28th April 2014 (2014-04-28)
- Description
- Author(s)
- Latest Version
- Documentation
- Requirements / Dependencies
- Installation
- Bugs, Feature requests etc.
- Copyright And Licensing
- Examples
A harvester to collect records from an OAI-PMH enabled provider.
The harvester can be used to carry out one-time harvesting of all records from a particular OAI-PMH provider by giving its base URL. It can also be used for selective harvesting, e.g. to harvest only records updated after, or before specified dates.
To assist in regular harvesting from one or more OAI-PMH providers, there's a provider registry. It is possible to associate a short memorable name for a provider with its base URLs, destination directory for harvested records, and the format (metadataPrefix) in which records should be harvested. The registry will also record the date and time of the most recent harvest, and automatically add this to subsequent requests in order to avoid repeatedly harvesting unmodified records.
This could be used in conjunction with a scheduler (e.g. CRON) to maintain a reasonably up-to-date copy of the record in one or more providers. Examples of how to accomplish these tasks are available below.
John Harrison <john.harrison@liv.ac.uk> at the University of Liverpool
The latest release version is available in the Python Packages Index:
https://pypi.python.org/pypi/oaiharvest
Source code is under version control and available from:
http://github.com/bloomonkey/oai-harvest
All executable commands are self documenting, i.e. you can get help on
how to use them with the -h or --help option.
At this time the only additional documentation that exists can be found in this README file!
pip install git+http://github.com/bloomonkey/oai-harvest.git#egg=oaiharvest
I recommend that you use virtualenv to isolate your development environment from system Python and any packages that may be installed there.
In GitHub, fork the repository
Clone your fork:
git clone git@github.com:<username>/oai-harvest.git
Install dependencies:
pip install -r requirements.txt
Install in develop / editable mode:
pip install -e .
Bug reports and feature requests can be submitted to the GitHub issue tracker: http://github.com/bloomonkey/oai-harvest/issues
If you'd like to contribute code, patches etc. please email the author, or submit a pull request on GitHub.
Copyright (c) University of Liverpool, 2013-2014
See LICENSE.rst for licensing details.
oai-harvest http://example.com/oai
oai-harvest --from 2013-01-01 http://example.com/oai
oai-harvest --set "some:set" http://example.com/oai
oai-harvest --limit 50 http://example.com/oai
oai-harvest --help
oai-reg add provider1 http://example.com/oai/1
If you don't supply --metadataPrefix and --directory options,
you will be interactively prompted to supply alternatives, or accept
the defaults.
oai-reg rm provider1 [provider2]
oai-reg list
You can harvest from one or more providers in the registry using the short names that they were registered with:
oai-harvest provider1 [provider2]
By default, this will harvest all records modified since the last
harvest from each provider. You can over-ride this behavior using the
--from and --until options.
You can also harvest from all providers in the registry:
oai-harvest all
In order to maintain a reasonably up-to-date copy of all the the records held by those providers, one could configure a scheduler to periodically harvest from all registered providers. e.g. to tell CRON to harvest all at 2am every day, one might add the following to crontab:
0 2 * * * oai-harvest all