Add an option to generate the page view ID according to changes in the page URL to account for events tracked before page views in SPAs (close #1307 and #1125)

.bundlemonrc.json

@@ -7,7 +7,7 @@
     },
     {
       "path": "./trackers/browser-tracker/dist/index.umd.min.js",
-      "maxSize": "15.5kb",
+      "maxSize": "16kb",
       "maxPercentIncrease": 10
     },
     {
@@ -22,7 +22,7 @@
     },
     {
       "path": "./libraries/browser-tracker-core/dist/index.module.js",
-      "maxSize": "27kb",
+      "maxSize": "28kb",
       "maxPercentIncrease": 10
     },
     {

api-docs/docs/browser-tracker/browser-tracker.api.md

@@ -83,6 +83,7 @@ export interface BrowserTracker {
     namespace: string;
     newSession: () => void;
     preservePageViewId: () => void;
+    preservePageViewIdForUrl: (preserve: PreservePageViewIdForUrl) => void;
     setBufferSize: (newBufferSize: number) => void;
     setCollectorUrl: (collectorUrl: string) => void;
     setCookiePath: (path: string) => void;
@@ -291,6 +292,9 @@ export type PostBatch = Record<string, unknown>[];
 // @public
 export function preservePageViewId(trackers?: Array<string>): void;
 
+// @public (undocumented)
+export type PreservePageViewIdForUrl = boolean | "full" | "pathname" | "pathnameAndSearch";
+
 // @public
 export function removeGlobalContexts(contexts: Array<ConditionalContextProvider | ContextPrimitive>, trackers?: Array<string>): void;
 
@@ -417,6 +421,7 @@ export type TrackerConfiguration = {
     retryFailedRequests?: boolean;
     onRequestSuccess?: (data: EventBatch) => void;
     onRequestFailure?: (data: RequestFailure) => void;
+    preservePageViewIdForUrl?: PreservePageViewIdForUrl;
 };
 
 // @public

common/changes/@snowplow/browser-tracker-core/issue-1307-page_view_id_2024-05-02-13-51.json

@@ -0,0 +1,10 @@
+{
+  "changes": [
+    {
+      "packageName": "@snowplow/browser-tracker-core",
+      "comment": "Add an option to generate the page view ID according to changes in the page URL to account for events tracked before page views in SPAs (#1307 and #1125)",
+      "type": "none"
+    }
+  ],
+  "packageName": "@snowplow/browser-tracker-core"
+}

common/changes/@snowplow/browser-tracker/issue-1307-page_view_id_2024-05-02-14-18.json

@@ -0,0 +1,10 @@
+{
+  "changes": [
+    {
+      "packageName": "@snowplow/browser-tracker",
+      "comment": "Add an option to generate the page view ID according to changes in the page URL to account for events tracked before page views in SPAs (#1307 and #1125)",
+      "type": "none"
+    }
+  ],
+  "packageName": "@snowplow/browser-tracker"
+}

common/changes/@snowplow/javascript-tracker/issue-1307-page_view_id_2024-05-15-06-21.json

@@ -0,0 +1,10 @@
+{
+  "changes": [
+    {
+      "packageName": "@snowplow/javascript-tracker",
+      "comment": "Fix running integration tests due to a problem with Micro",
+      "type": "none"
+    }
+  ],
+  "packageName": "@snowplow/javascript-tracker"
+}

libraries/browser-tracker-core/src/state.ts

@@ -52,6 +52,8 @@ export class SharedState {
   /* pageViewId, which can changed by other trackers on page;
    * initialized by tracker sent first event */
   pageViewId?: string;
+  /* URL of the page view which the `pageViewId` was generated for */
+  pageViewUrl?: string;
 }
 
 export function createSharedState(): SharedState {

libraries/browser-tracker-core/src/tracker/index.ts

@@ -48,6 +48,7 @@ import {
   ClientSession,
   ExtendedCrossDomainLinkerOptions,
   ParsedIdCookie,
+  PreservePageViewIdForUrl,
 } from './types';
 import {
   parseIdCookie,
@@ -292,8 +293,10 @@ export function Tracker(
       ),
       // Whether pageViewId should be regenerated after each trackPageView. Affect web_page context
       preservePageViewId = false,
-      // Whether first trackPageView was fired and pageViewId should not be changed anymore until reload
-      pageViewSent = false,
+      // Whether pageViewId should be kept the same until the page URL changes. Affects web_page context
+      preservePageViewIdForUrl = trackerConfiguration.preservePageViewIdForUrl ?? false,
+      // The pageViewId of the last page view event or undefined if no page view tracked yet. Used to determine if pageViewId should be regenerated for a new page view.
+      lastSentPageViewId: string | undefined = undefined,
       // Activity tracking config for callback and page ping variants
       activityTrackingConfig: ActivityTrackingConfig = {
         enabled: false,
@@ -719,6 +722,7 @@ export function Tracker(
     function resetPageView() {
       if (!preservePageViewId || state.pageViewId == null) {
         state.pageViewId = uuid();
+        state.pageViewUrl = configCustomUrl || locationHrefAlias;
       }
     }
 
