Defining Jobs using FireTasks

In the installation tutorial, we ran a simple script that performed echo "howdy, your job launched successfully!" >> howdy.txt". Looking inside fw_test.yaml, you might have noticed that command defined within a task labeled Subprocess Task:

fw_id: -1
spec:
  _tasks:
  - _fw_name: Subprocess Task
    parameters:
      script: echo "howdy, your job launched successfully!" >> howdy.txt
      use_shell: true

The Subprocess Task is one type of FireTask, which is a predefined job template written in Python. The Subprocess Task is Python code that runs an arbitrary shell script that you give it, so you can use the SubProcess Task to run almost any job (without worrying that it’s all done within a Python layer). In this section, we’ll demonstrate how to use and define FireTasks.

Note

In this tutorial, we will run examples on the central server for simplicity. One could just as easily run them on a FireWorker if you’ve set one up.

Running multiple FireTasks

You can run multiple tasks within the same FireWork. For example, the first step of your FireWork might write an input file that the second step processes. Let’s create a FireWork where the first step prints howdy.txt, and the second step counts the number of words in that file.

1. Navigate to the tasks tutorial directory:

   cd <INSTALL_DIR>/fw_tutorials/firetask

2. Look inside the file fw_multi.yaml. You should see two Subprocess Task; the second one runs the wc -w command to count the number of characters in howdy.txt and exports the result to words.txt:

   fw_id: -1
   spec:
     _tasks:
     - _fw_name: Subprocess Task
       parameters:
         script: echo "howdy, your job launched successfully!" > howdy.txt
         use_shell: true
     - _fw_name: Subprocess Task
       parameters:
         script: wc -w < howdy.txt > words.txt
         use_shell: true
   launch_data: []

3. Run this multi-step FireWork on the central server:

   launchpad_run.py initialize <TODAY'S DATE>
   launchpad_run.py insert_single_fw fw_multi.yaml
   rocket_run.py

Tip

You can run all three of these commands on a single line by separating them with a semicolon. This will allow you to reset the database, insert a FW, and run it within a single command.

You should see two files written out to the system, howdy.txt and words.txt, confirming that you successfully ran a two-step job!

Using SubprocessTask

While running arbitrary shell scripts is nice, it’s not particularly well-organized. The command (echo), its arguments ("howdy, your job launched successfully!"), and its output (howdy.txt) are all intermingled within the same line. If we separated these components, it would be easier to do a data-parallel task where the same commands are run for multiple arguments. Let’s explore a better way to define our multi-step job:

1. Staying in the firetasks tutorial directory, remove any output from the previous step:

   rm howdy.txt fw.json

2. Look inside the file fw_better_multi.yaml. You should see two FireTasks as before. However, this time, the text we are printing is separated into its own echo_text parameter. We just need to change the value of this parameter in order to perform the same commands (echo and wc) on different input data. Note also that the names of the input and output files are also clearly separated from the commands themselves within the FireWork specification:

   fw_id: -1
   spec:
     _tasks:
     - _fw_name: Subprocess Task
       parameters:
         script: cat -t
         stdin_key: echo_text
         stdout_file: howdy.txt
     - _fw_name: Subprocess Task
       parameters:
         script: wc -w
         stdin_file: howdy.txt
         stdout_file: words.txt
     echo_text: howdy, your job launched successfully!
   launch_data: []

3. Run the FireWork on the central server to confirm that this new formulation also works as intended:

   launchpad_run.py initialize <TODAY'S DATE>
   launchpad_run.py insert_single_fw fw_better_multi.yaml
   rocket_run.py

At this point, you might want to change the echo_text parameter, reinsert the FireWork, and re-run the Rocket. Your custom text should get printed to howdy.txt and the number of words should change appropriately.

Creating a custom FireTask

Because the Subprocess Task can run arbitrary shell scripts, it can in theory run any type of job and is an ‘all-encompassing’ FireTask. However, if you are comfortable with some basic Python, it is better to define your own custom FireTasks (job templates) for the codes you run. A custom FireTask can clarify the usage of your code and guard against unintended behavior by restricting the commands that can be executed.

Even if you plan to only use Subprocess Task, we suggest that you still read through the next portion before continuing with the tutorial. We’ll be creating a custom FireTask that adds one or more numbers using Python’s sum() function, and later building workflows with this FireTask:

1. Navigate to the tasks tutorial directory and remove any output from the previous step:

   cd <INSTALL_DIR>/fw_tutorials/firetask
   rm howdy.txt fw.json

2. Look inside the file fw_adder.yaml for a new FireWork definition. This FireWork references a new FireTask, Addition Task, that adds the numbers 1 and 2:

   fw_id: -1
   spec:
     _tasks:
     - _fw_name: Addition Task
       parameters: {}
     input_array:
     - 1
     - 2
   launch_data: []

3. Look inside the file addition_task.py which defines the Addition Task:

   class AdderTask(FireTaskBase, FWSerializable):
   
      _fw_name = "Addition Task"
   
      def run_task(self, fw):
          input_array = fw.spec['input_array']
          m_sum = sum(input_array)
   
          with open('sum_output.txt', 'w') as f:
              f.write("The sum of {} is: {}".format(input_array, m_sum))
   

4. It should be clear how the Addition Task is set up:

   1. the reserved fw_name parameter is set to Addition Task, which is how FireWorks knows to use this code when an Addition Task is specified inside the fw_adder.yaml FireWork file.
   2. the run_task() method is the code that gets executed by the Rocket. In this case, we sum the values in the field called input_array, and write the output to sum_output.txt. In our fw_adder.yaml file, the input_array was set to 1 and 2.

   Note

   The main method in addition_task.py is not necessary to define a FireTask. However, it demonstrates how we created the fw_adder.yaml file.

4. Run the FireWork on the central server to confirm that the Addition Task works:

   launchpad_run.py initialize <TODAY'S DATE>
   launchpad_run.py insert_single_fw fw_adder.yaml
   rocket_run.py

Next up: Workflows!

With custom FireTasks, you can now go beyond running shell commands and execute arbitrary Python code templates. Furthermore, these templates can operate on dynamic input from the spec of the FireWork. For example, the Addition Task used the input_array from the spec to decide what numbers to add.

While one could construct an entire workflow by chaining together FireTasks within a single FireWork, this is often not ideal. For example, we might want to switch between different FireWorkers for different parts of the workflow depending on the computing requirements for each step. Or, we might have a restriction on walltime that necessitates breaking up the workflow into more atomic steps. Finally, we might want to employ complex branching logic or error-correction that would be cumbersome to employ within a single FireWork. The next step in the tutorial is to explore connecting together FireWorks into a true workflow.