Funnelback Manager

The Funnelback Manager

The Funnelback Manager allows you to create and manage the Collections to be used on the Funnelback Search Page and is located under the System Management Folder, as highlighted in the figure to the right.

The tasks managed by the Funnelback Manager include:

• The creation and control of Collections
• The indexing configuration for the Funnelback Indexer
• The indexing of Collections that have been created
• The rebuilding of cache on the system
• The management of the search log archives 

Management of indexing, cache rebuilding and search query logging is done on the Details and Funnelback Indexer screens, while Collections are created on the Collections screen and maintained on the Thesaurus, Feature Pages, Contextual Navigation and Metadata Mapping screens. Each of these screens are explained in this chapter. For more information on the other screens available on the Funnelback Manager, refer to the Asset Screens manual.

Details Screen

The Details screen for the Funnelback Manager allows you to configure the indexing options and archive logging of the Funnelback Search Page.

Indexing Manager

This section allows you to enable indexing for Funnelback. The Indexing Manager section of the Details screen is shown in the figure below.

The Indexing Manager section of the Details screen

For the Funnelback Search Page to work, Funnelback indexing must be enabled in the system. Once indexing is enabled, the Rebuild Cache section will appear, allowing you to rebuild cache on the system. The fields on the Funnelback Indexer screen will also become available, allowing you to index your Collections.

To enable indexing, select Enabled in the Enable Indexing field. Please note that indexing will not be enabled if the Funnelback installation cannot be found in the System.

Indexing Options

This section allows you to control the settings of the Funnelback Indexer. The Indexing Options section of the Details screen is shown in the figure below.

The Indexing Options section of the Details screen

The fields available in this section are as follows:

• User: select a User account to use as the Funnelback user login account. This login will be used to index the content of private assets within the system. The content that is indexed is dependant on the Permissions of the User account selected and the content they have access to; all assets that this user has Write access on will be indexed. Additionally, Live assets that this user has Read access on will also be indexed.

  Tip: Rather than applying Write permission on each asset manually, you can apply these permissions using the give_funnelback_permission.php script in the <system_root>/funnelback/scripts directory. Running this script will allow you to select the root nodes to automatically apply Write permission on for the selected User account.

  If the User account selected in this field has Write access on an asset in Safe Edit, the edited content of this asset will be indexed. This may result in inaccurate results for users who do not have access to this Safe Edit content.

• Password: enter the password for the Funnelback user login account.
• Time Delay (Politeness): enter the time delay (in milliseconds) that the Funnelback Indexer will wait between URL documents when indexing.
• Turn off Speller: select whether or not to disable spelling indexing. If this field is disabled, the Funnelback search will not be able to suggest the correct spellings for misspelled words.  For more information on spelling suggestions, refer to the Displaying Spelling Suggestions example in this manual.
• Spelling Threshold: select the number of times a word must appear before being indexed for spelling suggestions. For example, if this field is set to 5, a word must appear five times during indexing before it will be considered for spelling suggestions. Increasing the number in this field will increase the quality of spelling suggestions on the Funnelback search, however not as many words will be indexed. A lower number will increase the amount of suggested words at the risk of reducing the quality, i.e. a misspelled word may be indexed.
• Index if binary: select whether or not to index documents that appear to be binary. If this field is set to Yes, binary documents will be indexed by the Funnelback Indexer and appear in search results. By default, this field is set to No.
• Index Comments: select whether or not to index comments found in HTML and XML documents. If this field is set to Yes, comments in these documents will be indexed by the Funnelback Indexer and appear in search results. By default, this field is set to No.

Rebuild Cache

This section allows you to rebuild the cache on either a selected root node or system-wide. The Rebuild Cache section of the Details screen is shown in the figure below. Please note that this section will only appear if indexing is enabled on the Funnelback Manager.

