Presto, Tomcat and Web Services

bniknar
User offline. Last seen 2 years 6 weeks ago. Offline
Joined: 12/21/2009
Points: 10

Hello,

I want to build a proof-of-concept with Presto. The goal: a web service which, when called with SoapUI, logs into a Intranet server with username and password, and scrapes data (name and phone number) from a web page. I hope to prove that:

-- Presto can host the web service itself

-- There is a simple GUI tool that can be used to setup the web service: point it to the Intranet server, "show" it the web page to scape, provide username and password for access

My question: what JackBe components to I need to download to set this up? Also: what are best practices for making this successful? If I can prove this out, I will look to Presto as an enterprise solution for scraping many different large Intranet websites.

Regards, Brian

 

 

  •  
0
Your rating: None
girish
girish's picture
User offline. Last seen 4 weeks 3 hours ago. Offline
Joined: 09/22/2008
Points: 70

Hi Brian, 

Yes, it's doable.

1. download Presto 2.7.0 (it comes with server component and other UI apps to manage/compose services)

2. follow the tutorials (you will see the links on the home page once you have your server running or you can follow "Resources" tab here (see menu above)

3. for web clipping example, see <presto_installation>/mashupclient/samples/webclipping.emml  (there is no UI tool for setting up the scraping for intranet services, but you can do using EMML)

4. You can publish this EMML in Presto as a service that you can invoke from SoapUI or other apps/tools. 

- Girish@JackBe

bniknar
User offline. Last seen 2 years 6 weeks ago. Offline
Joined: 12/21/2009
Points: 10

Thanks Girish.

Can Presto Wires help with the EMML creation? My goal is to enable a team to create EMML that scrapes an Intranet site, then call it as a Web service. I envision a set of Tomcat-hosted web services, each that represents a different web server to scrape.

 

Regards, Brian

smitchell
smitchell's picture
User offline. Last seen 16 hours 33 min ago. Offline
Joined: 08/29/2008
Points: 34

In most cases, the answer to your question is no, but it does depend on your specific situation. You can use the DirectInvoke action in Wires to return the HTML from a web page on an Intranet, as long as the page does not require authentication to access it.

Even then what typically happens is that pulling out the 'relevant information' in that web page that you want to work with as a service involves semi- to really-complex XPath expressions that  Wires does not support. For example, if you are primarily interested in <div> elements within the page that have a class attribute with a a value of 'customer' (//*:div[@custom='customer']).

You can use the Web Clipping sample as a simple template to copy and then tweak. But it does mean learning some XPath and perhaps also some EMML.

 

 

Sara, technical writer/jackbe

 

aishmishra
aishmishra's picture
User offline. Last seen 10 weeks 6 days ago. Offline
Joined: 09/24/2008
Points: 3

Hi Brian,

To try and answer your questions:

i. Wires is a tool for (Mashup) EMML creation. So yes we can create mashups with Wires

ii. There is in fact an EMML sample for screen scraping in the mashup samples that are packaged with Presto.

iii. You can expose mashups as REST or SOAP based service

You can find details on these in the documentation. Please refer to them and do let us know in case you have any queries.