Premium Web of Knowledge API python access

Vote count:

0

My university bought the premium Web Services of Thompson-Reuters (Web of Knowledge, a.k.a Web of Sciences, I've uploaded the latest guide book on Google Drive) and I tried to retrieve some publication information. I found the python script by domoritz and it's for Web Services Lite. I modified it a little bit:

#!/usr/bin/env python
# -*- coding: utf-8 -*-

from suds.client import Client
from suds.transport.http import HttpTransport
import urllib2

class HTTPSudsPreprocessor(urllib2.BaseHandler):
    def __init__(self, SID):
        self.SID = SID

    def http_request(self, req):
        req.add_header('cookie', 'SID="'+self.SID+'"')
        return req

    https_request = http_request


class WokmwsSoapClient():
    """
    main steps you have to do:
        soap = WokmwsSoapClient()
        results = soap.search(...)
    """
    def __init__(self):
        self.url = self.client = {}
        self.SID = ''

        self.url['auth'] = 'http://ift.tt/1tZgd5O'
        self.url['search'] = 'http://ift.tt/1qCUUrs'

        self.prepare()

    def __del__(self):
        self.close()

    def prepare(self):
        """does all the initialization we need for a request"""
        self.initAuthClient()
        self.authenticate()
        self.initSearchClient()

    def initAuthClient(self):
        self.client['auth'] = Client(self.url['auth'])

    def initSearchClient(self):
        http = HttpTransport()
        opener = urllib2.build_opener(HTTPSudsPreprocessor(self.SID))
        http.urlopener = opener
        self.client['search'] = Client(self.url['search'], transport = http)

    def authenticate(self):
        self.SID = self.client['auth'].service.authenticate()

    def close(self):
        self.client['auth'].service.closeSession()

    def search(self, query, chkpt):
        qparams = {
            'databaseId' : 'WOS',
            'userQuery' : query,
            'queryLanguage' : 'en',
            'editions' : [{
                'collection' : 'WOS',
                'edition' : 'SCI',
            },{
                'collection' : 'WOS',
                'edition' : 'SSCI',
            }]
        }

        rparams = {
            'count' : 100, # 1-100
            'firstRecord' : chkpt,
            'sortField' : [{
                'name' : 'RS',
                'sort' : 'D',
            }]
        }

        return self.client['search'].service.search(qparams, rparams)

    def citedReferences(self, query, chkpt):
        qparams = {
            'databaseId' : 'WOS',
            'uid' : query,
            'queryLanguage' : 'en'
        }

        rparams = {
            'count' : 100, # 1-100
            'firstRecord' : chkpt
        }

        return self.client['search'].service.citedReferences(qparams, rparams)

The search function for abstract info is fine but the citedReferences can't like suds.TypeNotFound: 'databaseId'. I'm sure databaseId is one of the parameters (page 12 in the guide).

<soapenv:Envelope xmlns:soapenv="http://ift.tt/sVJIaE"
       xmlns:woksearch="http://ift.tt/1xxc4tr">
       <soapenv:Header/>
       <soapenv:Body>
          <woksearch:citedReferences>
             <databaseId>WOS</databaseId>
             <uid>WOS:000270372400005</uid>
             <queryLanguage>en</queryLanguage>
             <retrieveParameters>
                <firstRecord>1</firstRecord>
                <count>100</count>
                <option>
                  <key>Hot</key>
                  <value>On</value>
                </option>
             </retrieveParameters>
          </woksearch:citedReferences>
       </soapenv:Body>
    </soapenv:Envelope>

The official example is a little different from the one for search (page 21):

<soapenv:Envelope xmlns:soapenv="http://ift.tt/sVJIaE"
       xmlns:woksearch="http://ift.tt/1xxc4tr">
       <soapenv:Header/>
       <soapenv:Body>
          <woksearch:search>
             <queryParameters>
                <databaseId>WOS</databaseId>
                <userQuery>TS=(cadmium OR lead)</userQuery>
                <editions>
                   <collection>WOS</collection>
                   <edition>SCI</edition>
                </editions>
                <timeSpan>
                   <begin>2000-01-01</begin>
                   <end>2011-12-31</end>
                </timeSpan>
                <queryLanguage>en</queryLanguage>
             </queryParameters>
             <retrieveParameters>
                <firstRecord>1</firstRecord>
                <count>5</count>
            <option>
                   <key>RecordIDs</key>
                   <value>On</value>
                </option>
                <option>
                   <key>targetNamespace</key>
                   <value>http://ift.tt/1q6EGJr;
                </option>
            </retrieveParameters>
          </woksearch:search>
       </soapenv:Body>
</soapenv:Envelope>

In citedReferences there's no queryParameters, all databaseId, uid, queryLanguage are parallel variables. I tried like this:

def citedReferences(self, query, chkpt):
        databaseId = 'WOS'
        uid = query
        queryLanguage = 'en'

        rparams = {
            'count' : 100, # 1-100
            'firstRecord' : chkpt
        }

        return self.client['search'].service.citedReferences(uid, rparams)

It returned suds.TypeNotFound: 'count'. Is there any line I should add/modify to make the script work? Thanks!

asked 2 mins ago

leoce

108

Premium Web of Knowledge API python access