Implementation:Ucbepic Docetl MOAR CliHelpers

Knowledge Sources	Ucbepic_Docetl
Domains	Data_Processing, Optimization, CLI
Last Updated	2026-02-08 00:00 GMT

Overview

Concrete tool for CLI helper functions that launch and configure MOAR optimizer runs provided by DocETL.

Description

The cli_helpers module provides functions to infer dataset information from YAML pipeline configs, load custom evaluation functions, and orchestrate a complete MOAR optimization run from the command line. It handles parameter extraction from optimizer_config sections, resolves relative file paths, validates required configuration fields, and initializes MOARSearch with the correct parameters before running the search.

Usage

Use these helpers when running the MOAR optimizer from the CLI via docetl optimize or when programmatically launching MOAR optimization from a YAML pipeline configuration that includes an optimizer_config section.

Code Reference

Source Location

• Repository: Ucbepic_Docetl
• File: docetl/moar/cli_helpers.py
• Lines: 1-305

Signature

def infer_dataset_info(yaml_path: str, config: dict) -> tuple[str, str]: ...

def load_evaluation_function(config: dict, dataset_file_path: str) -> callable: ...

def run_moar_optimization(
    yaml_path: str,
    optimizer_config: dict,
) -> Dict[str, Any]: ...

Import

from docetl.moar.cli_helpers import infer_dataset_info, load_evaluation_function, run_moar_optimization

I/O Contract

Inputs

Name	Type	Required	Description
yaml_path	str	Yes	Path to the YAML pipeline configuration file
config	dict	Yes	Full YAML config dictionary (for infer_dataset_info)
optimizer_config	dict	Yes	Dictionary from the YAML optimizer_config section containing save_dir, available_models, evaluation_file, metric_key, max_iterations
dataset_file_path	str	Yes	Path to the dataset file (for load_evaluation_function)

Outputs

Name	Type	Description
dataset_path	str	Resolved absolute path to the dataset file
dataset_name	str	Name of the dataset from the YAML config
evaluate_func	callable	Loaded evaluation function decorated with @docetl.register_eval
experiment_summary	Dict[str, Any]	Summary dictionary with optimization results, costs, and timings

Usage Examples

from docetl.moar.cli_helpers import run_moar_optimization
import yaml

# Load a pipeline YAML with optimizer_config
with open("my_pipeline.yaml", "r") as f:
    config = yaml.safe_load(f)

optimizer_config = config.get("optimizer_config", {})

# Run MOAR optimization
results = run_moar_optimization(
    yaml_path="my_pipeline.yaml",
    optimizer_config=optimizer_config,
)

print(f"Best pipeline cost: {results['best_cost']}")

Related Pages

• Environment:Ucbepic_Docetl_Python_Runtime