Recognize quotes when parsing urls in logs (#40508)

--------- Co-authored-by: Jacob Lee <seungmin_lee3@apple.com>
apache · Jul 2, 2024 · 2f9b23d · 2f9b23d
1 parent eb3e199
commit 2f9b23d
Show file tree

Hide file tree

Showing 2 changed files with 23 additions and 4 deletions.
diff --git a/airflow/www/static/js/dag/details/taskInstance/Logs/utils.test.tsx b/airflow/www/static/js/dag/details/taskInstance/Logs/utils.test.tsx
@@ -38,6 +38,7 @@ const mockTaskLog = `
 [2022-06-04 00:00:01,921] {dagbag.py:507} INFO - Filling up the DagBag from /files/dags/test_ui_grid.py
 [2022-06-04 00:00:01,964] {task_command.py:377} INFO - Running <TaskInstance: test_ui_grid.section_1.get_entry_group scheduled__2022-06-03T00:00:00+00:00 [running]> on host 5d28cfda3219
 [2022-06-04 00:00:02,010] {taskinstance.py:1548} WARNING - Exporting env vars: AIRFLOW_CTX_DAG_OWNER=*** AIRFLOW_CTX_DAG_ID=test_ui_grid
+[2024-07-01 00:00:02,010] {taskinstance.py:1548} INFO - Url parsing test => "https://apple.com", "https://google.com"
 `;
 
 describe("Test Logs Utils.", () => {
@@ -64,7 +65,7 @@ describe("Test Logs Utils.", () => {
   test.each([
     {
       logLevelFilters: [LogLevel.INFO],
-      expectedNumberOfLines: 11,
+      expectedNumberOfLines: 12,
       expectedNumberOfFileSources: 4,
     },
     {
@@ -111,7 +112,7 @@ describe("Test Logs Utils.", () => {
       "taskinstance.py",
     ]);
     const lines = parsedLogs!.split("\n");
-    expect(lines).toHaveLength(7);
+    expect(lines).toHaveLength(8);
     lines.forEach((line) => expect(line).toContain("taskinstance.py"));
   });
 
@@ -131,7 +132,25 @@ describe("Test Logs Utils.", () => {
       "taskinstance.py",
     ]);
     const lines = parsedLogs!.split("\n");
-    expect(lines).toHaveLength(7);
+    expect(lines).toHaveLength(8);
     lines.forEach((line) => expect(line).toMatch(/INFO|WARNING/));
   });
+
+  test("parseLogs function with quoted urls", () => {
+    const { parsedLogs } = parseLogs(
+      mockTaskLog,
+      null,
+      [LogLevel.INFO, LogLevel.WARNING],
+      ["taskinstance.py"],
+      []
+    );
+
+    const lines = parsedLogs!.split("\n");
+    expect(lines[lines.length - 1]).toContain(
+      '<a href="https://apple.com" target="_blank" style="color: blue; text-decoration: underline;">https://apple.com</a>'
+    );
+    expect(lines[lines.length - 1]).toContain(
+      '<a href="https://google.com" target="_blank" style="color: blue; text-decoration: underline;">https://google.com</a>'
+    );
+  });
 });
diff --git a/airflow/www/static/js/dag/details/taskInstance/Logs/utils.ts b/airflow/www/static/js/dag/details/taskInstance/Logs/utils.ts
@@ -74,7 +74,7 @@ export const parseLogs = (
   const ansiUp = new AnsiUp();
   ansiUp.url_allowlist = {};
 
-  const urlRegex = /((https?:\/\/|http:\/\/)[^\s]+)/g;
+  const urlRegex = /((https?:\/\/|http:\/\/)(?:(?!&#x27;|&quot;)[^\s])+)/g;
   // Detect log groups which can be collapsed
   // Either in Github like format '::group::<group name>' to '::endgroup::'
   // see https://docs.github.com/en/actions/using-workflows/workflow-commands-for-github-actions#grouping-log-lines