The Rebuild Cache section of the Details screen

Rebuilding cache allows the Funnelback Indexer to index the latest cached version of your site. You should rebuild cache if you make significant asset changes, move a system to a new URL or make database changes. To do this, select the root node under which to rebuild cache and click Commit. Alternatively, you can choose to rebuild the cache for the whole system by selecting the System Wide field and clicking Commit. This will rebuild the site map xml cache stored in the system. Please note that this process may take a while on larger systems.

Tip: The Funnelback Manager can be scheduled to automatically rebuild cache by using the funnelback_rebuild_cache.php script on the system crontab. This script takes the Squiz Matrix system root directory and an optional root node. Omitting this root node will rebuild cache system wide. 

Logging

Search query logs are archived as scheduled by the Funnelback Log Rotation Cron Job. By default, this Cron Job is set to run weekly, as shown in the figure below.

The Funnelback Log Rotation Cron Job

This section allows you to select the length of time to keep these archived logs. The Logging section of the Details screen is shown in the figure below.

The Logging section of the Details screen

In the Keep Archived Logs field, enter the number of months to keep the logs. By default, this is set to 6 months.

Collections Screen

The Collections screen for the Funnelback Manager allows you to create and configure Collections of assets, indexed together and made available for searching on the Funnelback Search Page.

New Collection

This section allows you to create new Collections. The New Collection section of the Collections screen is shown in the figure below.

The New Collection section of the Collections screen

To create a new Collection, first fill in the details of the Collection. The following fields are available:

• Collection ID: enter an ID for the Collection.
• Collection Name: enter a name for the Collection.
• Collection Type: select the type of Collection that will be created. The following Collection Types are available:  
  • Web Collection: an asset and all of the assets beneath it, including system folders (e.g. the Media and Users folders).
  • Meta Collection: a set of Web Collections (i.e. a set of assets).

Once you have done this, click Commit. The Collection will appear in the Edit Collections section, as described below.

Edit Collections

This section displays any Collections created on the system. By default, no Collections will be displayed in this section, as shown in the figure below.

The Edit Collections section of the Collections screen

When Collections have been created, they will be listed in this section, as shown in the figure below. For more information on creating Collections, refer to the New Collection section above.

The Edit Collections section listing invalid Collections

As these Collections do not yet have any defined data source (e.g. an asset or Web Collection), they are not Valid and cannot be used in the Funnelback Manager. When a data source has been selected on a Collection, it will be deemed Valid, as shown in the figure below.

The Edit Collections section listing valid Collections

To add and edit the settings of a Collection, click on the  icon. For more information, refer to the Edit Selected Collection section, below.

To delete a Collection, select the Delete? field and click Commit.

Edit Selected Collection

When editing a Collection, fields will appear in the Edit Selected Collection section, allowing you to edit the data source of the selected Collection. These options differ depending on the type of the Collection you are editing. 
The Edit Selected Collection section for a Web Collection is shown in the figure below.

The Edit Selected Collection section for a Web Collection

The following fields are available when editing a Web Collection:

• Collection Name: by default, the name that was entered when the Web Collection was first created will appear in this field. To change this name, enter a new name in this field
• Collection Source Asset:  select the data source of the Web Collection. This can be an asset or a Web Folder. Once you have selected the source, click Commit. The URLs of the selected source will be listed in the Collection URL field, as shown in the figure below.  

  
  URLs in the Collection URL field  

• Collection URL: select the URL(s) that will be used for the root path of data in the Collection. Only URLs under the selected root path will be indexed for searching. For example, a Site has both the http://example1.com and http://example2.com URLs assigned to it. If you select the example1.com URL in this field, only data under this URL will be indexed (e.g. example1.com/home, example1.com/news etc).

Once you have selected a Source Asset and URL(s) for the Web Collection, click Commit. The Web Collection will now be listed as Valid in the Edit Collection section of this screen.

The Edit Selected Collection section for a Meta Collection is shown in the figure below.

The Edit Selected Collection section for a Meta Collection

The following fields are available when editing a Meta Collection:

• Collection Name: by default, the name that was entered when the Meta Collection was first created will appear in this field. To change this name, enter a new name in this field.
• Collections: this field will list the Web Collections configured on the Funnelback Manager (in the figure above, the Library and News Web Collections). Select the Web Collections to assign on the Meta Collection in the corresponding Please Select fields.

