OARepo OAI-PMH harvester

An OAI-PMH harvesing library for Invenio 3.5+. The library provides initial transformation of OAI-PMH payload to an intermediary json representation which is later on transformed by a specific transformer to the format of invenio records.

Due to their generic nature, these transformers are not part of this library but have to be provided by an application.

The progress and transformation errors are captured within the database.

For now, the library does not provide error notifications, but these will be added. Sentry might be used for the logging & reporting.

Installation

pip install oarepo-oaipmh-harvester
pip install <your transformer library>

Configuration

All configuration is inside the OAIHarvesterRecord record. There is a command-line tool to add a new record:

invenio oarepo oai harvester add nusl \
    --name "NUSL harvester" \
    --url http://invenio.nusl.cz/oai2d/ \
    --set global \
    --prefix marcxml \
    --loader sickle \
    --transformer marcxml \
    --transformer nusl \
    --writer 'service{service=nr_documents}'

This will register an oai-pmh harvester with code "nusl", its url, oai set and metadata prefix. Records from this harvester will be loaded with the sickle loader (default loader if not specified) and at first transformed from marcxml to json format and subsequently by NUSL transformer to get nr_documents compatible json.

The json is then used by service writer to create/update the target record.

Usage

Command-line

On command line, invoke

invenio oarepo oai harvester run nusl <optional list of oai identifiers to harvest>

Options:

  --all-records      Re-harvest all records, not from the last timestamp
  --on-background    Run harvest on background (via celery task)
  --identifier       Harvest the passed identifier/s

You can also pass arguments for "havester add" to override the defaults from the configuration.

Harvest status

Harvester uses 3 additional types of record:

• Run represents a single run of the harvester
• Batch - the run consists of multiple batches, that is records, that are processed together. Each such batch is represented in the batch record
• OAI record is a link between harvested record and its original. If there are errors harvesting the record, it is still created and its errors field is filled with the error

Custom parsers and transformers

Transformer

A piece of code that gets a StreamEntry (or a StreamBatch) instance, processes it and returns modified StreamEntry. An example is a MarcXML transformer that takes the string with xml representation of the entry and transforms it into simple json representation {abcxy: value(s)}, where abcxy is marc field code.

See oarepo_runtime.datastream.transformers package for StreamEntry/StreamBatch interfaces.

The transformer needs to be registered:

# mypkg.transformers

from .impl import MyTransformer

my_transformer = {"class": MyTransformer, "params": {
       # default parameters that will go to the MyTransformer constructor
}}

And setup.cfg:

# setup.cfg

[options.entry_points]
oarepo.oaipmh.transformers =
    my_transformer = mypkg.transformers:my_transformer

Then you can use my_transformer when creating your harvester.

Reader

A reader is responsible for fetching records and creating a stream of StreamEntry items. See oarepo_runtime.datastreams.readers for details. Then register the reader into oarepo.oaipmh.readers entry point with the same syntax as above.