client = await (0, auth_js_1.getAuthenticatedClient)(context);
-        // Initialize TensorFlow.js
-        await tf.ready();
-        // Fetch change request data
-        const query = include_failed_changes ?
-            'state!=cancelled' :
-            'state=closed^close_code=successful';
-        const changes = await client.query('change_request', {
-            query: query,
-            limit: sample_size,
-            fields: ['short_description', 'risk', 'category', 'type', 'assignment_group', 'approval', 'test_plan', 'backout_plan', 'close_code', 'implementation_plan']
-        });
-        // Check data availability
-        const requiredRecords = 100;
-        const availableRecords = changes.length;
-        const canTrain = availableRecords >= requiredRecords;
-        if (!changes || changes.length === 0) {
-            return (0, error_handler_js_1.createErrorResult)('No change requests found for training', {
-                data_availability: {
-                    required_records: requiredRecords,
-                    available_records: 0,
-                    can_train: false,
-                    recommendation: 'No change request data available. Ensure your instance has change request records.'
-                }
-            });
-        }
-        if (!canTrain) {
-            return (0, error_handler_js_1.createErrorResult)(`Insufficient training data: found ${availableRecords} records, need at least ${requiredRecords}`, {
-                data_availability: {
-                    required_records: requiredRecords,
-                    available_records: availableRecords,
-                    can_train: false,
-                    recommendation: `Need at least ${requiredRecords} change records for reliable training. Currently have ${availableRecords}. Consider adjusting your query or waiting for more data.`
-                }
-            });
-        }
-        // Prepare training data
-        const { features, labels, riskLevels } = prepareChangeData(changes);
-        // Create neural network model
-        const model = tf.sequential({
-            layers: [
-                tf.layers.dense({
-                    units: 64,
-                    activation: 'relu',
-                    inputShape: [features.shape[1]]
-                }),
-                tf.layers.dropout({ rate: 0.3 }),
-                tf.layers.dense({
-                    units: 32,
-                    activation: 'relu'
-                }),
-                tf.layers.dropout({ rate: 0.3 }),
-                tf.layers.dense({
-                    units: riskLevels.length,
-                    activation: 'softmax'
-                })
-            ]
-        });
-        // Compile model
-        model.compile({
-            optimizer: tf.train.adam(0.001),
-            loss: 'categoricalCrossentropy',
-            metrics: ['accuracy']
-        });
-        // Train model
-        const history = await model.fit(features, labels, {
-            epochs,
-            validationSplit: validation_split,
-            batchSize: 32,
-            callbacks: {
-                onEpochEnd: (epoch, logs) => {
-                    const loss = logs?.loss ? logs.loss.toFixed(4) : 'N/A';