Once you have assigned a set of Web Collections for the Meta Collection, click Commit. The Meta Collection will now be listed as Valid in the Edit Collection section of this screen.

Thesaurus Screen

The Thesaurus screen allows you to configure the thesaurus of a Collection for indexing. This thesaurus lets you control the search results of a specific search by creating replacement queries for set terms. For more information on the usage of the thesaurus and how it is configured, refer to the Configuring the Thesaurus example in this manual.

Thesaurus

In the Choose a collection field, shown in the figure below, select a Collection to edit its thesaurus. If no Collections have been created on the Funnelback Manager, nothing will appear in this field. For more information on creating Collections, refer to the Collections Screen section of this chapter.

The Choose a collection field with Collections

Edit Thesaurus

Once you have selected a Collection, its thesaurus will appear in the Edit Thesaurus section, as shown in the figure below.

The Edit Thesaurus section of the Thesaurus screen displaying a new Thesaurus

The following fields are available:

• Query Type: select the match type of the thesaurus rule. This determines how the thesaurus Term is matched to the user’s search query. The following options are available:  
  • Term by Term Match: the thesaurus Term will match to the search query on an individual term basis. For example, if the search query is Squiz Matrix, the Term can be either Squiz OR Matrix to match. This is the default Query Type for a thesaurus rule.
  • Whole Query Match: the thesaurus Term will match to the search query on a full query basis. For example, if the search query is Squiz Matrix, the Term needs to be Squiz Matrix to match.
  • Regular Expression Match: the thesaurus Term will match to the search query against a regular expression term. This allows you to use the regular expression format to create term rules such as starts with a numeric (in regexp: \b\d). For example, the Term w.*the will match to a search query such as where are the documents.
• Term: enter a Term that will be matched against all search queries on the Funnelback Search Page. This Term can be one or more query terms.
• Replacement: enter a replacement search query to use if the thesaurus Term is matched. For example, if the thesaurus Term is Squiz, you could set the Replacement as Squiz Matrix. This means that if a user searches for Squiz, the search results will instead display the results of Squiz Matrix.  

  Tip: You can use query operators in this field to create a more advanced Replacement. For example, encasing two terms in square brackets will act as an OR rule, the search results displaying documents that contain at least one of the query terms.
 For example: Term=Matrix, Replacement=[“Squiz Labs” Matrix] – If a user searches for Matrix, the search results will display documents that contain at least one of eitherMatrix OR “Squiz Labs”.   

• Delete: to delete a thesaurus rule, select this field and click Commit.

Feature Pages Screen

The Featured Pages screen allows you to specify certain URLs to index as Feature Pages on a Collection. These Feature Pages are displayed on the results page of specific search queries, as set by a configured Trigger Word(s). For more information on creating a Feature Page, refer to the Creating a Feature Page example in this manual.

Feature Pages

In the Choose a collection field, shown in the figure below, select a Collection to create the Feature Page for. If no Collections have been created on the Funnelback Manager, nothing will appear in this field. For more information on creating Collections, refer to the Collections Screen section of this chapter.

The Choose a collection field with Collections

Once you have selected a Collection and clicked Commit, new fields will appear in the Edit Feature Page field, as shown in the figure below.

The Edit feature page field listing a blank Feature Page

A blank Feature Page is automatically created. To add more Feature Pages, enter the number of pages to create in the Add new feature page entries field and click Commit.

As these Feature Pages do not yet have any defined settings, they are not Valid and cannot be used by the Funnelback Manager. When the Feature Page has been configured correctly, it will be deemed valid, as shown in the figure below.

The Edit feature page field listing a valid Feature Page

To edit the settings of a Feature Page, click the  icon. For more information, refer to the Edit Feature Page section below.  

To delete a Feature Page, select the Delete? field and click Commit.

