/history/by_task_id : a method to link each task's history uniquely to the SQL database

doc/central_scheduler.rst

@@ -95,3 +95,14 @@ The task history has the following pages:
   a listing of all runs of the task ``{name}`` restricted to runs with ``params`` matching the given history.
   The ``params`` is a json blob describing the parameters,
   e.g. ``data={"foo": "bar"}`` looks for a task with ``foo=bar``.
+* ``/history/by_task_id/{task_id}``
+  the latest run of a task given the ``{task_id}``. It is different from just ``{id}``
+  and is a derivative of ``params``. It is available via ``{task_id}`` property of a 
+  ``luigi.Task`` instance or via `luigi.task.task_id_str
+  <https://luigi.readthedocs.io/en/stable/api/luigi.task.html#luigi.task.task_id_str>`_.
+  This kind of representation is useful for concisely recording URLs in a history tree.
+  Example screenshot:
+
+    .. figure:: history_by_task_id.png
+       :alt: By task_id screenshot
+

luigi/db_task_history.py

@@ -187,6 +187,13 @@ def find_task_by_id(self, id, session=None):
         with self._session(session) as session:
             return session.query(TaskRecord).get(id)
 
+    def find_task_by_task_id(self, task_id, session=None):
+        """
+        Find task with the given task ID.
+        """
+        with self._session(session) as session:
+            return session.query(TaskRecord).filter(TaskRecord.task_id == task_id).all()[-1]
+
 
 class TaskParameter(Base):  # type: ignore
     """

luigi/server.py

@@ -264,6 +264,13 @@ def get(self, id):
             self.render("show.html", task=task)
 
 
+class ByTaskIdHandler(BaseTaskHistoryHandler):
+    def get(self, task_id):
+        with self._scheduler.task_history._session(None) as session:
+            task = self._scheduler.task_history.find_task_by_task_id(task_id, session)
+            self.render("show.html", task=task)
+
+
 class ByParamsHandler(BaseTaskHistoryHandler):
     def get(self, name):
         payload = self.get_argument('data', default="{}")
@@ -314,6 +321,7 @@ def app(scheduler):
         (r'/history', RecentRunHandler, {'scheduler': scheduler}),
         (r'/history/by_name/(.*?)', ByNameHandler, {'scheduler': scheduler}),
         (r'/history/by_id/(.*?)', ByIdHandler, {'scheduler': scheduler}),
+        (r'/history/by_task_id/(.*?)', ByTaskIdHandler, {'scheduler': scheduler}),
         (r'/history/by_params/(.*?)', ByParamsHandler, {'scheduler': scheduler}),
         (r'/metrics', MetricsHandler, {'scheduler': scheduler})
     ]