From 0048bc8ed41f36e085bba6b7ec4ef1beb0d4f0ec Mon Sep 17 00:00:00 2001
From: James Gowdy <jgowdy@elastic.co>
Date: Tue, 22 Jun 2021 21:24:31 +0100
Subject: [PATCH] [ML] Fixing categorization tokens for multi-line messages

---
 .../models/job_service/new_job/categorization/examples.ts    | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/x-pack/plugins/ml/server/models/job_service/new_job/categorization/examples.ts b/x-pack/plugins/ml/server/models/job_service/new_job/categorization/examples.ts
index 5fecb3d9eb1ec..abda62097dfb7 100644
--- a/x-pack/plugins/ml/server/models/job_service/new_job/categorization/examples.ts
+++ b/x-pack/plugins/ml/server/models/job_service/new_job/categorization/examples.ts
@@ -145,10 +145,11 @@ export function categorizationExamplesProvider({
         for (let g = 0; g < sumLengths.length; g++) {
           if (t.start_offset <= sumLengths[g] + g) {
             const offset = g > 0 ? sumLengths[g - 1] + g : 0;
+            const start = t.start_offset - offset;
             tokensPerExample[g].push({
               ...t,
-              start_offset: t.start_offset - offset,
-              end_offset: t.end_offset - offset,
+              start_offset: start,
+              end_offset: start + t.token.length,
             });
             break;
           }