Edit Feature Page

When editing a Feature Page, fields will appear in the Edit Feature Page section, allowing you to edit the settings of the Feature Page. The Edit Feature Page section is shown in the figure below.

Additional fields in the Edit Feature Page section of the Feature Pages screen

The following fields are available:

• Query Type: select the match type of the Feature Page. This determines how the Trigger Word is matched to the user’s search query. The following options are available:  
  • Term by Term Match: the Trigger Word will match to the search query on an individual term basis. For example, if the search query is Squiz Matrix, the Trigger Word can be either Squiz OR Matrix to match. This is the default Query Type for a Feature Page.
  • Exact Match: the Trigger Word will match to the search query on a full query basis. For example, if the search query is Squiz Matrix, the Trigger Word needs to be Squiz Matrix to match.
  • Regular Expression Match: the Trigger Word will be matched to the search query against a regular expression term. This allows you to use the regular expression format to create term rules such as starts with a numeric (in regexp: \b\d). For example, the Trigger Word w.*the will match to a search query such as where are the documents.
  • Substring Match: the Trigger Word will be matched to the search query on a sub-string match of all terms against the full query. This means that all of the Trigger Word terms must be found within the query (full or partial word match). For example, the Trigger Word document site will match to search queries such as document site, site document and documentation website.
• Trigger Word: enter the term(s) to act as the Trigger Word that will be matched against search queries on the Funnelback Search Page. This Trigger Word can be one or more query terms.
• Page Title: enter a title for the Feature Page. This will be displayed when the Featured Title keyword replacement is used on the Featured Result Format Bodycopy.
• Page Description: enter a description of the Feature Page. This will be displayed when the Featured Summary keyword replacement is used on the Featured Result Format Bodycopy.
• Asset: select the asset source of the Feature Page. This can be any asset within the Collection you are editing. Once you have selected the asset, click Commit. The URLs of the asset will be listed in the URL field, as shown in the figure below.  

  
  URLs listed in the URL field  

• URL: select the URL(s) that will be used for the Feature Page in the search results.

Once you have filled out these fields, click Commit. The Feature Page will now be listed as Valid in the Edit Feature Page section of this screen.

Contextual Navigation Screen

The Contextual Navigation screen allows you to configure the index settings for Contextual Navigation. This will determine the related clusters that can be displayed on the search results page, giving users the ability to refine their search results by type, topic or site.

For more information on setting up Contextual Navigation, refer to the Setting Up Contextual Navigation example in this manual.

Contextual Navigation


In the Choose a collection field, shown in the figure below, select a Collection to configure Contextual Navigation for. If no Collections have been created on the Funnelback Manager, nothing will appear in this field. For more information on creating Collections, refer to the Collections Screen section of this chapter.

The Choose a collection field with Collections

Edit Contextual Navigation

Once you have selected a Collection and clicked Commit, new fields will appear in the Edit Contextual Navigation section, as shown in the figure below.

The Edit Contextual Navigation section of the Contextual Navigation screen

The following fields are available:

