From 6369d5f4627065f5285feecb54adbdde16bfd998 Mon Sep 17 00:00:00 2001
From: Predrag Stojadinovic <predrag@stojadinovic.net>
Date: Sat, 11 May 2024 18:43:55 +0200
Subject: [PATCH 1/3] chore: confluence data connector can now handle custom
 urls, in addition to default {subdomain}.atlassian.net ones

---
 .../utils/extensions/Confluence/index.js      | 44 ++++++++++++-------
 1 file changed, 29 insertions(+), 15 deletions(-)

diff --git a/collector/utils/extensions/Confluence/index.js b/collector/utils/extensions/Confluence/index.js
index 5a473f654cc..63094a1f99e 100644
--- a/collector/utils/extensions/Confluence/index.js
+++ b/collector/utils/extensions/Confluence/index.js
@@ -2,20 +2,30 @@ const fs = require("fs");
 const path = require("path");
 const { default: slugify } = require("slugify");
 const { v4 } = require("uuid");
+const UrlPattern = require("url-pattern");
 const { writeToServerDocuments } = require("../../files");
 const { tokenizeString } = require("../../tokenizer");
-const {
-  ConfluencePagesLoader,
-} = require("langchain/document_loaders/web/confluence");
+const { ConfluencePagesLoader } = require("langchain/document_loaders/web/confluence");
 
 function validSpaceUrl(spaceUrl = "") {
-  const UrlPattern = require("url-pattern");
-  const pattern = new UrlPattern(
-    "https\\://(:subdomain).atlassian.net/wiki/spaces/(:spaceKey)*"
-  );
-  const match = pattern.match(spaceUrl);
-  if (!match) return { valid: false, result: null };
-  return { valid: true, result: match };
+  // Atlassian default URL match
+  const atlassianPattern = new UrlPattern("https\\://(:subdomain).atlassian.net/wiki/spaces/(:spaceKey)/*");
+  const atlassianMatch = atlassianPattern.match(spaceUrl);
+  if (atlassianMatch) {
+    return { valid: true, result: atlassianMatch };
+  }
+
+  // Custom Confluence URL match
+  const customPattern = new UrlPattern("https\\://(:subdomain.):domain.:tld/wiki/spaces/(:spaceKey)/*");
+  const customMatch = customPattern.match(spaceUrl);
+  if (customMatch) {
+    customMatch.customDomain = (customMatch.subdomain ? `${customMatch.subdomain}.` : "") + //
+      (`${customMatch.domain}.${customMatch.tld}`);
+    return { valid: true, result: customMatch, custom: true };
+  }
+
+  // No match
+  return { valid: false, result: null };
 }
 
 async function loadConfluence({ pageUrl, username, accessToken }) {
@@ -31,15 +41,19 @@ async function loadConfluence({ pageUrl, username, accessToken }) {
   if (!validSpace.result) {
     return {
       success: false,
-      reason:
-        "Confluence space URL is not in the expected format of https://domain.atlassian.net/wiki/space/~SPACEID/*",
+      reason: "Confluence space URL is not in the expected format of https://domain.atlassian.net/wiki/space/~SPACEID/* or https://customDomain/wiki/space/~SPACEID/*",
     };
   }
 
-  const { subdomain, spaceKey } = validSpace.result;
-  console.log(`-- Working Confluence ${subdomain}.atlassian.net --`);
+  const { subdomain, customDomain, spaceKey } = validSpace.result;
+  let baseUrl = `https://${subdomain}.atlassian.net/wiki`;
+  if (customDomain) {
+    baseUrl = `https://${customDomain}/wiki`;
+  }
+
+  console.log(`-- Working Confluence ${baseUrl} --`);
   const loader = new ConfluencePagesLoader({
-    baseUrl: `https://${subdomain}.atlassian.net/wiki`,
+    baseUrl,
     spaceKey,
     username,
     accessToken,

From c73c5087e2f7a724d00a1186922f46c07622bc57 Mon Sep 17 00:00:00 2001
From: Predrag Stojadinovic <predrag@stojadinovic.net>
Date: Sat, 11 May 2024 19:04:28 +0200
Subject: [PATCH 2/3] chore: formatting as per yarn lint

