Skip to main content

snowdev: DevOps toolkit for Snowflake, facilitating seamless deployment of UDFs, stored procedures, and Streamlit apps using Snowpark's capabilities right from your local environment.

Project description

SnowDev - Snowpark Devops

Documentation Downloads

SnowDev is a command-line utility designed for deploying various components related to Snowflake such as UDFs, stored procedures, and Streamlit applications using Snowpark. This tool streamlines tasks like initializing directories, local testing, uploading, and auto create components code using AI.

Features

SnowDev currently supports the following components:

  • UDF (User-Defined Function)
  • Stored Procedure
  • Streamlit
  • Tasks

Setup

 pyenv install 3.10.0 
 pyenv local 3.10.0 
 pip install snowdev or poetry add snowdev
 snowdev init

Usage

snowdev <command> [options]

Commands

init

  • Description: Initializes the directory structure for the deployment.
  • Usage: snowdev init

test

  • Description: Test the deployment locally.
  • Usage: snowdev test --udf predict_sentiment

deploy

  • Description: Deploys the specified components, registers and tests using temp function before deploying to prod
  • Usage: snowdev deploy --udf predict_sentiment

upload

  • Description: Uploads specified items such as static content.
  • Usage: snowdev upload --upload <upload_item>

add

  • Description: Adds a package and optionally uploads it to stage.
  • Usage: snowdev add --package <package_name>

new

  • Description: Adds a new component.
  • Usage: snowdev new --sproc "test_script"

ai

  • Description: Interact with AI components and embeddings. It can also help in creating new AI components code based on the description, make sure to have executed snowdev ai --embed to generate embeddings.
  • Usage: snowdev ai --streamlit "Want to see a bar chart on the order table"

Options

  • --udf <udf_name>: Name or identifier for the UDF you want to deploy.
  • --sproc <sproc_name>: Name or identifier for the Stored Procedure you want to deploy.
  • --streamlit <streamlit_name>: Name or identifier for the Streamlit application you want to deploy. (This is still in PrPr)
  • --task <task_name>: Name of the snowflake task you want to deploy.
  • --upload <upload_item>: Specifies what to upload. Currently supported options: static.
  • --package <package_name>: Specifies the name of the package to zip and upload to the static folder.
  • --embed: Used with the ai command to run embeddings.

Requirements

  • Python: >=3.10.0, <3.11.0
  • Dependencies:
    • "snowflake-snowpark-python" = { version = "1.5.1", extras = ["pandas"] }

Notes

  • For the ai command, when specifying a component type (using --udf, --sproc, or --streamlit), ensure that the respective name or identifier is provided.
  • When using the add command, the user will be prompted to decide if they want to upload the zip package to stage.
  • The AI functionality in SnowDev is optimized with GPT-4, ensuring a better relevance in code suggestions and interactions.

Roadmap

  • Support for UDFs and Stored Procedures
  • Support for Streamlit
  • AI interactions for embedding and suggestions
  • Support for snowflake Tasks
  • AI to add the packages to toml file
  • Support for multiple python scripts in stored procedures
  • Use AI to modify existing code for optimization
  • Adding more granularity for AI commands

Contributions

Feel free to contribute to SnowDev by submitting pull requests or opening issues on the project's GitHub repository.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

snowdev-0.2.0.tar.gz (34.8 kB view hashes)

Uploaded source

Built Distribution

snowdev-0.2.0-py3-none-any.whl (44.8 kB view hashes)

Uploaded py3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page