• Cannot End With: the list of words and phrases that will prevent all clusters ending in these terms from appearing on the Funnelback Search Page. This allows you to manage the clusters that will appear for users on the results page. For example, if the word says appears on this list, a cluster titled Squiz says would be excluded on the results page. The default list of words is: says, needs, wants, require, requires, and contains. To add a word or phrase to this list, enter it in one of the empty fields. To remove a word or phrase from the list, click on the  icon next to its corresponding field.
• Case Sensitive: enter any case-sensitive words that should retain their capitalisation when appearing in clusters. For example, Australia, Mr. and NSW.
• Categorise Clusters: select whether or not clusters should be categorised into sites, topics and types. This will allow the Funnelback Manager to determine whether a cluster is a site, topic or type relating to the search query and display appropriate grouping accordingly. Please note that this categorisation is not always accurate, but is usually beneficial when providing a set of categorised clusters. Select True to enable this feature or False to disable it; by default this field is set to True.
• Kill List: the list of words and phrases that will prevent all clusters containing these terms from appearing on the Funnelback Search Page. This allows you to manage the clusters that will appear for users on the results page. For example, if the word more appears on this list, a cluster titled show more would be excluded on the results page. The default list of words is: more, govt, http, https, has, was, is, but, says, au, browse, were, or, may, which, printable, provide, 95, back to top and ago. To add a word or phrase to this list, enter it in one of the empty fields. To remove a word or phrase from the list, click on the  icon next to its corresponding field.
• Max Phrase Length: enter the maximum word limit of cluster phrases. For example, if this field is set to 2, a cluster titled News Items would be indexed, however a cluster titled Latest News Items would not. By default, this field is set to 5.
• Max Phrases: enter the maximum number of clusters that will be indexed for Contextual Navigation. Increasing this value will improve the quality of suggestions; please note, however, that this may lead to a lower response time. By default, this field is set to 750.
• Max Results To Examine: enter the maximum number of results to examine when Contextual Navigation suggestions are generated on the results page. By default, this field is empty, which results in an internal system default of 150. Please note that the maximum number of results that can be examined is 200.
• Site Granularity: specify the type of granularity to be used for Contextual Navigation site suggestions. The options available include:  
  • host: the site suggestion will be displayed as the hostname of the site, for example, www.example.com.
  • domain: the site suggestion will be displayed as the domain of the site, for example, example.com.
  • tldir: the site suggestion will be displayed as the top level directory of the site, for example, www.example.com/home.
  By default, this field is set to host.
• Summary Fields: enter the metadata classes (configured on the Metadata Mapping screen) that are to be included in Contextual Navigation indexing. This means that if you have configured metadata class a to contain author metadata information, you can add this class to the list in order to include this information in clusters. By default, the metadata classes c (description) and s (dc.subject) are entered into this field. For more information, refer to the Metadata Mapping section of this chapter.
• Timeout Seconds: enter the amount of time (in seconds) that Contextual Navigation should generate results for. If Contextual Navigation times out, the results that were generated in that time will be displayed.

Metadata Mapping Screen

The Metadata Mapping screen allows you to group the metadata tags of a Collection into classes, indexed by the Funnelback Manager for querying on a Funnelback Search Page. This allows users to use metadata scoping to refine their search results.

For more information on configuring and using Metadata Mapping, refer to the Using Metadata Mapping example in this manual.

Metadata Mapping

 
In the Choose a collection field, shown in the figure below select a Collection to configure Metadata Mapping for. If no Collections have been created on the Funnelback Manager, nothing will appear in this field. For more information on creating Collections, refer to the Collections Screen section of this chapter.

The Choose a collection field with Collections

Edit HTML Meta Tag Mapping

Once you have selected a Collection and clicked Commit, new fields will appear in the Edit HTML Meta Tag Mapping section, as shown in the figure below.

Additional fields in the Edit HTML Meta Tag Mapping section

Thedefault and reserved Metadata Classes outlined in the table below will be listed in the Edit HTML Meta Tag Mapping section.

Metadata Class	Flag	Metadata Tag(s)
Metadata Class S	Reserved Metadata Class
Metadata Class a	Index	author, dc.author, dc.creator, dc.contributor
Metadata Class b	Index	dc.rights
Metadata Class c	No Index	dc.description, description
Metadata Class d	Index	dc.date, date, lastSaved, trim.datereg
Metadata Class e	Index	dc.type
Metadata Class f	Index	dc.format, text/html
Metadata Class g	Index	dc.relation
Metadata Class j	Index	dc.identifier, trim.number, agls.availability
Metadata Class l	Index	dc.language
Metadata Class n	Index	dc.source
Metadata Class o	Index	dc.coverage
Metadata Class p	Index	dc.publisher
Metadata Class q	Index	agls.function
Metadata Class r	Index	agls.audience
Metadata Class s	No Index	keywords, dc.subject
Metadata Class t	No Index	title, dc.title, trim.title
Metadata Class w	Index	agls.mandate

In most cases, it is possible to make additions to the default classes. Funnelback reserved classes, however, are not available for alteration.

