Constellio
Constellio search in Tempus Serva installations

Activate the search servlet in your installation
The search servlet is deactivated by default. 

 

 Edit the <tomcat>/webapps/<Tempus Serva>/WEB-INF/web.xml 

 Remove comments from the search servlet 

 Remove comments from the search filter 

 

 If you are using web container security please remove it from the search servlet: The servlet filter will handle authentication of crawler robots using a specialized form of basic authentication (normal users will be redirected to the main servlet instead). 

 Option: Create a user for crawling 

 You will need at least 1 user for crawling the content in Tempus Serva, possibly more if content restrictions apply to different search user groups. 

 The following applies to crawling users 

 

 All group and policies will be respected through the indexing 

 No codeunits will be activated 

 No log entries will be created 

 

 You can test what the crawler will se by logging in with an extra parameter: 

 /login?SearchIndexing=true

Prepare Constellio
Install Constellio 

 

 Download the 1.3 installer 

 Run the installer by doubleclicking the .jar file 

 Install to MySQL database 

 Run the  Start constellio 

 

 Setting up a connector 

 Before setting up a connector create or choose a valid search scope 

 

 Choose connector type:  auth-http-connector 

 Ensure that  Use security  is checked 

 Set start URL to:  http://<server name>/TempusServa/search 

 Include the same URL in include patterns 

 Enter username for the crawler user (a valid TS user) 

 Enter password for the crawler user (a valid TS user) 

 After submitting the new connector, crawling/indexing will start by itself 

 

 No further actions are needed: 

 The search servlet will automatically redirect real users after they click on a search result. 

 Option: Tweak search results 

 The search servlet will automatically deliver content in a crude form, without any extra html such as wrappers. It will also provide the crawler with information about when it was last updated, and document Title will be se to current records Resume value. 

 You might consider excluding the  command=list pages for better (less redundant) search results.