Using the Google APIs to Spell Check
by Jeremy Ehrmann
Overview
The Developers at Google have been kind enough to offer a web API for developers using
the SOAP protocol. When you do a search using Google, you may have noticed that you are
prompted with possible alternatives to any words you may have misspelled.
The Google web API "spell check" allows you to send a string of text and
receive alternatives for misspelled words. The power in this web API is that the
Google dictionary includes technology words that are used in website searches, but
may not have been included in a Standard English dictionary.
Setup
Setting up the Google SDK on your server is as simple as downloading the API from
http://www.google.com/apis/. You'll need to register with the website which will
give you your own key. You'll need the key for Google to accept SOAP connections from your server.
There are some limitations to be mentioned as well. The Google web API allows 10
words to be sent at a time and a limit of 1000 connections per key per day.
The following script works around the 10 word limit, however is still limited by the 1000 connections.
The Code
<%
'SEARCH ENGINE using GOOGLE API for RLC search engine
'Define Variables for Script
Dim objSoapClient, results
Dim googleAPIKey
Dim xmldocument
Dim xsldocument
Dim estimatedTotalResultsCount
Dim endIndex, startIndex
Dim searchQuery
Dim i 'as loop counter
'Instantiate the application variable for the GoogleAPI
application("googleSearchNumber") = application("googleSearchNumber") + 1
'Set the Google API Key you received when registering.
googleAPIKey = "00000000000000000000000000000000"
'Set the string to send to Google for spell checking
query = "The Lazy Bron Dog Jumpd Over The Fnce"
'Need to make an array with blocks of 10 words because the
'Google API will only accept the first 10 words of a string.
Dim arrTemp
Dim strTemp
Dim wordCount
Dim ArrayLength
Dim dicWords
i = 0
arrTemp = Split(query," ")
set dicWords = Server.CreateObject("Scripting.Dictionary")
wordCount = 1
'Google only allows up to 10 words at a time. Must make it 10 word strings.
If UBound(arrTemp) < 10 Then
dicWords.Add i, query
Else
For each Item in arrTemp
IF UBound(arrTemp) < 10 Then
dicWords.Add i, query
Else
If wordCount = 10 Then
dicWords.Add i, strTemp
wordCount = 1
i = i + 1
strTemp = Item & " "
Else
wordCount = wordCount + 1
strTemp = strTemp & Item & " "
End If
End If
Next
'Get any left overs
IF strTemp <> "" Then
dicWords.Add i, strTemp
'response.write "Aux: " & strTemp & ":" & i & "<BR>"
End If
End If
strTemp = ""
' set up a Soap Client object from MS Soap
set objSoapClient = Server.CreateObject("MSSOAP.SoapClient30")
' set up an XML document
set xmldocument = server.CreateObject("MSXML2.DOMDocument.4.0")
' set up an xsldocument
set xsldocument = server.CreateObject("MSXML2.DOMDocument.4.0")
' load the style sheet
'xsldocument.load(Server.MapPath("search.xsl"))
' initialise the SOAP object with the WSDL file from Google.
objSoapClient.mssoapinit Server.MapPath("GoogleSearch.wsdl")
'If you need to, set proxy settings here; Otherwise, comment out this line.
'objSoapClient.ConnectorProperty("ProxyServer") = "my.proxy.com:8000"
' Call the Google API search. This can only be done 1000 times a day
For i = 0 to UBound(dicWords.Items)
Results = objSoapClient.doSpellingSuggestion(googleAPIKey, dicWords.Item(i))
If Results = "" Then
strResults = strResults & dicWords.Item(i)
Else
strResults = strResults & Results
End If
Next
'Display the alternative String, if any.
response.write strResults
%>
|
Note: If the script gives you an error about not being able to create the
MSSOAP.SoapClient30 object (on line 66 or so) you may not have the
SOAP Toolkit 3.0
installed. It's a free download and doesn't even require a reboot. Everything else you need (WSDL file, etc.)
you'll find in the API you downloaded from http://www.google.com/apis/.
Conclusion
Although limitations exist in the Google web API, it makes a pretty handy spell check
for small sites or intranet applications. The dictionary and algorithm are all stored
off site so you won't need to update word lists. The best part is that technology words
are corrected as well, For Example: Google is an error in English, however the web API
sees this as a legitimate word.
Good Luck!
Jeremy (aka gigamanx)
Related Articles
- Working with Google APIs to Fetch Results using ASP