Transcript Inetium and New Horizons: MOSS Search Deep Dive

SharePoint 2007:
Search Deep Dive
Erik Mau
Inetium
Agenda
•
•
•
•
•
•
Search Overview
User Experience
Extensibility
Administration
Scenarios
Q&A
SEARCH OVERVIEW
Topics
•
•
•
•
•
Topology
High-level Architecture
Performance
WSS vs. MOSS Search Comparison
Search Server 2008
Topology
• Single Server
• Separate Index Server with WFE playing Query
Role
• Separate Index and Separate Query Server
• Index Server can crawl itself or a specific WFE
Architecture
•
•
•
•
•
•
•
•
Index Engine
Query Engine
Protocol Handlers
IFilters
Content Index
Property Store
Search Configuration
Wordbreakers
Performance
•
•
•
•
Change Log / Security Change Only Crawl
Continuous Propagation
10s of Millions of documents per Indexer
WSS sites use MOSS Search
WSS vs. MOSS Search Comparison
Windows
SharePoint
Services
Microsoft Office
SharePoint Server
Local SharePoint
content
SharePoint, Web,
Exchange File
Shares, Notes,
LOB…
Rich, relevant results
X
X
Alerts, RSS, DYM, Dup collapse
X
X
Can Index
Scopes, Managed Properties
X
Best Bets, Result Removal, Query Reports
X
Tabs
X
People Search, *KN
X
**BDC Search
X
APIs Provided
Query
* Knowledge Network is a separate product
** BDC is Enterprise CAL
Query + Admin
Introducing Microsoft Search Server 2008 Express
USER EXPERIENCE
Topics
•
•
•
•
Search Center
Search Usability Features
Power User Features
Demo
Search Center
• Site Template
• Tabbed Interface
• Multiple Entrypoints
• Keywords,
Definitions, and
Best Bets
• Extensible
Search Usability Features
• Hit Highlighting
• “Did you mean” feature
• Duplicate collapsing
“Power User” Features
•
•
Property-based Searches
•
Customer:Acme (find all items where Customer metadata =
Acme)
•
Author:”Sally Jones” (find all items authored by Sally Jones)
•
ContentType:task (find all items of a specific content type)
•
Site:http://www.microsoft.com
Keyword Searches
•
+Solution (keyword inclusion)
•
-Plumtree (keyword exclusion)
•
Quotes for Exact Match (i.e., “User Group”)
USER EXPERIENCE DEMO
EXTENSIBILITY
Topics
•
•
•
•
Search Web Parts
Search Tabs
Search API
Demo
Browser-Based Extensibility Options
• Search Web Parts
–
–
–
–
–
–
–
Search Box / Advanced Search
Search Summary
Search Statistics
Search Paging
Search Actions
Search Best Bets / High Confidence Results
Search Results
• Search Tab Pages
Search Input Web Parts
• Search Box
–
–
–
–
Scope options
Additional query terms
Advanced Search Page
Search Results Page
• Advanced Search
– Input types
– Properties (metadata)
– Search Results Page
Search Utility Web Parts
• Search Summary
– “Did you mean”
– “You searched for”
• Search Statistics
• Search Paging
– Next / Prev Text
– Next / Prev Images
• Search Action Links
– Determine what links to display
Search Best Bets / High Confidence Results
•
•
•
•
•
•
Same web part, different configuration
XSLT for display
Show Keywords?
Show Best Bets?
Show High Confidence Matches
Best Bet Limit
Search Core Results
•
•
•
•
•
•
Results per Page
Enable removal of duplicates
Enable stemming
Selected Columns
Configurable XSLT
Scope
Search Tabs
• Based off of lists in Search Center Site
– Tabs in Search Pages
– Tabs in Search Results
• Process:
–
–
–
–
–
Create Entry Page
Create Results Page
Configure Web Parts on Entry and Results Page
Add Entry Page to Tabs in Search Pages
Add Results Page to Tabs in Search Results
Bringing it all together…
Search API
• KeywordQuery
– Allows basic property-based and free text queries
– Namespace: Microsoft.Office.Server.Search.Query
– Example:
KeyWordQuery query = new KeyWordQuery();
query.SelectProperties.Add(“Customer”);
query.QueryText = “Author:Erik SharePoint”;
ResultTableCollection results = query.Execute();
Search API
• FullTextSqlQuery
– Allows highly customized MSFT Full text search
syntax queries
– Namespace: Microsoft.Office.Server.Search.Query
– Example:
FullTextSqlQuery query = new FullTextSqlQuery();
query.QueryText = “SELECT Customer, Path FROM Scope() WHERE …”;
ResultTableCollection results = query.Execute();
EXTENSIBILITY DEMO
SEARCH ADMINISTRATION
Topics
•
•
•
•
SSP Administration Features
SSP Administration Demo
Site Collection Administration Features
Site Collection Administration Demo
Search Administration
SSP ADMINISTRATION
SSP Administration Overview
• Search is managed at the Shared Service
Provider (SSP)
• Provides management of:
– Content Sources (What is being indexed)
– Managed Properties (How users want to perform
advanced searches)
– Scopes (How users want to search)
– Relevance (What content sources are more relevant)
– Health (What are users searching for and how are my
crawls performing)
Content Source Management
• What can we index?
