A simple mechanism for storing a (potentially large) set of immutable documents in an organized way. Metadata is stored along with the data as a YAML file.

barn is inspired by OFS, BagIt and Pairtree. It replaces a previous project, docstash.

The easiest way of using barn is via PyPI:

$ pip install barn

Alternatively, check out the repository from GitHub and install it locally:

$ git clone https://github.com/pudo/barn.git
$ cd barn
$ python setup.py develop

barn manages Packages which are part of a Collection.

from barn import open_collection, Source

# open a collection of packages
collection = open_collection('file', path='/tmp')

# or via S3:
collection = open_collection('s3', aws_key_id='..', aws_secret='..',
                             bucket_name='test.pudo.org')

# import a file from the local working directory:
collection.ingest('README.md')

# import an http resource:
collection.ingest('http://pudo.org/index.html')
# ingest will also accept file objects and httplib/urllib/requests responses

# iterate through each document and set a metadata
# value:
for package in collection:
    for source in package.all(Source):
        with source.fh() as fh:
            source.meta['body_length'] = len(fh.read())
    package.save()

The code for this library is very compact, go check it out.

If AWS credentials are not supplied for an S3-based collection, the application will attempt to use the AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY environment variables. AWS_BUCKET_NAME is also supported.

barn is open source, licensed under a standard MIT license (included in this repository as LICENSE).