Skip to content

Configure a task in Python

A Task is the basic building block in Otaro. Similar to DSPy, Otaro treats LLM calls like functions with inputs and outputs.

Quickstart

quickstart.py
from otaro import Field, Task

# Define a task that takes a topic and returns a list of quotes
task = Task(
    model="gemini/gemini-2.0-flash-001",
    inputs=[
        "topic", # same as Field(name="topic", type=str)
    ],
    outputs=[
        Field(
            name="quotes",
            type=list[str],
            # type takes native Python types or Pydantic BaseModel classes
        )
    ],
)

response = task.run(topic="discworld")
for quote in response.quotes:
    print(quote)
uv run quickstart.py
The pen is mightier than the sword if the sword is very short, and the pen is very sharp.
In ancient times cats were worshipped as gods; they have not forgotten this.
Wisdom comes from experience. Experience is often a result of lack of wisdom.
Words in the heart cannot be taken.

Defining a task

At minimum, a Task needs to be initialized with:

  • model (str): A model name, used by litellm under the hood
  • inputs (list[str | Field]): A list of inputs that the task is expected to receive
  • outputs (list[str | Field]): A list of outputs to be produced

Fields

A Field is used to specify the type of output that the LLM is supposed to generate.

The Field definitions will also be used to parse the LLM response and return an output with attributes of the correct types.

For instance, in the above example, response.quotes will be of type list[str].

For convenience, a Field can also be defined with a single string, which will be interpreted as the field name and default to type=str. For example, inputs is set as ["topic"] in quickstart.py above.

Adding rules

Instead of trying to figure out the best prompt, define the desired output with rules. Then, whenever task.run is called, the task will be automatically optimized to enforce the rules where possible.

For illustration, rewrite the task above to enforce 3 quotes:

three_quotes.py
from otaro import Field, Task

# Create a task that takes a topic and returns a list of quotes
task = Task(
    model="gemini/gemini-2.0-flash-001",
    inputs=["topic"],
    outputs=[
        Field(
            name="quotes",
            type=list[str],
        )
    ],
    # Add rule to enforce len(quotes) == 3
    rules=[
        "otaro.rules.length_eq(quotes, 3)",
    ]
)

response = task.run(topic="discworld")
for quote in response.quotes:
    print(quote)

Running the script will automatically optimize the task to output three quotes.

Note: The optimized config will only be saved for future use if the task was loaded via YAML. See Configure a task in YAML for more details.

uv run three_quotes.py
INFO:otaro.task:Evaluating rulesINFO:otaro.task:Rule 1: otaro.rules.length_eq(quotes, 3) - False
INFO:otaro.task:Optimizing task...
INFO:otaro.task:Evaluating prompt: "Write three quotes about {topic}."
INFO:otaro.task:Evaluating rulesINFO:otaro.task:Rule 1: otaro.rules.length_eq(quotes, 3) - True
INFO:otaro.task: Score: 1.0
INFO:otaro.task:All scores: [0.0, 1.0]
INFO:otaro.task:Selecting prompt #1 with score 1.0

The pen is mightier than the sword if the sword is very short, and the pen is very sharp.
In ancient times cats were worshipped as gods; they have not forgotten this.
Wisdom comes from experience. Experience is often a result of lack of wisdom.