-                    const accuracy = logs?.acc ? (logs.acc * 100).toFixed(2) : 'N/A';
-                    console.log(`Epoch ${epoch + 1}/${epochs} - Loss: ${loss}, Accuracy: ${accuracy}%`);
-                }
-            }
-        });
-        // Clean up tensors
-        features.dispose();
-        labels.dispose();
-        // Calculate final metrics
-        const finalAccuracyValue = history.history.acc[history.history.acc.length - 1];
-        const finalAccuracy = typeof finalAccuracyValue === 'number' ? finalAccuracyValue : Array.isArray(finalAccuracyValue) ? finalAccuracyValue[0] : 0;
-        const finalLossValue = history.history.loss[history.history.loss.length - 1];
-        const finalLoss = typeof finalLossValue === 'number' ? finalLossValue : Array.isArray(finalLossValue) ? finalLossValue[0] : 0;
-        return (0, error_handler_js_1.createSuccessResult)({
-            status: 'success',
-            message: 'Change risk predictor trained successfully',
-            training_summary: {
-                samples: changes.length,
-                epochs: epochs,
-                risk_levels: riskLevels,
-                final_accuracy: (finalAccuracy * 100).toFixed(2) + '%',
-                final_loss: finalLoss.toFixed(4)
-            },
-            data_availability: {
-                required_records: requiredRecords,
-                available_records: availableRecords,
-                can_train: true,
-                recommendation: `Successfully trained on ${availableRecords} records. Model is ready for risk predictions.`
-            }
-        });
-    }
-    catch (error) {
-        return (0, error_handler_js_1.createErrorResult)(error.message);
-    }
-}
-/**
- * Prepare change request data for neural network training
- */
-function prepareChangeData(changes) {
-    const riskLevels = ['low', 'moderate', 'high'];
-    const features = [];
-    const labels = [];
-    for (const change of changes) {
-        // Extract features
-        const feature = [
-            // Risk encoding
-            riskLevels.indexOf(change.risk || 'moderate'),
-            // Category encoding (simple hash)
-            simpleHash(change.category || ''),
-            // Type encoding
-            simpleHash(change.type || 'standard'),
-            // Has test plan
-            change.test_plan === 'true' ? 1 : 0,
-            // Has backout plan
-            change.backout_plan === 'true' ? 1 : 0,
-            // Approval count
-            parseInt(change.approval) || 0,
-            // Description length (complexity indicator)
-            (change.short_description || '').length / 100,
-            // Implementation plan exists
-            change.implementation_plan ? 1 : 0
-        ];
-        features.push(feature);
-        // One-hot encode risk level
-        const risk = change.risk || 'moderate';
-        const riskIndex = riskLevels.indexOf(risk);
-        const label = new Array(riskLevels.length).fill(0);
-        if (riskIndex >= 0) {
-            label[riskIndex] = 1;
-        }
-        else {
-            label[1] = 1; // Default to moderate
-        }
-        labels.push(label);
-    }
-    return {
-        features: tf.tensor2d(features),