@@ -727,10 +731,42 @@ export function Tracker(
      * Generates it if it wasn't initialized by other tracker
      */
     function getPageViewId() {
-      if (state.pageViewId == null) {
+      if (shouldGenerateNewPageViewId()) {
         state.pageViewId = uuid();
+        state.pageViewUrl = configCustomUrl || locationHrefAlias;
+      }
+      return state.pageViewId!;
+    }
+
+    function shouldGenerateNewPageViewId() {
+      // If pageViewId is not initialized, generate it
+      if (state.pageViewId == null) {
+        return true;
+      }
+      // If pageViewId should be preserved regardless of the URL, don't generate a new one
+      if (preservePageViewId || !preservePageViewIdForUrl) {
+        return false;
       }
-      return state.pageViewId;
+      // If doesn't have previous URL in state, generate a new pageViewId
+      if (state.pageViewUrl === undefined) {
+        return true;
+      }
+      const current = configCustomUrl || locationHrefAlias;
+      // If full preserve is enabled, compare the full URL
+      if (preservePageViewIdForUrl === true || preservePageViewIdForUrl == 'full' || !('URL' in window)) {
+        return state.pageViewUrl != current;
+      }
+      const currentUrl = new URL(current);
+      const previousUrl = new URL(state.pageViewUrl);
+      // If pathname preserve is enabled, compare the pathname
+      if (preservePageViewIdForUrl == 'pathname') {
+        return currentUrl.pathname != previousUrl.pathname;
+      }
+      // If pathname and search preserve is enabled, compare the pathname and search
+      if (preservePageViewIdForUrl == 'pathnameAndSearch') {
+        return currentUrl.pathname != previousUrl.pathname || currentUrl.search != previousUrl.search;
+      }
+      return false;
     }
 
     /**
@@ -943,11 +979,11 @@ export function Tracker(
 
     function logPageView({ title, context, timestamp, contextCallback }: PageViewEvent & CommonEventProperties) {
       refreshUrl();
-      if (pageViewSent) {
-        // Do not reset pageViewId if previous events were not page_view
+      if (lastSentPageViewId && lastSentPageViewId == getPageViewId()) {
+        // Do not reset pageViewId if a page view was not tracked yet or a different page view ID was used (in order to support multiple trackers with shared state)
         resetPageView();
       }
-      pageViewSent = true;
+      lastSentPageViewId = getPageViewId();
 
       // So we know what document.title was at the time of trackPageView
       lastDocumentTitle = document.title;
@@ -1291,6 +1327,10 @@ export function Tracker(
         preservePageViewId = true;
       },
 
+      preservePageViewIdForUrl: function (preserve: PreservePageViewIdForUrl) {
+        preservePageViewIdForUrl = preserve;
+      },
+
       disableAnonymousTracking: function (configuration?: DisableAnonymousTrackingConfiguration) {
         trackerConfiguration.anonymousTracking = false;
 

libraries/browser-tracker-core/src/tracker/types.ts

@@ -46,6 +46,9 @@ export type ExtendedCrossDomainLinkerAttributes = {
 
 export type ExtendedCrossDomainLinkerOptions = boolean | ExtendedCrossDomainLinkerAttributes;
 
+/* Setting for the `preservePageViewIdForUrl` configuration that decides how to preserve the pageViewId on URL changes. */
+export type PreservePageViewIdForUrl = boolean | 'full' | 'pathname' | 'pathnameAndSearch';
+
 /**
  * The configuration object for initialising the tracker
  * @example
@@ -266,6 +269,17 @@ export type TrackerConfiguration = {
    * @param data - The data associated with the event(s) that failed to send
    */
   onRequestFailure?: (data: RequestFailure) => void;
+
+  /**
+   * Decide how the `pageViewId` should be preserved based on the URL.
+   * If set to `false`, the `pageViewId` will be regenerated on the second and each following page view event (first page view doesn't change the page view ID since tracker initialization).
+   * If set to `true` or `'full'`, the `pageViewId` will be kept the same for all page views with that exact URL (even for events tracked before the page view event).
+   * If set to `'pathname'`, the `pageViewId` will be kept the same for all page views with the same pathname (search params or fragment may change).
+   * If set to `'pathnameAndSearch'`, the `pageViewId` will be kept the same for all page views with the same pathname and search params (fragment may change).
+   * If `preservePageViewId` is enabled, the `preservePageViewIdForUrl` setting is ignored.
+   * Defaults to `false`.
+   */
+  preservePageViewIdForUrl?: PreservePageViewIdForUrl;
 };
 
 /**
@@ -603,6 +617,17 @@ export interface BrowserTracker {
    */
   preservePageViewId: () => void;
 
+  /**
+   * Decide how the `pageViewId` should be preserved based on the URL.
+   * If set to `false`, the `pageViewId` will be regenerated on the second and each following page view event (first page view doesn't change the page view ID since tracker initialization).
+   * If set to `true` or `'full'`, the `pageViewId` will be kept the same for all page views with that exact URL (even for events tracked before the page view event).
+   * If set to `'pathname'`, the `pageViewId` will be kept the same for all page views with the same pathname (search params or fragment may change).
+   * If set to `'pathnameAndSearch'`, the `pageViewId` will be kept the same for all page views with the same pathname and search params (fragment may change).
+   * If `preservePageViewId` is enabled, the `preservePageViewIdForUrl` setting is ignored.
+   * Defaults to `false`.
+   */
+  preservePageViewIdForUrl: (preserve: PreservePageViewIdForUrl) => void;
+
   /**
    * Log visit to this page
    *

libraries/browser-tracker-core/test/helpers/index.ts

@@ -75,7 +75,7 @@ export function createTestSessionIdCookie(params?: CreateTestSessionIdCookie) {
   return `_sp_ses.${domainHash}=*; Expires=; Path=/; SameSite=None; Secure;`;
 }
 
-export function createTracker(configuration?: TrackerConfiguration) {
+export function createTracker(configuration?: TrackerConfiguration, sharedState?: SharedState) {
   let id = 'sp-' + Math.random();
-  return addTracker(id, id, '', '', new SharedState(), configuration);
+  return addTracker(id, id, '', '', sharedState ?? new SharedState(), configuration);
 }