The ISFDB currently provides 2 web APIs: getpub.cgi which is used to obtain publication metadata, and submission.cgi which is used to submit data to the ISFDB. There are currently no methods for obtaining author, title, series, or awards information from the ISFDB.
Since registration is required to edit the ISFDB, the web API enforces that policy. Submitting data to the ISFDB requires a registered user ID, as well as a license key which is your password for accessing the web API. If you are logged into the ISFDB, a license key can be generated by visiting http://www.isfdb.org/cgi-bin/edit/keygen.cgi and clicking on [Generate New Key]. For data submissions this key must appear within a LicenseKey XML tag following the Submitter tag.
The getpub.cgi application takes an ISBN for an argument, and returns an XML structure containing the metadata for that particular publication. If more than one publication is associated with that particular ISBN, multiple records will be generated. The Records tag indicates the number of records found, followed by some number of Publication records. A license key is not required to use this API. The URL for getpub.cgi should appear as follows:
Example XML returned via call to getpub.cgi:
<?xml version="1.0" encoding="iso-8859-1" ?> <ISFDB> <Records>1</Records> <Publications> <Publication> <Record>325837</Record> <Title>The Winchester Horror</Title> <Authors> <Author>William F. Nolan</Author> </Authors> <Year>1998-00-00</Year> <Isbn>1881475530</Isbn> <Publisher>Cemetery Dance Publications</Publisher> <PubSeries>Cemetery Dance Novella</PubSeries> <PubSeriesNum>6</PubSeriesNum> <Price>$30.00</Price> <Pages>111</Pages> <Binding>hc</Binding> <Type>CHAPTERBOOK</Type> <Tag>THWNCHSTRH1998</Tag> <CoverArtists> <Artist>Eric Powell</Artist> </CoverArtists> <Note>Hardcover Limited Edition of 450 signed and numbered copies bound in full-cloth and Smyth sewn</Note> </Publication> </Publications> </ISFDB>
Example python script utilizing getpub.cgi
Although getpub.cgi can be invoked via a browser, a more practical application is to invoke the URL via a programming interface, parsing the result with an XML parser, and then doing something of use with the resulting data. The following example shows the python code needed to obtain the XML data; it does not show an example of parsing the XML:
import httplib host = "www.isfdb.org" def GetXml(isbn): webservice = httplib.HTTP(host) command = '/cgi-bin/rest/getpub.cgi?%s' % isbn webservice.putrequest("GET", command) webservice.putheader("Host", host) webservice.putheader("User-Agent", "Wget/1.9+cvs-stable (Red Hat modified)") webservice.endheaders() errcode, errmsg, headers = webservice.getreply() if errcode != 200: resp = webservice.getfile() print "Error:", errmsg print "Resp:", resp.read() resp.close() return '' else: resp = webservice.getfile() raw = resp.read() resp.close() index = raw.find('<?xml') return raw[index:]
In the event of an error condition, getpub.cgi will return an XML structure with the number of records found set to zero:
<?xml version="1.0" encoding="iso-8859-1" ?> <ISFDB> <Records>0</Records> <Publications> </Publications> </ISFDB>
The submission.cgi application allows a remote user to create ISFDB submissions. Note that the use of submission.cgi requires prior authorization by the ISFDB administrator. If you would like to be added to the list of authorized users, post a request on ISFDB:Moderator noticeboard.
Any of the data payloads defined in the Data Submission Formats article are accepted. The format of the XML data is identical to that used for local submissions, with the exception that the user's license key must be inserted after the Submitter tag. The user's registered login name should be contained within the Submitter tag, and the user's current license key should be contained with the LicenseKey tag. The URL for submission.cgi should appear as follows:
Submissions are placed in the moderator queue as with an ISFDB edit, and are not integrated into the database until a moderator approves it. As such it is important that you not flood the moderator queue with hundreds of submissions. In general you should only post no more than 20 submissions per session, until they are cleared out by a moderator.
An example submission XML payload follows, with the LicenseKey x'd out:
<?xml version="1.0" encoding="iso-8859-1" ?> <IsfdbSubmission> <NewPub> <Submitter>Dissembler</Submitter> <LicenseKey>xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx</LicenseKey> <Subject>Lady of the Glen: A Novel of 17th-Century Scotland and the Massacre of Glencoe</Subject> <Isbn>1575661292</Isbn> <Title>Lady of the Glen: A Novel of 17th-Century Scotland and the Massacre of Glencoe</Title> <Year>1997-01-01</Year> <Price>$14.95</Price> <Language>English</Language> <Publisher>Kensington</Publisher> <PubSeries>Time Romances</PubSeries> <PubSeriesNum>17</PubSeriesNum> <Pages>432</Pages> <Authors> <Author>Jennifer Roberson</Author> </Authors> <Binding>tp</Binding> <PubType>NOVEL</PubType> <Image>http://ecx.images-amazon.com/images/I/519xedh9nrL.jpg</Image> <Note>Data from Amazon.com as of 2008-05-08</Note> </NewPub> <IsfdbSubmission>
Example python script utilizing submission.cgi
The submission.cgi application uses the http POST method, so it can not be invoked via a web browser. Once a proper XML payload has been constructed as described above, a connection needs to be made to the ISFDB web server, and the appropriate http header needs to be sent, including content type and length information. This example below shows how this can be implemented in python:
import httplib host = "www.isfdb.org" def SendXml(xmlmsg): webservice = httplib.HTTP(host) webservice.putrequest("POST", "/cgi-bin/rest/submission.cgi") webservice.putheader("Host", host) webservice.putheader("Content-type", "text/xml; charset=\"UTF-8\"") webservice.putheader("Content-length", "%d" % len(xmlmsg)) webservice.endheaders() webservice.send(xmlmsg) errcode, errmsg, headers = webservice.getreply() if errcode != 200: resp = webservice.getfile() print "Error:", errmsg else: resp = webservice.getfile() resp.close()
In the event of an error condition, submission.cgi will return an XML structure with a Status of FAIL (success will return OK). The cause of the error can be found within the Error tag:
<?xml version="1.0" encoding="iso-8859-1" ?> <ISFDB> <Status>FAIL</Status> <Error>Bad XML data</Error> </ISFDB>