-        labels: tf.tensor2d(labels),
-        riskLevels
-    };
-}
-/**
- * Simple hash function for string encoding
- */
-function simpleHash(str) {
-    let hash = 0;
-    for (let i = 0; i < str.length; i++) {
-        hash = ((hash << 5) - hash) + str.charCodeAt(i);
-        hash = hash & hash;
-    }
-    return Math.abs(hash) % 100;
-}
-exports.version = '1.0.0';
-//# sourceMappingURL=ml_train_change_risk.js.map

package/dist/mcp/servicenow-mcp-unified/tools/machine-learning-MIGRATED/ml_train_change_risk.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"ml_train_change_risk.js","sourceRoot":"","sources":["../../../../../src/mcp/servicenow-mcp-unified/tools/machine-learning-MIGRATED/ml_train_change_risk.ts"],"names":[],"mappings":";AAAA;;;;;;;;;;;;;;;;;GAiBG;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AA2CH,0BAgIC;AAxKD,kDAA8D;AAC9D,oEAAuF;AACvF,qDAAuC;AAE1B,QAAA,cAAc,GAAsB;IAC/C,IAAI,EAAE,sBAAsB;IAC5B,WAAW,EAAE,0KAA0K;IACvL,gDAAgD;IAChD,QAAQ,EAAE,cAAc;IACxB,WAAW,EAAE,kBAAkB;IAC/B,SAAS,EAAE,CAAC,UAAU,EAAE,iBAAiB,EAAE,UAAU,CAAC;IACtD,UAAU,EAAE,UAAU;IACtB,SAAS,EAAE,KAAK;IAChB,WAAW,EAAE;QACX,IAAI,EAAE,QAAQ;QACd,UAAU,EAAE;YACV,WAAW,EAAE;gBACX,IAAI,EAAE,QAAQ;gBACd,WAAW,EAAE,+CAA+C;gBAC5D,OAAO,EAAE,GAAG;aACb;YACD,sBAAsB,EAAE;gBACtB,IAAI,EAAE,SAAS;gBACf,WAAW,EAAE,yCAAyC;gBACtD,OAAO,EAAE,IAAI;aACd;YACD,MAAM,EAAE;gBACN,IAAI,EAAE,QAAQ;gBACd,WAAW,EAAE,iBAAiB;gBAC9B,OAAO,EAAE,EAAE;aACZ;YACD,gBAAgB,EAAE;gBAChB,IAAI,EAAE,QAAQ;gBACd,WAAW,EAAE,4BAA4B;gBACzC,OAAO,EAAE,GAAG;aACb;SACF;KACF;CACF,CAAC;AAEK,KAAK,UAAU,OAAO,CAAC,IAAS,EAAE,OAA0B;IACjE,MAAM,EACJ,WAAW,GAAG,GAAG,EACjB,sBAAsB,GAAG,IAAI,EAC7B,MAAM,GAAG,EAAE,EACX,gBAAgB,GAAG,GAAG,EACvB,GAAG,IAAI,CAAC;IAET,IAAI,CAAC;QACH,MAAM,MAAM,GAAG,MAAM,IAAA,gCAAsB,EAAC,OAAO,CAAC,CAAC;QAErD,2BAA2B;QAC3B,MAAM,EAAE,CAAC,KAAK,EAAE,CAAC;QAEjB,4BAA4B;QAC5B,MAAM,KAAK,GAAG,sBAAsB,CAAC,CAAC;YACpC,kBAAkB,CAAC,CAAC;YACpB,oCAAoC,CAAC;QAEvC,MAAM,OAAO,GAAG,MAAM,MAAM,CAAC,KAAK,CAAC,gBAAgB,EAAE;YACnD,KAAK,EAAE,KAAK;YACZ,KAAK,EAAE,WAAW;YAClB,MAAM,EAAE,CAAC,mBAAmB,EAAE,MAAM,EAAE,UAAU,EAAE,MAAM,EAAE,kBAAkB,EAAE,UAAU,EAAE,WAAW,EAAE,cAAc,EAAE,YAAY,EAAE,qBAAqB,CAAC;SAC5J,CAAC,CAAC;QAEH,0BAA0B;QAC1B,MAAM,eAAe,GAAG,GAAG,CAAC;QAC5B,MAAM,gBAAgB,GAAG,OAAO,CAAC,MAAM,CAAC;QACxC,MAAM,QAAQ,GAAG,gBAAgB,IAAI,eAAe,CAAC;QAErD,IAAI,CAAC,OAAO,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACrC,OAAO,IAAA,oCAAiB,EAAC,uCAAuC,EAAE;gBAChE,iBAAiB,EAAE;oBACjB,gBAAgB,EAAE,eAAe;oBACjC,iBAAiB,EAAE,CAAC;oBACpB,SAAS,EAAE,KAAK;oBAChB,cAAc,EAAE,oFAAoF;iBACrG;aACF,CAAC,CAAC;QACL,CAAC;QAED,IAAI,CAAC,QAAQ,EAAE,CAAC;YACd,OAAO,IAAA,oCAAiB,EAAC,qCAAqC,gBAAgB,2BAA2B,eAAe,EAAE,EAAE;gBAC1H,iBAAiB,EAAE;oBACjB,gBAAgB,EAAE,eAAe;oBACjC,iBAAiB,EAAE,gBAAgB;oBACnC,SAAS,EAAE,KAAK;oBAChB,cAAc,EAAE,iBAAiB,eAAe,yDAAyD,gBAAgB,2DAA2D;iBACrL;aACF,CAAC,CAAC;QACL,CAAC;QAED,wBAAwB;QACxB,MAAM,EAAE,QAAQ,EAAE,MAAM,EAAE,UAAU,EAAE,GAAG,iBAAiB,CAAC,OAAO,CAAC,CAAC;QAEpE,8BAA8B;QAC9B,MAAM,KAAK,GAAG,EAAE,CAAC,UAAU,CAAC;YAC1B,MAAM,EAAE;gBACN,EAAE,CAAC,MAAM,CAAC,KAAK,CAAC;oBACd,KAAK,EAAE,EAAE;oBACT,UAAU,EAAE,MAAM;oBAClB,UAAU,EAAE,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;iBAChC,CAAC;gBACF,EAAE,CAAC,MAAM,CAAC,OAAO,CAAC,EAAE,IAAI,EAAE,GAAG,EAAE,CAAC;gBAChC,EAAE,CAAC,MAAM,CAAC,KAAK,CAAC;oBACd,KAAK,EAAE,EAAE;oBACT,UAAU,EAAE,MAAM;