- 15 Aug 2024
- Print
- DarkLight
- PDF
Workflow Nodes
- Updated On 15 Aug 2024
- Print
- DarkLight
- PDF
Workflow Nodes
My Task Node
The My Task node allows you to create an annotation or QA task in the pipeline.
- Items that are added into the task node, will be added to the task defined in this node.
- Items will flow out of the task node based on the status assigned to them by the annotator or reviewer (such as Complete, Approved, Discard, etc.).
- The task created from the Task node, will be listed in project tasks (Dataloop main menu > Labeling).
- Edit task settings in View Mode: The Task Node supports editing of the task settings when the pipeline is running (in View mode).
Details
When you click on a My Task node, its details, such as Configuration, Executions, Logs, and available Actions are shown on the right-side panel.
For the actions available on each node in the right-side panel, see the Pipeline Node Actions.
The My Task node details are presented in three tabs as follows:
Config Tab
- Task Details: Click on Create Task to create a task. Once the task is created, the following task details are displayed and click on the Edit Task Settings to edit them.
- Name: It displays the name of the task. Click on the copy icon to copy the Task ID.
- Type: The type of the task, Labeling or QA task.
- Quality: It displays the quality task type details, if available.
- Owner: The avatar of the task owner. Click on the copy icon to copy the Owner email ID.
- Dataset: The dataset name of the task.
- Click on the dataset link to open it in the Data browser.
- Click on the copy icon to copy the Dataset ID.
- Recipe: The recipe name of the task.
- Click on the recipe link to open it in the recipe page.
- Click on the copy icon to copy the Recipe ID.
- Assignments: It displays the number of task assignees. Click on the link to view more details of the assignee, including email ID and Project Role.
- Node Settings: By default, the process to rerun items once their status changes are enabled. Click here for detailed review.
- Node Input: Delivered items will be added to the Task managed by this pipeline node. Click Set Parameter to set input parameter for the Task node. For more information, see the Node Inputs article.
- Node Output: Statuses will be set automatically according to Task Settings. Click on the plus icon to add the following parameters:
- Output name: The of the task output type.
- Output type: Task output type, whether it is a task or assignment.
- Descriptions (Optional): An optional description.
- Actions: By default, the completed action is displayed. It is system generated.
- Trigger (Optional): An Event/Cron Triggers can be set on this node, enabling you to initiate the pipeline run from this specific point.
For information on the Executions and Logs tabs, see the Node Details article.
Rerun on Status Change
A checkbox is available on the task node's Config tab to enable re-sending items as the task node outputs when their status becomes updated. As a result, the subsequent pipeline nodes will be re-executed with the same items.
This option is enabled by default and will be updated automatically to all existing pipelines.
For example;
If there are two tasks to complete; the first task has an issue and something has changed in its status, the process won’t go back. Every change in status from the task is a kind of trigger and will cause its running to reoccur.
For example, if there are two annotation tasks, such as:
- Annotation of a cat
- Annotation of a dog
The annotator inaccurately labeled the 'dog,' prompting the QA team to raise an issue. Consequently, the item's status will be changed to 'In progress' in the pipeline and returned to the annotator. The status will only transition to 'Completed' once the second task is finished. This procedure is also applicable when the annotator updates the status.
The code in the pipeline below will be executed only once, even if the item has been completed.
Checkpoint for Task/Assignment Completion
You have the option to create a temporary checkpoint between a Workflow node (task) and its next node, which will remain until the task or assignment managed by the node is completed. This will cause the execution of the subsequent node in the cycle to be put on hold until the completion event occurs.
To use this functionality, follow these steps:
- Add a Workflow node to the canvas and configure the task as desired.
- Create another output on the Workflow node you just added, specifying its type as either a task or an assignment (according to the checkpoint you wish to create).
- Connect the output of the Task/Assignment to any Task/Assignment input in the nodes you wish to delay the running of until the Task/Assignment is complete. You will see a dotted line indicating the connection between the nodes. Task/Assignment inputs can be set on Code and FaaS nodes.
- By setting the above, every cycle created in the pipeline will pause at the Workflow node, waiting for the completion of the Task/Assignment before executing the subsequent node. The cycle's status will remain as "In Progress".
- Keep in mind that this functionality alters the pipeline context from running on your data (items) to running on the task/assignment. The items can extracted from the task/assignment context.
Dotted Line Connection
The dotted line connecting the nodes also indicates that a new cycle will be initiated with a task/assignment context once the subsequent node is triggered by the completion event.
Checkpoint Example
In the following example, the initial node (workflow) receives items as inputs, a pipeline cycle will be generated for each item. Upon node execution, the item is passed to the task created by the node. The code snippet node is triggered only after all of the task items were given a status by the task assignees, and an event is fired. A new pipeline cycle is created to indicate this special flow with a new context.