AppVeyor: CircleCI:

Coveralls: Codecov:

This is a Python3 tool that can mass-download files from Zenodo records.

The code is hosted at Github.

It is recommended to use uv for managing Python environments and installing this package. zenodo-get requires Python 3.10 or newer.

0. The most simple way is to use it as a tool, no installation is needed:

   uv tool run zenodo_get RECORD_ID_OR_DOI

1. Install uv (if you haven't already):

   # On macOS and Linux
   curl -LsSf https://astral.sh/uv/install.sh | sh
   # On Windows
   powershell -c "irm https://astral.sh/uv/install.ps1 | iex"

2. Create a virtual environment and install zenodo-get:

   • From PyPI:

     uv venv
     uv pip install zenodo-get
     source .venv/bin/activate # Or .venv\Scripts\activate on Windows

   • Or from a local source checkout:

     uv venv
     uv pip install .
     source .venv/bin/activate # Or .venv\Scripts\activate on Windows

Traditional pip installation is also supported:

pip install zenodo-get # Ensure pip is for Python 3.10+

Afterwards, you can query the command line options:

zenodo_get -h

but the default settings should work for most use cases:

zenodo_get RECORD_ID_OR_DOI

As stated above, the most simple way is to use uv tool run or uvx. If you don't use these, but you use uv: Once your project is set up with uv (either by installing dependencies via uv pip install . or by just having the pyproject.toml present), you can use uv run to execute the zenodo_get command directly without needing to activate the virtual environment in your current shell:

# Example: Show help message
uv run zenodo_get --help

# Example: Download a record (replace YOUR_RECORD_ID)
uv run zenodo_get YOUR_RECORD_ID -o output_directory

# Example: Using a script defined in pyproject.toml (zenodo_get is defined there)
# uv run zenodo_get YOUR_RECORD_ID

Pip and pipx also work.

The tool itself is simple, and the help message is reasonable:

zenodo_get -h

but if you need more, open a github ticket and explain what is missing.

Basic usage:

zenodo_get RECORD_ID_OR_DOI

You can use the -g or --glob option to specify file patterns. To download multiple specific file types, provide a comma-separated list of glob patterns:

zenodo_get RECORD_ID_OR_DOI -g "*.txt,*.pdf,images/*.png"

Other Special parameters:

• -m : generate md5sums.txt for verification. Beware, if md5sums.txt is present in the dataset, it will overwrite this generated file. Verification example: md5sum -c md5sums.txt
• -w FILE : instead of downloading the record files, it will generate a FILE (or print to stdout if FILE is -) which contains direct links to the Zenodo site. These links could be downloaded with any download manager, e.g. with wget: wget -i urls.txt
• -e : continue on error. It will skip the files with errors, but it will try to download the rest of the files.
• -k : keep files: it will keep files with invalid md5 checksum. The main purpose is debugging.
• -R N: retry on error N times.
• -p N: Waiting time in sec before retry attempt. Default: 0.5 sec.
• -n : do not continue. The default behaviour is to download only the files which are not yet download or where the checksum does not match with the file. This flag disables this feature, and it will force download existing files, and assigning a new name to the files (e.g. file(1).ext )

Remark for batch processing: the program always exits with non-zero exit code, if any error has happened, for instance, checksum mismatch, download error, time-out, etc. Only perfectly correct downloads end with 0 exit code.

You don't really need to cite this software, except if you use it for another academic publication. E.g. if you download something from Zenodo with zenodo-get: no need to cite anything. If you download a lot from Zenodo, and you publish about Zenodo, and my tool is integral part of the methodology, then you could cite it. You could always ask the code to print the most up-to-date reference producing plain text and bibtex references too:

zenodo_get --cite