iBACnB,CAAC;gBACF,EAAE,CAAC,MAAM,CAAC,OAAO,CAAC,EAAE,IAAI,EAAE,GAAG,EAAE,CAAC;gBAChC,EAAE,CAAC,MAAM,CAAC,KAAK,CAAC;oBACd,KAAK,EAAE,UAAU,CAAC,MAAM;oBACxB,UAAU,EAAE,SAAS;iBACtB,CAAC;aACH;SACF,CAAC,CAAC;QAEH,gBAAgB;QAChB,KAAK,CAAC,OAAO,CAAC;YACZ,SAAS,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC;YAC/B,IAAI,EAAE,yBAAyB;YAC/B,OAAO,EAAE,CAAC,UAAU,CAAC;SACtB,CAAC,CAAC;QAEH,cAAc;QACd,MAAM,OAAO,GAAG,MAAM,KAAK,CAAC,GAAG,CAAC,QAAQ,EAAE,MAAM,EAAE;YAChD,MAAM;YACN,eAAe,EAAE,gBAAgB;YACjC,SAAS,EAAE,EAAE;YACb,SAAS,EAAE;gBACT,UAAU,EAAE,CAAC,KAAa,EAAE,IAAU,EAAE,EAAE;oBACxC,MAAM,IAAI,GAAG,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC;oBACvD,MAAM,QAAQ,GAAG,IAAI,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,GAAG,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC;oBACjE,OAAO,CAAC,GAAG,CAAC,SAAS,KAAK,GAAG,CAAC,IAAI,MAAM,YAAY,IAAI,eAAe,QAAQ,GAAG,CAAC,CAAC;gBACtF,CAAC;aACF;SACF,CAAC,CAAC;QAEH,mBAAmB;QACnB,QAAQ,CAAC,OAAO,EAAE,CAAC;QACnB,MAAM,CAAC,OAAO,EAAE,CAAC;QAEjB,0BAA0B;QAC1B,MAAM,kBAAkB,GAAG,OAAO,CAAC,OAAO,CAAC,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC,GAAG,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;QAC/E,MAAM,aAAa,GAAG,OAAO,kBAAkB,KAAK,QAAQ,CAAC,CAAC,CAAC,kBAAkB,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,kBAAkB,CAAC,CAAC,CAAC,CAAC,kBAAkB,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QAClJ,MAAM,cAAc,GAAG,OAAO,CAAC,OAAO,CAAC,IAAI,CAAC,OAAO,CAAC,OAAO,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;QAC7E,MAAM,SAAS,GAAG,OAAO,cAAc,KAAK,QAAQ,CAAC,CAAC,CAAC,cAAc,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,cAAc,CAAC,CAAC,CAAC,CAAC,cAAc,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QAE9H,OAAO,IAAA,sCAAmB,EAAC;YACzB,MAAM,EAAE,SAAS;YACjB,OAAO,EAAE,4CAA4C;YACrD,gBAAgB,EAAE;gBAChB,OAAO,EAAE,OAAO,CAAC,MAAM;gBACvB,MAAM,EAAE,MAAM;gBACd,WAAW,EAAE,UAAU;gBACvB,cAAc,EAAE,CAAC,aAAa,GAAG,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,GAAG;gBACtD,UAAU,EAAE,SAAS,CAAC,OAAO,CAAC,CAAC,CAAC;aACjC;YACD,iBAAiB,EAAE;gBACjB,gBAAgB,EAAE,eAAe;gBACjC,iBAAiB,EAAE,gBAAgB;gBACnC,SAAS,EAAE,IAAI;gBACf,cAAc,EAAE,2BAA2B,gBAAgB,gDAAgD;aAC5G;SACF,CAAC,CAAC;IAEL,CAAC;IAAC,OAAO,KAAU,EAAE,CAAC;QACpB,OAAO,IAAA,oCAAiB,EAAC,KAAK,CAAC,OAAO,CAAC,CAAC;IAC1C,CAAC;AACH,CAAC;AAED;;GAEG;AACH,SAAS,iBAAiB,CAAC,OAAc;IACvC,MAAM,UAAU,GAAG,CAAC,KAAK,EAAE,UAAU,EAAE,MAAM,CAAC,CAAC;IAE/C,MAAM,QAAQ,GAAe,EAAE,CAAC;IAChC,MAAM,MAAM,GAAe,EAAE,CAAC;IAE9B,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC7B,mBAAmB;QACnB,MAAM,OAAO,GAAG;YACd,gBAAgB;YAChB,UAAU,CAAC,OAAO,CAAC,MAAM,CAAC,IAAI,IAAI,UAAU,CAAC;YAC7C,kCAAkC;YAClC,UAAU,CAAC,MAAM,CAAC,QAAQ,IAAI,EAAE,CAAC;YACjC,gBAAgB;YAChB,UAAU,CAAC,MAAM,CAAC,IAAI,IAAI,UAAU,CAAC;YACrC,gBAAgB;YAChB,MAAM,CAAC,SAAS,KAAK,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;YACnC,mBAAmB;YACnB,MAAM,CAAC,YAAY,KAAK,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;YACtC,iBAAiB;YACjB,QAAQ,CAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC;YAC9B,4CAA4C;YAC5C,CAAC,MAAM,CAAC,iBAAiB,IAAI,EAAE,CAAC,CAAC,MAAM,GAAG,GAAG;YAC7C,6BAA6B;YAC7B,MAAM,CAAC,mBAAmB,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;SACnC,CAAC;QAEF,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;QAEvB,4BAA4B;QAC5B,MAAM,IAAI,GAAG,MAAM,CAAC,IAAI,IAAI,UAAU,CAAC;QACvC,MAAM,SAAS,GAAG,UAAU,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;QAC3C,MAAM,KAAK,GAAG,IAAI,KAAK,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACnD,IAAI,SAAS,IAAI,CAAC,EAAE,CAAC;YACnB,KAAK,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC;QACvB,CAAC;aAAM,CAAC;YACN,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,sBAAsB;QACtC,CAAC;QACD,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IACrB,CAAC;IAED,OAAO;QACL,QAAQ,EAAE,EAAE,CAAC,QAAQ,CAAC,QAAQ,CAAC;QAC/B,MAAM,EAAE,EAAE,CAAC,QAAQ,CAAC,MAAM,CAAC;QAC3B,UAAU;KACX,CAAC;AACJ,CAAC;AAED;;GAEG;AACH,SAAS,UAAU,CAAC,GAAW;IAC7B,IAAI,IAAI,GAAG,CAAC,CAAC;IACb,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACpC,IAAI,GAAG,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,GAAG,GAAG,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC;QAChD,IAAI,GAAG,IAAI,GAAG,IAAI,CAAC;IACrB,CAAC;IACD,OAAO,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,GAAG,CAAC;AAC9B,CAAC;AAEY,QAAA,OAAO,GAAG,OAAO,CAAC"}

package/dist/mcp/servicenow-mcp-unified/tools/machine-learning-MIGRATED/ml_train_incident_classifier.d.ts

@@ -1,23 +0,0 @@
-/**
- * ml_train_incident_classifier - Train LSTM neural networks LOCALLY on ServiceNow incident data
- *
- * ⚠️ IMPORTANT: This trains models LOCALLY on your machine, NOT in ServiceNow.
- * This is an alternative to ServiceNow Predictive Intelligence (PI) for dev/testing.
- *
- * This tool trains a deep learning classifier using TensorFlow.js with:
- * - LSTM layers for sequence processing
- * - Embedding layers for text representation
- * - Dropout for regularization
- * - Intelligent data selection and optimization
- *
- * How it works:
- * 1. Fetches incident data from ServiceNow via OAuth2 API
- * 2. Trains TensorFlow.js model locally (Node.js environment)
- * 3. Saves model to .snow-flow/ml-models/ directory
- * 4. NOT importable into ServiceNow PI
- */
-import { MCPToolDefinition, ServiceNowContext, ToolResult } from '../../shared/types.js';
-export declare const toolDefinition: MCPToolDefinition;
-export declare function execute(args: any, context: ServiceNowContext): Promise<ToolResult>;
-export declare const version = "1.0.0";
-//# sourceMappingURL=ml_train_incident_classifier.d.ts.map

package/dist/mcp/servicenow-mcp-unified/tools/machine-learning-MIGRATED/ml_train_incident_classifier.d.ts.map

@@ -1 +0,0 @@
-{"version":3,"file":"ml_train_incident_classifier.d.ts","sourceRoot":"","sources":["../../../../../src/mcp/servicenow-mcp-unified/tools/machine-learning-MIGRATED/ml_train_incident_classifier.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;GAiBG;AAEH,OAAO,EAAE,iBAAiB,EAAE,iBAAiB,EAAE,UAAU,EAAE,MAAM,uBAAuB,CAAC;AAMzF,eAAO,MAAM,cAAc,EAAE,iBA8D5B,CAAC;AAEF,wBAAsB,OAAO,CAAC,IAAI,EAAE,GAAG,EAAE,OAAO,EAAE,iBAAiB,GAAG,OAAO,CAAC,UAAU,CAAC,CAwNxF;AAyED,eAAO,MAAM,OAAO,UAAU,CAAC"}

package/dist/mcp/servicenow-mcp-unified/tools/machine-learning-MIGRATED/ml_train_incident_classifier.js

@@ -1,355 +0,0 @@
-"use strict";
-/**
- * ml_train_incident_classifier - Train LSTM neural networks LOCALLY on ServiceNow incident data
- *
- * ⚠️ IMPORTANT: This trains models LOCALLY on your machine, NOT in ServiceNow.
- * This is an alternative to ServiceNow Predictive Intelligence (PI) for dev/testing.
