Add Format Searching in Orbis New Books RSS Feed
- The condition in SQL query for format has a few "or". This will be slow down the searching speed a lot. So avoid these conditions, and get the whole result set, filter them in the programing loop.
- Add the format type as another element in RSS description display.
- Add the format type drop down list into the interface.
- Add submit parameter for format type in the Javascript.
- Add "not searching" restrictions in the Javascript. In this way, it will avoid searching a huge results set, and returning to browser diaplay; otherwise could cause two problems: consume the tremendous time to search database, and browser doesn't have capability to populate the huge data set.
- Add Java UTF-8 method to all of RSS element fields, except for guid, and link element fields.
Develop Finding Aids Pre-Built PDF from EAD Files
- This is a new application that replaces the existing one which has problems. Meet with different groups to understand Finding Aids ingest process and the problem.
- Read the existing codes to understand how it converts EAD XML file to PDF file.
- Learn from colleague about how to execute Apache FOP API on command line with requirement settings, and test it.
- Develop the codes:
- go through the directory for each EAD file,
- apply the fop.jar in Java's execution method to generate PDF file from EAD file,
- use getErrorStream to catch error and warning messages, and write them into log file,
- delete the PDF file if there is error message "OutOfMemoryError" during this file's converting process. So there will be no corrupt PDF files feed into ingest process.
- Add "-Xmx300m -Xms300m" into java execution command to increase the virtual memory in order to increase the probability for converting the large EAD into PDF.
- Test on springer and local desktop. This application needs big memory to support. The bigger memory the computer has, the larger EAD file size can be converted into PDF.
- Coordinate to run on production server. From assigned this project to successfully ran on production, I finished it all in one week.
Update OAI Map to Improve the Response Time
- Re-evaluate the open source code that OAI map is using, and find the sections that can improve.
- Change the external SQL query to just include the basic, no function, such as decode, tochar etc.
- Get and filter the records in programing loop, instead in SQL query by function.
- Remove the redundancy that repeating running the same query in different methods. The main reason that slows down the searching speed is query's running speed.
- Re-code to process records for adjusting the changes from the queries. This includes changes in SQL queries that are dynamically built internally.
- Make other changes that associate with the major changes.
Update RSS Program to Accept the Combination Conditions Searching
- Change mandatory inputs are only RSS feed channel title and new books week, instead of channel title, week, location and language.
- Optional inputs can be one of location, language, one or multiple call number range; or the combinations from any of them. But must have one optional input with mandatory inputs.
Troubleshooting Bursar Transfer Problem
- Figure out two weeks' bursar transfer files are the same.
- Figure out that process bursar sif files program didn't run.
- Analyze circulation client fine fee data related with bursar debit and credit transfering to compare with the debit and credit files which are sent to SFAS.
- Provide the solution about how to re-run the program of processing bursar sif files.
- Figure out the reason of stoping program running. This is caused by the carriage return and newline characters in the fine fee notes that mess the sif file which is generated by Voyager Pbursar program.
- Provide the solution about how to clean up that mess in the sif file.
- Tutor the colleague about the whole bursar transfer procedures, including how to read the Pbursar sif files, SFAS bursar transfer files, meaning of bursar data from circulation client, unix script, programs functions and how to troubleshooting when the problem happens.
Develop Orbis New Books RSS feed
- Develop two ways to output the XML file. One is Java Servelet which locates in WEB_INF classes. It needs to build up the whole XML file before it displays in browser. Another one is JSP which locates under root or any sub directory. This way can display XML file on browser while generating the XML file. If XML file is not big, the two methods are no difference. But if the XML file is big, to avoid blank web page shows up too long, the JSP method is more appealing.
- Generate new books' feed elements: title, link, guid, pub date, description, category.
- Create artificial customized elements in the feed's description.
- Develop the interface to accept limitation search conditions: new books' weeks, location, language, call number range.
- The spaces among normalized call number have some problems. This can lead some missing results. So use the display call number to search.
- Location element includes item location and call number.
Return to the top