make-mp-data 2.1.11 → 3.0.2

package/dungeons/benchmark-light.js

@@ -0,0 +1,141 @@
+/**
+ * Benchmark Dungeon Configuration
+ * Generates one million events across 10,000 users
+ * ensure we have multiple property types and distributions
+ */
+
+
+
+
+import Chance from 'chance';
+let chance = new Chance();
+import dayjs from "dayjs";
+import utc from "dayjs/plugin/utc.js";
+dayjs.extend(utc);
+import { uid, comma } from 'ak-tools';
+import { pickAWinner, weighNumRange, date, integer, weighChoices } from "../lib/utils/utils.js";
+import { createTextGenerator } from '../lib/generators/text.js';
+
+/** @type {import('../types.js').Dungeon} */
+const config = {
+	// token: "",
+	name: "5M-Events-Light",
+	format: 'json', //csv or json
+	seed: "one million events",
+	numDays: 92, //how many days worth of data
+	numEvents: 5_000_000, //how many events
+	numUsers: 10_000, //how many users
+	strictEventCount: true,
+	
+	region: "US",
+	hasAnonIds: false, //if true, anonymousIds are created for each user
+	hasSessionIds: false, //if true, hasSessionIds are created for each user
+	hasAdSpend: false,
+	makeChart: false,
+	hasLocation: false,
+	hasAndroidDevices: false,
+	hasIOSDevices: false,
+	hasDesktopDevices: false,
+	hasBrowser: false,
+	hasCampaigns: false,
+	isAnonymous: false,
+	alsoInferFunnels: true,
+	// concurrency automatically set to 1 when strictEventCount is enabled
+	writeToDisk: false,
+	batchSize: 12_500_000,
+
+	events: [
+		{
+			event: "Page View",
+			weight: 50,
+
+		},
+		{
+			event: "Product View",
+			weight: 30,
+		},
+		{
+			event: "Add to Cart",
+			weight: 10,
+		},
+		{
+			event: "Checkout Started",
+			weight: 5,
+		},
+		{
+			event: "Purchase",
+			weight: 1,
+		},
+		{
+			event: "Product Review",
+			weight: 3,
+		},
+		{
+			event: "Search",
+			weight: 17,
+		},
+		{
+			event: "Browse",
+			weight: 25,
+
+		},
+		{
+			event: "Save for Later",
+			weight: 4,
+		},
+		{
+			event: "Remove from Cart",
+			weight: 2,
+		},
+		{
+			event: "Empty Cart",
+			weight: 4,
+		}
+	],
+	funnels: [
+		{
+			"name": "Purchase Funnel",
+			"sequence": [
+				"Page View",
+				"Product View",
+				"Add to Cart",
+				"Checkout Started",
+				"Purchase"
+			],
+			requireRepeats: true,
+			"conversionRate": 30,
+			"order": "fixed",
+			"weight": 1,
+			"isFirstFunnel": false,
+			"timeToConvert": 5,
+			"experiment": true,
+		}
+	],
+	superProps: {
+	},
+	/*
+	user properties work the same as event properties
+	each key should be an array or function reference
+	*/
+	userProps: {},
+	scdProps: {},
+	mirrorProps: {},
+
+	/*
+	for group analytics keys, we need an array of arrays [[],[],[]] 
+	each pair represents a group_key and the number of profiles for that key
+	*/
+	groupKeys: [],
+	groupProps: {},
+	lookupTables: [],
+	hook: function (record, type, meta) {
+		// if (type === "everything") {
+		// 	debugger;
+		// }
+		return record;
+	}
+};
+
+
+
+export default config;

package/dungeons/big.js

@@ -1,34 +1,35 @@
 import Chance from 'chance';
 import dayjs from 'dayjs';
-import utc from 'dayjs/plugin/utc';
+import utc from 'dayjs/plugin/utc.js';
 import { uid, comma, makeName } from 'ak-tools';
-import { pickAWinner, weighNumRange, integer, date, choose } from '../lib/utils/utils.js';
+import { pickAWinner, weighNumRange, integer, date, choose } from "../lib/utils/utils.js";
 
 const seed = "lets go big";
 const chance = new Chance();
 dayjs.extend(utc);
 