–
–
–
–
–
–
SharePoint Sites (v2, v3)
Web Sites
File Shares
Exchange Public Folders
Lotus Notes
Business Data (BDC)
Business Data Catalog
• MOSS 2007 (e-cal) provides access to business
data to users
– Data is accessed through the Business Data
Catalog
•
•
Databases
Web Services
– Exposed to many features in SharePoint
•
•
Web Parts (Lists and Details)
Metadata Columns
Business Data Catalog
• Search can easily index the data
– No need to write code
– Highly customizable results
– Launching pad into other web-based, querystringaddressable applications.
– Integrated search results in the search center:
•
•
•
Documents
Web Content
Business Data
Metadata Property Mappings
• Crawled Properties
– SharePoint “discovers” properties during crawl
– Organized in categories
– Can be included/excluded in the index
• Managed Properties
– Out-of-the-box Managed Properties
– Custom-defined Properties (i.e. CustomerName)
• CustomerName:Acme
– One Managed Property can be associated to 1 or
more Crawled Properties
Scope Management
• Logical organization of crawled content
• Rule-based
– Web Address / Folder
– Property Query (Author = John Doe)
– Specific Content Source
• Supports multiple rules (joined via ANDs / Ors)
• Supports content exclusion
• Site Collections “opt-in” to SSP-defined scopes
Relevancy
• Authoritative Pages
– Three levels of “authority”
– Ability to demote sites
• Relevancy API
– API Available to change the ranking of managed
properties
Crawl Log
• Administrative interface for viewing crawl history
• Summary View
• Details View
Search Usage Reporting
• Out of the box usage reporting
in MOSS
– Query volume trends, top queries,
click through rates, queries with zero
results, etc.
– At both SSP and site collection levels
– Export data for extended reporting in Excel
Search Usage Reporting
Search Administration
SSP ADMINISTRATION DEMO
Search Administration
SITE COLLECTION ADMINISTRATION
Site Collection Administration Overview
•
•
•
•
•
Site Collection Scopes
Scope Visibility
Search Visibility Settings
Keywords / Best Bets Management
Site Collection Search Usage Reports
Site Collection Scopes
• Similar interface to SSP Scope Management
• Organize by Display Group
• Same flexibility for rules
Scope Visibility
• Define Display Groups
• Groups Control Visibility on Search Web Parts
Search Visibility
• Control if a site shows up in search results
• Sites are already security trimmed
• Example: Search Center Set to “No”
Keywords & Best Bets
• Use query reporting
to help define
keywords
• Use keywords to
provide access to
content inaccessible
to the crawler
Search Administration
SITE COLLECTION ADMINISTRATION DEMO
SCENARIOS
Scenario 1: Creating a Search Tab
The sales department at Litware, Inc. would like easier
access to their documents in SharePoint, documents on a
legacy file share, and information on one of their competitor
sites.
Solution:
-New Content Sources
-New Scope
-New Search Center Tab
Scenario 2: Customizing Search Results
The sales department at Litware, Inc. is raving about their new search
tab, but would like to display additional information about their
documents in the search results. They have several document types
and are using the business data catalog for associating customers to
documents. They would also like to search for customer records in the
same search tab.
Solution:
-BDC
-New Managed Properties
-Custom XSLT for Search Results Web Part
Scenario 3: Searching by Metadata
The power users on the sales team regularly use the property-based
syntax to find documents for a specific customer or by document type.
The sales team would like to make it easier to find content using
metadata.
Solution:
-Managed Property
-Advanced Search Web Part Or Content Editor Web Part
Scenario 4: Keywords and Best Bets
The sales team is about to begin creating their sales plans for fiscal
2008. Many team members use the search functionality to find their
plans from other fiscal years. The management team wants to ensure
that team members use the new planning template for F’08.
Solution:
-Define Keyword and Synonyms
-Define Best Bet
Scenario 5: Relevant Results
Jerry from engineering is frustrated that every time he performs
searches for certain keywords, there are links to one of their
competitors web sites in the search results.
Solution:
-Demote competitor site
-Exclude competitor site from All Sites scope
Q&A