Quickstart: Dewey Client

Overview

Dewey makes it easy to download large datasets reliably with multi-threaded downloads, robust error handling, and date partitioning.

Just create a download directory and you can start in seconds with a one-line command — no need to install anything into your Python environment:

uvx --from deweypy dewey --api-key <YOUR_API_KEY> speedy-download <FOLDER_ID>

📘

What is uvx?

uvx is part of uv, a modern package manager. It lets you run Python tools like dewey in a temporary isolated environment — no installs, no dependency conflicts. uv install instructions here.

Or, if you prefer to install the client locally for repeated use, follow the full workflow below. For advanced filtering on Parquet datasets, check out how to use DuckDB with Dewey.

Full Workflow

If you want to install the client and use it repeatedly, follow these steps:

1. Install the client

pip install deweypy

2. Create a download directory

mkdir dewey-downloads/

3. Locate FOLDER_ID

FOLDER_ID can be extracted from the end of the end of the API URL after data/.

Example: api.deweydata.io/api/v1/external/data/fldr_1234567 ➡️ FOLDER_ID = fldr_1234567

4. Run client

python -m deweypy --api-key <YOUR_API_KEY> speedy-download <FOLDER_ID>

A few notes...

• CLI first: Please use the CLI for downloads (notebooks will be supported soon).
• Downloads are multi-threaded: Increasing the number of workers for multi-threaded downloads yields diminishing returns, as API requests are limited both by our bucket’s rate limits and your own. We recommend the default of 8 workers, but you can override this with:

--num-workers <INT>

• Date partitioning: For partitioned datasets, filter by date with:

--partition-key-before YYYY-MM-DD --partition-key-after YYYY-MM-DD

• --partition-key-before includes all partitions up to and including the given date.
• --partition-key-after includes all partitions from and including the given date onward.
• You can use these flags whether you’re running the client with python -m deweypy or using the one-liner uvx command.

Working with data post-download

For guidance on analyzing your downloaded data, check out the provided notebook tutorial. It demonstrates:

• Reading Dewey data into Polars, Pandas, and DuckDB
• Filtering and transforming datasets
• Exporting to Parquet for efficient downstream workflows