 
-const eventsPerQuarter = 5_000_000_000 // ~5 billion
-const numQuarters = 8; // 24 months
-const parallelism = 5000; 
-const totalEvents = Math.floor((eventsPerQuarter * numQuarters) / parallelism);
+const eventsPerQuarter = 10_000_000 // 50M
+const numQuarters = 4; // 12 months
+// const parallelism = 5000; 
+const totalEvents = Math.floor((eventsPerQuarter * numQuarters));
 const eventPerUser = 500;
 const totalUsers = Math.floor(totalEvents / eventPerUser);
 const totalDays = (numQuarters * 90) + 10;
 
 /** @type {import('../types').Dungeon} */
 const config = {
-	token: "",
+	// token: "",
 	seed: seed,
 	numDays: totalDays,
 	numEvents: totalEvents,
 	numUsers: totalUsers,
+	strictEventCount: true,
 	format: 'json', //csv or json
 	region: "US",
 	hasAnonIds: false, //if true, anonymousIds are created for each user
 	hasSessionIds: false, //if true, hasSessionIds are created for each user
-	hasLocation: true,
+	hasLocation: false,
 	hasAndroidDevices: false,
 	alsoInferFunnels: false,
 	batchSize: 2_000_000,
@@ -38,7 +39,7 @@ const config = {
 	hasCampaigns: false,
 	hasDesktopDevices: false,
 	hasIOSDevices: false,
-	writeToDisk: "gs://dungeon_master_4/big_data",
+	writeToDisk: false,
 	funnels: [
 		{
 			"sequence": ["foo", "bar", "baz", "qux", "garply", "durtle", "linny", "fonk", "crumn", "yak"],

package/dungeons/business.js

@@ -1,20 +1,21 @@
 /**
- * This is the default configuration file for the data generator in COMPLEX mode
- * notice how the config object is structured, and see it's type definition in ./types.d.ts
- * feel free to modify this file to customize the data you generate
- * see helper functions in utils.js for more ways to generate data
+ * Video platform dungeon (business/complex mode)
  */
 
+import Chance from 'chance';
+import dayjs from 'dayjs';
+import utc from 'dayjs/plugin/utc.js';
+dayjs.extend(utc);
+import { weighNumRange, pickAWinner, exhaust } from '../lib/utils/utils.js';
+import * as v from 'ak-tools';
 
-const Chance = require('chance');
 const chance = new Chance();
-const { weighNumRange, date, integer, pickAWinner, exhaust } = require('../lib/utils/utils.js');
-const u = require('ak-tools');
+const integer = (min, max) => chance.integer({ min, max });
 
 const channel_ids = [...Array(1234).keys()].map(i => i + 1).map(n => `channel_id_${n}`);
-const channel_names = chance.n(u.makeName, 1234);
+const channel_names = chance.n(v.makeName, 1234);
 const video_ids = [...Array(50000).keys()].map(i => i + 1).map(n => n.toString());
-const video_names = chance.n(u.makeName, 50000);
+const video_names = chance.n(v.makeName, 50000);
 
 const EVENTS = 50_000
 const USERS = EVENTS / 100
@@ -22,7 +23,7 @@ const USERS = EVENTS / 100
 
 /** @type {import('../types.js').Dungeon} */
 const config = {
-	token: "",	
+	token: "",
 	seed: "it's business time...",
 	numDays: 90, //how many days worth of data
 	numEvents: EVENTS, //how many events
@@ -41,6 +42,7 @@ const config = {
 	hasAdSpend: true,
 
 	hasAvatar: false,
+	makeChart: true,
 
 	batchSize: 500_000,
 	concurrency: 500,
@@ -257,6 +259,52 @@ const config = {
 	],
 