To create or make additions to a Metadata Class, use the fields available at the bottom of the Edit HTML Meta Tag Mapping section. The fields available are as follows:

• Metadata Class: select the Metadata Class you want to create a new metadata tag for. There are 62 class types available: letters (in both upper and lowercase) and numbers. Please note that some of these class types are reserved by Funnelback and cannot be modified, for example Metadata Class S.
• Flag: select whether or not the metadata class should also be indexed as part of the document’s content or not. If this field is set to Index, the metadata information will be classified as content within the document and can be queried on the search page via the main search bar. If this field is set to No Index, metadata information will only be able to be searched in the Metadata Scope search field.
• Tag Name: enter the name of the metadata tag you want to add to the selected Metadata Class. For example, if you have a metadata field titled Version applied to your Site. Entering Version in this field will add this metadata tag to the Metadata Class.

Asset Exclusions Screen

The Asset Exclusions screen allows you to specify assets to exclude from Funnelback indexing. Please note that the exclusions configured on this screen will be applied across all collections on the system.

The Global Asset Exclusions section of the Asset Exclusions Screen

Assets can be excluded by either Asset ID, Asset Type or Root Node. The fields available are as follows:

• Exclude By Asset ID: select the Asset IDs to exclude from Funnelback indexing. To add more than one asset ID to exclude, click the More... button.
• Exclude By Asset Type: select the Asset Types to exclude from Funnelback indexing. For example, selecting Standard Page will mean that all Standard Page assets will not be indexed by Funnelback. To add more than one asset type to exclude, click the More... button.
• Exclude By Root Node: select a Root Node to exclude from Funnelback indexing. For example, selecting your Home page in this field would mean that this page and its children would not be indexed by Funnelback.To add more than one root node to exclude, click the More... button.

Tip: When an exclusion is added, changed or deleted on this screen, you will be required to rebuild cache and reindex the collections on your system to enable the changes.

Funnelback Indexer Screen

The Funnelback Indexer screen allows you to manually index content for searching on either a Collection or system-wide basis. The Funnelback Indexer screen is shown in the figure below. Please note that the fields on this screen will only be available if indexing is enabled on the Details screen of the Funnelback Manager.

The Funnelback Indexer screen with Collections

In the Root Node field, select the Collection you want to index and click Commit. If no Collections have been configured on the Funnelback Manager, nothing will appear in the Root Node field, as shown in the figure below. For more information on Collections, refer to the Collections Screen section in this chapter.

The Funnelback Indexer screen with no Collections

Alternatively, you can choose to index the whole system by selecting the System Wide field and clicking Commit. Please note that this process may take a while on larger systems.

Tip: You can configure the information that is indexed on a page in the contents of the page and in the parse file of its Design, for example, the navigation menu. To do this, append no index tags around the content that is not be indexed, as follows:

          <!-- noindex -->
          Content NOT to be indexed.
          <!-- endnoindex -->

Please note that if you change the settings of a Collection on the Funnelback Manager (Metadata Mapping and Contextual Navigation) or alter the assets within a Collection, these changes will not be reflected within the search results of the Funnelback Search Page until a re-index is performed. This can be done manually or via the funnelback_reindex.php script. 

funnelback_reindex.php

The funnelback_reindex.php script indexes the Collections within your system. The script is located in <system_root>/packages/funnelback/scripts, where <system_root> is the location of the Squiz Matrix  system you are using. It can be added to the system crontab to run on a schedule and takes the Squiz Matrix system root directory and the ID of the Collection to index. Not specifying a Collection ID will mean that the script will re-index all Collections within the system. Please note that you will need to run this script as the Apache User in order to have the permission to re-index the Collections.

It is recommended that this script be set to run once a day if your Collections are regularly updated. Please note however, that the changes made to your Collections will not be reflected within search results until the script is run. For example, if you have set the script to run daily at 2am, changes made to a Collection at 5pm will not be shown in the search results until the script is run again at 2am, the next day.

PreviousNext