Presentation is loading. Please wait.

Presentation is loading. Please wait.

How to Build a Database- driven MIRC Teaching File System – A Case Study Stephen Moore Mallinckrodt Institute of Radiology.

Similar presentations


Presentation on theme: "How to Build a Database- driven MIRC Teaching File System – A Case Study Stephen Moore Mallinckrodt Institute of Radiology."— Presentation transcript:

1 How to Build a Database- driven MIRC Teaching File System – A Case Study Stephen Moore Mallinckrodt Institute of Radiology

2 Financial Disclaimer Stephen Moore is employed by Washington University and has received no extramural funding for this project. The Mallinckrodt Institute does not charge for access to the NM teaching file or for the storage service software.

3 Intended Audience Those who wish to add a Storage Service to an existing Teaching File Those who wish to build a Teaching File that includes a Storage Service Not someone who wants to install/run someone elses software –You need a different session or exhibit

4 Who Knows How the MIRC Community Works?

5 MIRC Internet Server Index Server Index MIRC site Server Index Query Service MIRC site Server Index RSNA site User MIRC site

6 MIRC Terminology Query Service –A point of access to the entire MIRC community. It provides a query form to the user, distributes the search criteria to all selected storage services, collates the responses, and presents them to the user. Storage Service –Responds to the query received from the query service, searches its index for documents meeting the search criteria, and returns abstracts and locations of the matching documents to the query service

7 Approaches to Creating a Storage Service Install, manage RSNA software –Attractive solution if you dont already have a teaching file Write new software to interface to your existing teaching file (live link to existing database) –Allows existing teaching file to continue and you get automatic updates to Storage Service Map data to MIRC-based database; write new software to interface to new database –Snapshot approach; minimal interference with existing teaching file

8 Storage Service Model DB Query Service Storage Service User Web Server Control Logic DB Interface Output Generator XML Parser

9 Storage Service Processing Steps Web Server passes query to Control Logic Control Logic invokes XML parser to produce a query Control Logic invokes DB Interface (or DB directly) to perform search Output generator produces legal XML that is passed back to Query Service

10 Software Development Steps Define system architecture Make technology choices for web server, parser, database Map MIRC Document Schema to your database schema (4-12 hours) Map XML query to existing database schemas (2- 6 hours) Extend MIR software based on new schemas (2 days)

11 MIR MIRC Components Java Tomcat Web Server PostgreSQL Relational Database http post, XMLServlet API SQL Java Servlet (MIR)

12 Java Servlets Are a Technology Choice Java Servlet technology provides Web developers with a simple, consistent mechanism for extending the functionality of a Web server and for accessing existing business systems. A servlet can almost be thought of as an applet that runs on the server side -- without a face.

13 Sun Java 2 Software Development Kit (J2SDK) The J2SDK provides core technology for the Tomcat web server and for servlets http://java.sun.com We are using an older version (1.2.2) We intend to upgrade to 1.4.2

14 Apache Tomcat Server Tomcat is the servlet container that is used in the official Reference Implementation for the Java Servlet and JavaServer Pages technologies. The Java Servlet and JavaServer Pages specifications are developed by Sun under the Java Community Process.

15 Apache Tomcat Server Provides a simple interface to obtain the http commands and return results http://jakarta.apache.org/tomcat/index.html We use version 4.1.29 (today) We have used 4.0.3 and 4.1.12 previously

16 Apache Xerces XML Parser Xerces2 is a fully conforming XML Schema processor. http://xml.apache.org/xerces2-j/index.html We use Xerces-J-bin-2.5.0. I see a 2.6.0 release is available as of 11/20

17 PostgreSQL Relational Database A robust, relational database supporting SQL queries Software is freely available without licensing fees We use version 7.1, 7.3 http://www.postgresql.org

18 Java Interface to PostgreSQL PostgreSQL contains an optional package that implements JDBC This is compiled and installed one time It allows a Java class (servlet) to execute SQL operations

19 Other Technology to Consider Apache web server / CGI / PHP Microsoft IIS / ASP Relational databases –MySQL –Oracle, SQL Server, Sybase XML index

20 MIR Storage Service (Moore/MIR) Web ServerApache Control LogicCustom (Java servlets) XML ParserXerces-J (2.5.0) DB InterfaceJDBC (PostgreSQL) DatabasePostgreSQL / Unix Output GeneratorCustom (Java servlets)