 	hook: function (record, type, meta) {
+		// --- user hook: tag users by their experiment variant ---
+		if (type === "user") {
+			const exp = record.experiment;
+			if (exp && Array.isArray(exp) && exp[0]) {
+				record.experimentGroup = exp[0].variant || "unknown";
+			}
+			return record;
+		}
+
+		// --- event hook: weekend watch time boost + premium quality bias ---
+		if (type === "event") {
+			if (record.event === "watch video" && record.time) {
+				const day = dayjs(record.time).day();
+				if (day === 0 || day === 6) {
+					record["watch time (sec)"] = Math.round((record["watch time (sec)"] || 60) * 1.8);
+					record.is_weekend_session = true;
+				}
+			}
+			// app errors on weekends are more severe (skeleton crew)
+			if (record.event === "app error" && record.time) {
+				const day = dayjs(record.time).day();
+				if (day === 0 || day === 6) {
+					record.weekend_incident = true;
+				}
+			}
+			return record;
+		}
+
+		// --- everything hook: binge watchers (5+ watch events) get a "binge_session" event ---
+		if (type === "everything") {
+			const watches = record.filter(e => e.event === "watch video");
+			if (watches.length >= 5) {
+				const totalWatchTime = watches.reduce((sum, e) => sum + (e["watch time (sec)"] || 0), 0);
+				const lastWatch = watches[watches.length - 1];
+				record.push({
+					event: "binge_session",
+					time: dayjs(lastWatch.time).add(1, "minute").toISOString(),
+					user_id: lastWatch.user_id,
+					videos_watched: watches.length,
+					total_watch_time_sec: totalWatchTime,
+					avg_watch_time_sec: Math.round(totalWatchTime / watches.length)
+				});
+			}
+			return record;
+		}
+
 		return record;
 	}
 };
@@ -266,7 +314,7 @@ const config = {
 function makeHashTags() {
 	const possibleHashtags = [];
 	for (let i = 0; i < 20; i++) {
-		possibleHashtags.push('#' + u.makeName(2, ''));
+		possibleHashtags.push('#' + v.makeName(2, ''));
 	}
 
 	const numHashtags = integer(integer(1, 5), integer(5, 10));
@@ -341,4 +389,4 @@ function designExperiment() {
 
 
 
-module.exports = config;
+export default config;

package/dungeons/complex.js

@@ -8,7 +8,10 @@
 
 import Chance from 'chance';
 const chance = new Chance();
-import { weighNumRange, date, integer } from '../lib/utils/utils.js';
+import dayjs from "dayjs";
+import utc from "dayjs/plugin/utc.js";
+dayjs.extend(utc);
+import { weighNumRange, date, integer } from "../lib/utils/utils.js";
 import * as u from 'ak-tools';
 
 /** @type {import('../types.js').Dungeon} */
@@ -33,6 +36,7 @@ const config = {
 	hasAdSpend: true,
 
 	hasAvatar: true,
+	makeChart: false,
 
 	batchSize: 500_000,
 	concurrency: 10,
@@ -280,6 +284,36 @@ const config = {
 	],
 
 	hook: function (record, type, meta) {
+		// event hook: weekend watch time boost — videos watched on weekends get 1.5x duration
+		if (type === "event") {
+			if (record.event === "watch video" && record.time) {
+				const day = dayjs(record.time).day();
+				if (day === 0 || day === 6) {
+					record.watchTimeSec = Math.round((record.watchTimeSec || 60) * 1.5);
+					record.is_weekend = true;
+				}
+			}
+			// support tickets on high-severity get escalated flag
+			if (record.event === "support ticket" && record.severity === "high") {
+				record.escalated = true;
+			}
+		}
+
+		// everything hook: simulate cart abandonment — users who "add to cart" but never "checkout" lose their last add-to-cart
+		if (type === "everything") {
+			const hasCheckout = record.some(e => e.event === "checkout");
+			const hasAddToCart = record.some(e => e.event === "add to cart");
+			if (hasAddToCart && !hasCheckout) {
+				// mark all their add-to-cart events as abandoned
+				for (const e of record) {
+					if (e.event === "add to cart") {
+						e.abandoned = true;
+					}
+				}
+			}
+			return record;
+		}
+
 		return record;
 	}
 };