Developers

Documentation

Everything you need to integrate eomer.ai's forecasting engine into your data pipeline.

Getting Started

eomer-forecasting is a production-grade time series forecasting engine for production-grade time series forecasting. It supports local, parallel, and cloud execution backends.

Installation

Installbash

# Core install
pip install -e .

# With parallel execution support
pip install -e ".[parallel]"

# With S3 support
pip install -e ".[s3]"

# With API server
pip install -e ".[api]"

# Everything (dev included)
pip install -e ".[all]"

Run Your First Forecast

The fastest way to get started is with the CLI. Create a YAML config file (or use one of the examples) and run:

CLIbash

# Zero-shot univariate forecasting
eomer run configs/example_local_univariate.yaml

# Forecasting with covariates
eomer run configs/example_local_covariates.yaml

# With parallel execution
eomer run configs/example_parallel.yaml

# Validate config without running inference
eomer validate configs/example_local_univariate.yaml

Python API

Pythonpython

from eomer_forecasting.pipeline.runner import PipelineRunner

runner = PipelineRunner()
result = runner.run("configs/example_local_univariate.yaml")

print(f"Forecasted {result.forecasts[0].num_items} items")
print(f"Elapsed: {result.total_elapsed:.1f}s")

HTTP API

Start the API server and submit forecasting jobs via REST:

API Serverbash

# Start the server
uvicorn eomer_forecasting.api.app:app --host 0.0.0.0 --port 8000

# Submit a forecast
curl -X POST http://localhost:8000/forecast \
  -H "Content-Type: application/json" \
  -d '{"config_path": "configs/example_local_univariate.yaml"}'

Architecture

The pipeline follows a four-stage architecture with a pluggable runtime executor:

Pipeline Architecturetext

Ingestion ──> Preparation ──> Inference ──> Output
(CSV/XLSX/S3)  (Validate,      (eomer        (CSV/Parquet,
                Transform       forecasting   Local/S3)
                -> TSDF)        engine)
                                     │
                           ┌─────────┴─────────┐
                           │  Runtime Executor  │
                           │                    │
                           │ local │ parallel │ │
                           │ cloud              │
                           └────────────────────┘