21 UCSF Storage Service (Tellis/UCSF) Web Server4 th Dimension Control LogicCustom (scripting language) XML ParserExpat4D DB Interface4 th Dimension Database4 th Dimension Output GeneratorCustom (scripting language)

22 Software Development Steps Define system architecture Make technology choices for web server, parser, database Map MIRC Document Schema to your database schema (4-12 hours) Map XML query to existing database schemas (2- 6 hours) Extend MIR software based on new schemas (2 days)

23 One MIRC Document = One NM Teaching File Case We decided not to replicate Teaching File functions with this Storage Service The MIRC Document that we return gives two pointers –One reference to the specific case that is chosen –Second reference to the Nuclear Medicine Teaching File main page

24 MIRC Query Schema 21 first tier elements –Title –Author –Abstract –Keywords Patient and image elements have sub- elements (9 such child elements)

25 MIR Nuclear Med /MIRC TF Schema We mapped the 30 MIRC schema elements to 12 columns in one table Remaining 18 elements are not implemented (but should be)

26 MIRCQuery MIR DB Columns MIRC QueryNM Database MIR MIRC DB titlebrief historytitle author abstractfull_historyabstr diagnosis diag_dx

27 Loading Data MIRC Database MIR Nuclear Medicine Teaching File can be exported as a series of HTML files We wrote a perl script to scan each file/record and produce an equivalent record in our SQL table

28 Mapping NM Database to MIRC Database (Snapshot) perl script HTML Files SQL insert statements PostgreSQL Relational DB

29 Questions to Ask When Mapping to Your Schema If starting from scratch, will you create a flat database design that matches the MIRC Document schema? –Do you have different tables for different concepts (patient, study, image)? How will your system change in the future should the query schema change? If you have an existing system, can you create a database view to get close to the query requirements?

30 A Simple MIRC Document Can Be Returned Attributes that we return in response include –docref –title –author –abstract RSNA Query Service will take care of formatting the output for us

31 Software Development Steps Define system architecture Make technology choices for web server, parser, database Map MIRC Document Schema to your database schema (4-12 hours) Map XML query to existing database schemas (2-6 hours) Extend MIR software based on new schemas (2 days)

32 Mapping MIRCQuery to Your Database If underlying technology is a SQL database, you need only parse the XML query and produce appropriate SQL and harvest results select title, author, abstr from doc_reference where keywords ilike %paget%;

33 Other Considerations Is there any private data in your existing database that you need to protect? Can you easily run case-insensitive queries against your database? Is it simple to provide a document + URL that will get the user directly to your teaching file case? Is this a system to also manage your teaching file or is it designed to provide another entry point?

34 MIR Implementation Details

35 Lines of Java Code Class to parse XML600 lines * Setup JDBC interface50 lines JDBC/SQL query70 lines Publish results30 lines * The XML parsing code uses the Xerces-J software. Our software could be more efficient.

36 Java Parser Sample public void doPost (HttpServletRequest req, HttpServletResponse res){ DocumentBuilder b = DocumentBuilderFactory.newInstance.newDocumentBuilder(); BufferedReader rdr = req.getReader(); org.xml.sax.InputSource is = new org.xml.sav.InputSource(rdr); Document d = builder.parse(is);

37 JDBC Code Sample Class.forName(org.postgresql.Driver); mDB = DriverManager.getConnection( jdbc:postgresql:doc_reference?user=postgres); Statement stmt = mDB.createStatement(); Stmt.execute(select title, author, … where …); ResultSet rs = stmt.getResultSet(); while (rs.next()) { // process row from database String title = rs.getString(1).trim(); String author = rs.getString(2).trim(); … }

38 Setup Steps Install J2SDK5 min Install Tomcat server5 min Compile/install PostgreSQL (Solaris) 3 hours ** Install/configure PostgreSQL (Linux) 1 hour **

39 Conclusion The process to build a Storage Service interface is relatively easy Free tools exist (web server, servlets, relational database) that assist the developer Our software available next week http://www.erl.wustl.edu/RSNA-MIRC Questions?


Download ppt "How to Build a Database- driven MIRC Teaching File System – A Case Study Stephen Moore Mallinckrodt Institute of Radiology."

Similar presentations


Ads by Google