Quick Start
PlaygroundData Catalogue

[Legacy] Submission

Finter Model Submission Helper Guide

This comprehensive guide is designed to streamline the process of preparing and submitting your Alpha Model (AM) through the use of the NotebookSubmissionHelper provided by the finter package. By leveraging this helper, you can concentrate on refining your model while the submission nuances are efficiently handled in the background.

Import NotebookSubmissionHelper

First, import the NotebookSubmissionHelper class from the finter.framework_model module

from finter import NotebookSubmissionHelper

Configure Your Model

It's crucial to detail your model's specifications. Specify the notebook's name, your model's name, its target universe, and type.

# Name of your Jupyter notebook
notebook_name = "dummy"

# Name of your model
model_name = "alpha/dummy"

# Universe your model is targeting
model_universe = "kr_stock"

# Type of model you are developing
model_type = "alpha"

# Benchmark for your model. Set to None for the default benchmark, or False for no benchmark.
benchmark = None

Submission

Create an instance of NotebookSubmissionHelper with the configuration you've defined.

processor = NotebookSubmissionHelper(
    notebook_name, model_name, model_universe, model_type, benchmark
)

Once you have completed these steps, your model is ready for submission. Below are additional features you might find useful:

Processing Your Model

To execute your model's code, run simulations, validate outputs, and prepare for submission, use the process method with appropriate parameters.

# Define the start and end dates for your model's evaluation
start = 20200101  # Start date in YYYYMMDD format
end = 20220101    # End date in YYYYMMDD format

# Process the model
processor.process(start, end, position=True, simulation=True, validation=True, submit=False, git=True, docker_submit=True)

The process method accepts several parameters:

  • start: The evaluation period's start date.

  • end: The evaluation period's end date.

  • position: Processes position data if set to True.

  • simulation: Runs a model simulation if True.

  • validation: Validates the model's output if True.

  • submit: Submits the model to Finter. Set to False for a review before submission.

  • git: Synchronizes your model with GitHub if True, facilitating version control and collaboration.

  • docker_submit: If set to True, the model will be submitted using the Docker submit. If set to False, the model will be submitted using the legacy submit.

Docker submit

The Docker submit allows you to build and submit models using your preferred Python development environment through Docker, resolving the fixed Python environment issue that was present in the legacy submit process.

The Python environment must be specified using Poetry, and the submit folder must include the pyproject.toml and poetry.lock files. If you are working from the Jupyter files in the Home directory of the Quantit Jupyterlab environment, these files will be automatically generated. Alternatively, you can create the Poetry files manually.

For more details about Poetry, please refer to this link: https://python-poetry.org/docs.

Review, Update, and Submit

After processing, thoroughly review the output at each step. Once you're satisfied and ready to submit, set the submit parameter to True and rerun the process method to finalize your submission.

# Final submission after review and updates
processor.process(start, end, position=True, simulation=True, validation=True, submit=True, git=True)

By following this guide, you can efficiently prepare and submit your model to Finter, ensuring a smooth transition through necessary checks and simulations. The added GitHub synchronization feature enhances collaboration and version control, making your model development process more effective and organized.

Last updated