- *
- * This tool trains a deep learning classifier using TensorFlow.js with:
- * - LSTM layers for sequence processing
- * - Embedding layers for text representation
- * - Dropout for regularization
- * - Intelligent data selection and optimization
- *
- * How it works:
- * 1. Fetches incident data from ServiceNow via OAuth2 API
- * 2. Trains TensorFlow.js model locally (Node.js environment)
- * 3. Saves model to .snow-flow/ml-models/ directory
- * 4. NOT importable into ServiceNow PI
- */
-var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
-    if (k2 === undefined) k2 = k;
-    var desc = Object.getOwnPropertyDescriptor(m, k);
-    if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
-      desc = { enumerable: true, get: function() { return m[k]; } };
-    }
-    Object.defineProperty(o, k2, desc);
-}) : (function(o, m, k, k2) {
-    if (k2 === undefined) k2 = k;
-    o[k2] = m[k];
-}));
-var __setModuleDefault = (this && this.__setModuleDefault) || (Object.create ? (function(o, v) {
-    Object.defineProperty(o, "default", { enumerable: true, value: v });
-}) : function(o, v) {
-    o["default"] = v;
-});
-var __importStar = (this && this.__importStar) || (function () {
-    var ownKeys = function(o) {
-        ownKeys = Object.getOwnPropertyNames || function (o) {
-            var ar = [];
-            for (var k in o) if (Object.prototype.hasOwnProperty.call(o, k)) ar[ar.length] = k;
-            return ar;
-        };
-        return ownKeys(o);
-    };
-    return function (mod) {
-        if (mod && mod.__esModule) return mod;
-        var result = {};
-        if (mod != null) for (var k = ownKeys(mod), i = 0; i < k.length; i++) if (k[i] !== "default") __createBinding(result, mod, k[i]);
-        __setModuleDefault(result, mod);
-        return result;
-    };
-})();
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.version = exports.toolDefinition = void 0;
-exports.execute = execute;
-const auth_js_1 = require("../../shared/auth.js");
-const error_handler_js_1 = require("../../shared/error-handler.js");
-const data_fetch_safety_js_1 = require("../../../../utils/data-fetch-safety.js");
-const tf = __importStar(require("@tensorflow/tfjs"));
-exports.toolDefinition = {
-    name: 'ml_train_incident_classifier',
-    description: '⚠️ LOCAL ML TRAINING: Trains LSTM neural networks on your machine using ServiceNow incident data fetched via API. NOT in ServiceNow. Alternative to PI license for dev/testing. Fetches up to 5000 records.',
-    // Metadata for tool discovery (not sent to LLM)
-    category: 'ml-analytics',
-    subcategory: 'machine-learning',
-    use_cases: ['training', 'classification', 'local-ml'],
-    complexity: 'advanced',
-    frequency: 'low',
-    inputSchema: {
-        type: 'object',
-        properties: {
-            sample_size: {
-                type: 'number',
-                description: 'Number of incidents to use for training. If not specified, automatically uses all available data (up to 5000).'
-            },
-            auto_maximize_data: {
-                type: 'boolean',
-                description: 'Automatically use all available incident data for best model accuracy',
-                default: true
-            },
-            epochs: {
-                type: 'number',
-                description: 'Training epochs',
-                default: 50
-            },
-            validation_split: {
-                type: 'number',
-                description: 'Validation data percentage',
-                default: 0.2
-            },
-            query: {
-                type: 'string',
-                description: 'Custom ServiceNow query for selecting training data'
-            },
-            intelligent_selection: {
-                type: 'boolean',
-                description: 'Let Snow-Flow intelligently select balanced training data',
-                default: true
-            },
-            focus_categories: {
-                type: 'array',
-                items: { type: 'string' },
-                description: 'Specific categories to focus on for training'
-            },
-            batch_size: {
-                type: 'number',
-                description: 'Process data in batches to prevent memory overload',
-                default: 100
-            },
-            max_vocabulary_size: {
-                type: 'number',
-                description: 'Maximum vocabulary size using feature hashing',
-                default: 10000
-            },
-            streaming_mode: {
-                type: 'boolean',
-                description: 'Enable streaming mode for very large datasets',
-                default: true
-            }
-        }
-    }
-};
-async function execute(args, context) {
-    const { sample_size, auto_maximize_data = true, epochs = 50, validation_split = 0.2, query = '', intelligent_selection = true, focus_categories = [], batch_size = 100, max_vocabulary_size = 10000, streaming_mode = true } = args;
-    try {
-        const client = await (0, auth_js_1.getAuthenticatedClient)(context);
-        // Ensure max_vocabulary_size is ALWAYS valid
-        const validVocabSize = Math.max(1000, max_vocabulary_size);
-        // Initialize TensorFlow.js
-        await tf.ready();
-        // Determine optimal sample size
-        let actualSampleSize = sample_size || 2000;
-        if (auto_maximize_data || !sample_size) {
-            const countQuery = query || (intelligent_selection ?
