Using the Vivo harvester from Python with Pyjnius

vivoharvester.py

"""
Use the Vivo harverster with Python and Jnius.

You need to set the Java classpath before running this. 
e.g.
    export CLASSPATH=".:/work/javalib/*"

The path you specify needs to contain the Vivo Harvester jar.  

https://github.com/vivo-project/VIVO-Harvester

"""

from jnius import autoclass


SDBJenaConnect = autoclass('org.vivoweb.harvester.util.repo.SDBJenaConnect')
SimpleSelector = autoclass('com.hp.hpl.jena.rdf.model.SimpleSelector')

class VivoConn(object):
    """
    A connection to Vivo SDB using the Vivo harvester.

    Pass in a dictionary with the connection details.

    {
     'db_url': 'vivo.school.edu/vivodb',
     'namespace': 'http://vivo.school.edu/individual/',
     'user': 'admin'
     'password': 'pass',
    }

    """

    def __init__(self, conf):
        db_url = conf.get('db_url')
        self.conn = SDBJenaConnect(
            'jdbc:mysql://%s' % db_url,
            conf.get('user'),
            conf.get('password'),
            conf.get('db_type', 'MySQL'),
            #class
            conf.get('db_driver', 'com.mysql.jdbc.Driver'),
            #layout
            conf.get('layout', 'layout2'),
            #modelname
            conf.get('layout', 'http://vitro.mannlib.cornell.edu/default/vitro-kb-2')
        )
        self.namespace = conf.get('namespace')

    def close(self):
        self.conn.close()

    def get_next_uri(self, max=9999999):
        """
        Get a random identifer to use as a resource uri.

        Interpretation of getNextURI here:
        http://svn.code.sf.net/p/vivo/vitro/code/branches/dev-sdb/webapp/src/edu/cornell/mannlib/vitro/webapp/utils/jena/JenaIngestUtils.java
        """
        import random
        model = self.conn.getJenaModel()
        next_uri = None
        if self.namespace is None:
            raise Exception('Namespace is None; set it.')
        while True:
            next_uri = self.namespace + 'n' + str(random.randint(1, max))
            resource = model.createResource(next_uri)
            #Check if this uri is the subject of any statements.
            sub_selector = SimpleSelector(resource, None, None)
            sub_iter = model.listStatements(sub_selector)
            if sub_iter.hasNext():
                sub_iter.close()
                continue
            else:
                #Check if this uri is the object of any statements.
                obj_selector = SimpleSelector(None, None, resource)
                obj_iter = model.listStatements(obj_selector)
                if obj_iter.hasNext():
                    obj_iter.close()
                    continue
                else:
                    break
        return next_uri


    def get_or_next(self, vtype, label):
        """
        Pass in a vivo type and a label to fetch the matching
        resource or return a next random identifier for creating
        resource.
        """
        query = """
        PREFIX vivo: <http://vivoweb.org/ontology/core#>
        PREFIX xsd:   <http://www.w3.org/2001/XMLSchema#>
        PREFIX rdfs:  <http://www.w3.org/2000/01/rdf-schema#>
        select ?uri
        where {
               ?uri a %(vtype)s.
               ?uri rdfs:label "%(label)s".
        }
        LIMIT 1
        """ % {'vtype': vtype, 'label': label}
        result_iter = self.conn.executeSelectQuery(query)
        while True:
            if result_iter.hasNext():
                next_result = result_iter.next()
                uri = next_result.get('?uri').toString()
                return (uri, False)
            else:
                break
        #If we are here, we need to generate a new uri.
        next_uri = self.get_next_uri()
        return (next_uri, True)