Cache location

Every fetcher downloads its source files to a local cache and reuses them on subsequent calls. The cache root defaults to the OS-appropriate user cache directory (via pooch.os_cache):

OS	Default location
Linux	~/.cache/extracts/
macOS	~/Library/Caches/extracts/
Windows	%LOCALAPPDATA%\extracts\extracts\Cache\

Each dataset gets its own subdirectory under the root (e.g. ~/.cache/extracts/barrett2020/), and within that subdirectory extracts keeps both the raw downloaded file (e.g. table1.tsv) and the parquet-cached parsed DataFrame (e.g. table1.parquet).

Inspecting the current cache root

import extracts

extracts.get_location()

PosixPath('/home/runner/.cache/extracts')

Overriding the cache root

There are two ways to point extracts at a different directory:

1. Set the EXTRACTS_DATA_DIR environment variable before launching Python:

export EXTRACTS_DATA_DIR=/path/to/my/cache  # Linux/macOS
$env:EXTRACTS_DATA_DIR = "C:\path\to\my\cache"  # PowerShell

2. Call extracts.set_location(path) inside Python to set the env var for the current process:

import tempfile
from pathlib import Path

with tempfile.TemporaryDirectory() as tmp:
    extracts.set_location(tmp)
    print(extracts.get_location())

/tmp/tmpfslmvmpz

set_location expands ~ and resolves the path to absolute form.

The environment variable is read on every fetch, so changes take effect immediately — no need to re-import extracts.