skip to navigation
skip to content

pyoai 2.5.0

The oaipmh module is a Python implementation of an "Open Archives Initiative Protocol for Metadata Harvesting" (version 2) client and server. The protocol is described here:

.. -*- restructuredtext -*-


.. image::

The oaipmh module is a Python implementation of an "Open Archives
Initiative Protocol for Metadata Harvesting" (version 2) client and
server. The protocol is described here:

Below is a simple implementation of an OAIPMH client:

>>> from oaipmh.client import Client
>>> from oaipmh.metadata import MetadataRegistry, oai_dc_reader

>>> URL = ''

>>> registry = MetadataRegistry()
>>> registry.registerReader('oai_dc', oai_dc_reader)
>>> client = Client(URL, registry)

>>> for record in client.listRecords(metadataPrefix='oai_dc'):
>>> print record

The pyoai package also contains a generic server implementation of the
OAIPMH protocol, this is used as the foundation of the `MOAI Server Platform`_

.. _MOAI Server Platform:

2.5.1 (unreleased)

2.5.0 (2017-07-03)

- Added Python 3 compatibility (contributed by Tobias Kurze, Uli Köhler and Michał Pasternak)
- Travis support and badges (Michał Pasternak)

2.4.5 (2015-12-23)

- Added switch in client to force harvesting using HTTP Get method (contributed by Stefan Oderbolz).

- Added unofficial GetMetadata verb in server and client. GetMetadata is
identical to GetRecord, but only returns the first element below the
oai:metadata element, it does not return the oai enveloppe.

2.4.4 (2010-09-30)

* Changed contact info, Migrated code from Subversion to Mercurial

2.4.3 (2010-08-19)


* Convert lxml.etree._ElementUnicodeResult and ElementStringResult
to normal string and unicode objects, to prevent errors when these
objects get pickled. (lp #617439)

2.4.2 (2010-05-03)


* OAI_DC and DC namespace declarations should not be declared on the document
root, but on the child of the metadata element. According to the OAI spec

2.4.1 (2009-11-16)


* When specifying a date (not a datetime) for the until parameter,
default to 23:59:59 instead of 00:00:00

2.4 (2009-05-04)


* Included support for description elements in OAI Identify headers,
added 'toolkit' description by default.

2.3.1 (2009-04-24)


* Raise correct error when from and until parameters have different granularities

2.3 (2009-04-23)


* Fixed bug and added tests for handling invalid dateTime formats,
the server will now respond with a BadArgument (XML) error instead
of a python traceback.

* Use buildout to create testrunner and environment as opposed to
```` script.

Install buildout by:

$ python
$ bin/buildout

Run the tests by doing:

$ bin/test

To get a python interpreter with the ``oaipmh`` library importable::

$ bin/devpython

2.2.1 (2008-04-04)


* Added xml declaration to server output
* Prettyprint xml output
* compatibility fix: should be compatible with lxml 2.0 now
* server resumption tokens now work with POST requests.
* Fix for client code that handles 503 response from server.

2.2 (2006-11-20)


* Support for BatchingServer. A BatchingServer implements the
IBatchingOAI interface. This is very similar to IOAI, but methods
get a 'cursor' and 'batch_size' argument. This can be used to
efficiently implement batching OAI servers on top of relational

* Make it possible to explicitly pass None as the from or until parameters
for a OAIPMH client.

* an extra nsmap argument to Server and BatchingServer allows the
programmer to specify either namespace prefix to namespace URI mappings
that should be used in the server output.

* fixed a bug where the output wasn't encoded properly as UTF-8.

2.1.5 (2006-09-18)


* compatibility fix: it should work with lxml 1.1 now.

2.1.4 (2006-06-16)


* Distribute as an egg.



* Add infrastructure to deal with non-XML compliant OAI-PMH feeds;
an XMLSyntaxError is raised in that case.

* added tolerant_datestamp_to_datetime which is a bit more tolerant
than the normal datestamp_to_datetime when encountering bad

* Split off datestamp handling into separate datestamp module.



* Add support for day-only granularity (YYYY-MM-DD) in client.
calling 'updateGranularity' with the client will check with the
server (using identify()) to see what granularity the server
supports. If the server only supports day level granularity, the
client will make sure only YYYY-MM-DD timestamps are sent.



* Added framework for implementing OAI-PMH compliant servers.

* Changed package structure: now a oaipmh namespace package. Client
functionality now in oaipmh.client.

* Refactoring of module to reuse code for both client and

* Extended testing infrastructure.

* Switched over from using libxml2 Python wrappers to the lxml binding.

* Use generators instead of hacked up __getitem__. This means that the
return from listRecords, listIdentifiers and listSets are now not
normal lists but iterators. They can easily be turned into a normal
list by using list() on them, however.


Bugs fixed

* Typo in


Bugs fixed

* Added an encoding parameter to the serialize call, which fixes a
unicode bug.


Bugs fixed

* A harvest can return records with <header status~"deleted"=""> that
contain no metadata and are merely an indication that that
metadata-set for that resource is no longer on the OAI service.
These records should be used to remove metadata from the catalog if
it is there, bur should never be stored or catalogued
themselves. They aren't now. (Fixed in zope/OAICore/


Initial public release.  
File Type Py Version Uploaded on Size
pyoai-2.5.0.tar.gz (md5) Source 2017-07-03 104KB