---
 .../utils/extensions/Confluence/index.js      | 20 +++++++++++++------
 1 file changed, 14 insertions(+), 6 deletions(-)

diff --git a/collector/utils/extensions/Confluence/index.js b/collector/utils/extensions/Confluence/index.js
index 63094a1f99e..5bb44deeb75 100644
--- a/collector/utils/extensions/Confluence/index.js
+++ b/collector/utils/extensions/Confluence/index.js
@@ -5,22 +5,29 @@ const { v4 } = require("uuid");
 const UrlPattern = require("url-pattern");
 const { writeToServerDocuments } = require("../../files");
 const { tokenizeString } = require("../../tokenizer");
-const { ConfluencePagesLoader } = require("langchain/document_loaders/web/confluence");
+const {
+  ConfluencePagesLoader,
+} = require("langchain/document_loaders/web/confluence");
 
 function validSpaceUrl(spaceUrl = "") {
   // Atlassian default URL match
-  const atlassianPattern = new UrlPattern("https\\://(:subdomain).atlassian.net/wiki/spaces/(:spaceKey)/*");
+  const atlassianPattern = new UrlPattern(
+    "https\\://(:subdomain).atlassian.net/wiki/spaces/(:spaceKey)/*"
+  );
   const atlassianMatch = atlassianPattern.match(spaceUrl);
   if (atlassianMatch) {
     return { valid: true, result: atlassianMatch };
   }
 
   // Custom Confluence URL match
-  const customPattern = new UrlPattern("https\\://(:subdomain.):domain.:tld/wiki/spaces/(:spaceKey)/*");
+  const customPattern = new UrlPattern(
+    "https\\://(:subdomain.):domain.:tld/wiki/spaces/(:spaceKey)/*"
+  );
   const customMatch = customPattern.match(spaceUrl);
   if (customMatch) {
-    customMatch.customDomain = (customMatch.subdomain ? `${customMatch.subdomain}.` : "") + //
-      (`${customMatch.domain}.${customMatch.tld}`);
+    customMatch.customDomain =
+      (customMatch.subdomain ? `${customMatch.subdomain}.` : "") + //
+      `${customMatch.domain}.${customMatch.tld}`;
     return { valid: true, result: customMatch, custom: true };
   }
 
@@ -41,7 +48,8 @@ async function loadConfluence({ pageUrl, username, accessToken }) {
   if (!validSpace.result) {
     return {
       success: false,
-      reason: "Confluence space URL is not in the expected format of https://domain.atlassian.net/wiki/space/~SPACEID/* or https://customDomain/wiki/space/~SPACEID/*",
+      reason:
+        "Confluence space URL is not in the expected format of https://domain.atlassian.net/wiki/space/~SPACEID/* or https://customDomain/wiki/space/~SPACEID/*",
     };
   }
 

From ec4177340b097286a6924bb58b9b6113ba1d5cb6 Mon Sep 17 00:00:00 2001
From: Predrag Stojadinovic <predrag@stojadinovic.net>
Date: Thu, 16 May 2024 18:20:53 +0200
Subject: [PATCH 3/3] chore: adding /display/ url matching to confluence data
 connector

---
 collector/utils/extensions/Confluence/index.js | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/collector/utils/extensions/Confluence/index.js b/collector/utils/extensions/Confluence/index.js
index da918b2d219..ff91773cbe3 100644
--- a/collector/utils/extensions/Confluence/index.js
+++ b/collector/utils/extensions/Confluence/index.js
@@ -24,7 +24,14 @@ function validSpaceUrl(spaceUrl = "") {
     "https\\://(:subdomain.):domain.:tld/wiki/spaces/(:spaceKey)/*"
   );
   const customMatch = customPattern.match(spaceUrl);
-  if (customMatch) {
+
+  // Custom "display" Confluence URL match
+  const customDisplayPattern = new UrlPattern(
+    "https\\://(:subdomain.):domain.:tld/display/(:spaceKey)/*"
+  );
+  const customDisplayMatch = customDisplayPattern.match(spaceUrl);
+
+  if (customMatch || customDisplayMatch) {
     customMatch.customDomain =
       (customMatch.subdomain ? `${customMatch.subdomain}.` : "") + //
       `${customMatch.domain}.${customMatch.tld}`;