-                'categoryISNOTEMPTY^descriptionISNOTEMPTY^sys_created_onONLast 6 months' :
-                '');
-            try {
-                // Try to count available incidents
-                const response = await client.query('incident', {
-                    query: countQuery,
-                    limit: 1,
-                    count: true
-                });
-                const totalAvailable = response.count || 0;
-                const maxRecommended = 5000;
-                const optimalSize = Math.min(totalAvailable, maxRecommended);
-                if (totalAvailable > 0) {
-                    actualSampleSize = sample_size ?
-                        Math.min(sample_size, totalAvailable) :
-                        optimalSize;
-                    console.log(`Using ${actualSampleSize} incidents for training (optimal for this dataset)`);
-                }
-            }
-            catch (error) {
-                actualSampleSize = sample_size || 1000;
-            }
-        }
-        // Safety check: warn about large data fetches
-        const approval = (0, data_fetch_safety_js_1.requestApproval)({
-            table: 'incident',
-            estimatedRecords: actualSampleSize,
-            query: query || 'Last 6 months, non-empty category/description',
-            purpose: 'Train LSTM incident classifier locally'
-        });
-        if (!approval.approved) {
-            return (0, error_handler_js_1.createErrorResult)('Data fetch cancelled by user');
-        }
-        const fetchStartTime = Date.now();
-        // Fetch incident data
-        const finalQuery = query || (intelligent_selection ?
-            'categoryISNOTEMPTY^descriptionISNOTEMPTY^sys_created_onONLast 6 months' :
-            'categoryISNOTEMPTY');
-        const incidents = await client.query('incident', {
-            query: finalQuery,
-            limit: actualSampleSize,
-            fields: ['short_description', 'description', 'category', 'priority', 'impact', 'urgency']
-        });
-        const fetchTime = Date.now() - fetchStartTime;
-        // Log fetch summary
-        const summary = (0, data_fetch_safety_js_1.formatFetchSummary)({
-            table: 'incident',
-            estimatedRecords: actualSampleSize,
-            purpose: 'Train LSTM incident classifier'
-        }, fetchTime, incidents.length);
-        console.log(summary);
-        // Check data availability
-        const requiredRecords = 100;
-        const availableRecords = incidents.length;
-        const canTrain = availableRecords >= requiredRecords;
-        if (!incidents || incidents.length === 0) {
-            return (0, error_handler_js_1.createErrorResult)('No incidents found for training', {
-                data_availability: {
-                    required_records: requiredRecords,
-                    available_records: 0,
-                    can_train: false,
-                    recommendation: 'No incident data available. Ensure your instance has incident records with category and description fields populated.'
-                }
-            });
-        }
-        if (!canTrain) {
-            return (0, error_handler_js_1.createErrorResult)(`Insufficient training data: found ${availableRecords} records, need at least ${requiredRecords}`, {
-                data_availability: {
-                    required_records: requiredRecords,
-                    available_records: availableRecords,
-                    can_train: false,
-                    recommendation: `Need at least ${requiredRecords} incident records for reliable training. Currently have ${availableRecords}. Consider adjusting your query or waiting for more data.`
-                }
-            });
-        }
-        // Prepare training data
-        const { features, labels, tokenizer, categories } = prepareIncidentData(incidents, validVocabSize);
-        const vocabularySize = tokenizer.get('_vocabulary_size') || validVocabSize;
-        // Create LSTM model
-        const model = tf.sequential({
-            layers: [
-                // Embedding layer for text
-                tf.layers.embedding({
-                    inputDim: vocabularySize,
-                    outputDim: 128,
-                    inputLength: 100
-                }),
-                // LSTM for sequence processing
-                tf.layers.lstm({
-                    units: 64,
-                    returnSequences: false,
-                    dropout: 0.2,
-                    recurrentDropout: 0.2
-                }),
-                // Dense layers
-                tf.layers.dense({
-                    units: 32,
-                    activation: 'relu'
-                }),
-                tf.layers.dropout({ rate: 0.3 }),
