I'm excited to announce that we have made some significant changes to the way we are handeling stop searching for
GetStopsBySearch. I think these changes will be for the positive, but I wanted to explain them so as not to catch anyone off guard.
First, we have updated the algorithm that we are using to do searching on the back end. Previously we were using our own implementation of the
Damerau-Levenshtein Distance to compare all the words in a stop's name with all the words in the query. This worked well and allowed us to do fuzzy searching, but it was not very flexible or overly efficient. With this update, we have switched to
Lucene.net, a popular (and powerful) searching library. This allows us more flexibility and we are able to do some pretty cool things with it. For example, if there are multiple stops that match your query equally well, we now return them in order of popularity.
Another exciting feature of this searching update is our, completely new,
Auto Completing API. This API is not a part of our existing Developer API and does not require an API key. It is designed to be lightweight and fast. In addition to not requiring an API key, it differs from our regular searching API in that it does not return full stop objects. It simply returns the matched stop_id, stop_name, and sms_code. Additionally, we are not providing a WS endpoint for the Auto Completing API only a RESTful JSON and XML interface.
This is exactly what we are using to do stop searching on our website. So, if you would like to see what it can do, head over to
www.cumtd.com. As always, if you have any questions or feedback regarding these updates, feel free to post them here or email
develoepr@cumtd.com.
Edited by user Wednesday, May 24, 2017 10:56:18 AM(UTC)
| Reason: Not specified