-                // Output layer
-                tf.layers.dense({
-                    units: categories.length,
-                    activation: 'softmax'
-                })
-            ]
-        });
-        // Compile model
-        model.compile({
-            optimizer: tf.train.adam(0.001),
-            loss: 'categoricalCrossentropy',
-            metrics: ['accuracy']
-        });
-        // Train model
-        const history = await model.fit(features, labels, {
-            epochs,
-            validationSplit: validation_split,
-            batchSize: 32,
-            callbacks: {
-                onEpochEnd: (epoch, logs) => {
-                    const loss = logs?.loss ? logs.loss.toFixed(4) : 'N/A';
-                    const accuracy = logs?.acc ? (logs.acc * 100).toFixed(2) : 'N/A';
-                    console.log(`Epoch ${epoch + 1}/${epochs} - Loss: ${loss}, Accuracy: ${accuracy}%`);
-                }
-            }
-        });
-        // Clean up tensors
-        features.dispose();
-        labels.dispose();
-        // Calculate final metrics
-        const finalAccuracyValue = history.history.acc[history.history.acc.length - 1];
-        const finalAccuracy = typeof finalAccuracyValue === 'number' ? finalAccuracyValue : Array.isArray(finalAccuracyValue) ? finalAccuracyValue[0] : 0;
-        const finalLossValue = history.history.loss[history.history.loss.length - 1];
-        const finalLoss = typeof finalLossValue === 'number' ? finalLossValue : Array.isArray(finalLossValue) ? finalLossValue[0] : 0;
-        return (0, error_handler_js_1.createSuccessResult)({
-            status: 'success',
-            message: 'Incident classifier trained successfully',
-            training_summary: {
-                samples: incidents.length,
-                epochs: epochs,
-                categories: categories.length,
-                vocabulary_size: vocabularySize,
-                final_accuracy: (finalAccuracy * 100).toFixed(2) + '%',
-                final_loss: finalLoss.toFixed(4)
-            },
-            categories: categories,
-            data_availability: {
-                required_records: requiredRecords,
-                available_records: availableRecords,
-                can_train: true,
-                recommendation: `Successfully trained on ${availableRecords} records. Model is ready for predictions.`
-            }
-        });
-    }
-    catch (error) {
-        return (0, error_handler_js_1.createErrorResult)(error.message);
-    }
-}
-/**
- * Prepare incident data for neural network training
- */
-function prepareIncidentData(incidents, maxVocabularySize) {
-    const validVocabSize = Math.max(1000, maxVocabularySize);
-    const hasher = createFeatureHasher(validVocabSize);
-    let categories = [...new Set(incidents.map((i) => i.category))].filter((c) => c);
-    // Ensure at least 2 categories
-    if (categories.length < 2) {
-        if (categories.length === 0) {
-            categories = ['uncategorized', 'other'];
-        }
-        else {
-            categories.push('other');
-        }
-    }
-    const sequences = [];
-    const labels = [];
-    for (const incident of incidents) {
-        const text = `${incident.short_description || ''} ${incident.description || ''}`;
-        const sequence = hasher(text);
-        sequences.push(sequence);
-        // One-hot encode category
-        const category = incident.category || 'uncategorized';
-        const categoryIndex = categories.indexOf(category);
-        const label = new Array(categories.length).fill(0);
-        if (categoryIndex >= 0) {
-            label[categoryIndex] = 1;
-        }
-        else {
-            label[0] = 1;
-        }
-        labels.push(label);
-    }
-    const tokenizerMap = new Map();
-    tokenizerMap.set('_vocabulary_size', validVocabSize);
-    return {
-        features: tf.tensor2d(sequences),
-        labels: tf.tensor2d(labels),
-        tokenizer: tokenizerMap,
-        categories
-    };
-}
-/**
- * Create feature hasher for vocabulary management
- */
-function createFeatureHasher(vocabSize) {
-    return (text) => {
-        const words = text.toLowerCase().split(/\s+/).slice(0, 100);
-        const sequence = new Array(100).fill(0);
-        for (let i = 0; i < words.length && i < 100; i++) {
-            let hash = 0;
-            for (let j = 0; j < words[i].length; j++) {
-                hash = ((hash << 5) - hash) + words[i].charCodeAt(j);
-                hash = hash & hash;
-            }
-            sequence[i] = Math.abs(hash) % vocabSize;
-        }
-        return sequence;
-    };
-}
-exports.version = '1.0.0';
-//# sourceMappingURL=ml_train_incident_classifier.js.map