April 10, 2007
----------------------------------------------------------------------------
1 Add CIPRES_BIN_HOME env variable in .cshrc or .bashrc to pick up the
location of cipres binaries including PAUP;
2 to get around the limitation of PAUP not accepting file name length > 100:
+ change CIPRES_WEB_HOME in .cshrc or .bashrc to a different location to short file directory path
+ shorten the name of UPLOADED_NEXUS_DIR variable in src/main/resources/cipres.properties
3 update the directory locations where cron job uses to remove files:
crontab -e
April 17, 2007
----------------------------------------------------------------------------
1. Terri made changes to *garli* command files to increase the memory to 2G to
fix the memory problem on large dataset;
availablememory = 2048
2. handle files without a starting tree... (-DtreeInferTool=paup-infer.xml)
3. allow user upload PHILIP format file;
4. Add StartTime, StopTime for each analysis run (input.xml)
5. make *.status file" downloadable;
6. change to in command files:
recidcm3_garli-gtr-g.xml
recidcm3_garli.xml
recidcm3_paup.xml
recidcm3_raxml.xml
rid3-rax-12.5-pct.xml
rid3-rax-25-pct.xml
rid3-rax-50-pct.xml
7. add 12.5%, 25%, 50% for PAUP(recidcm3_paup.xml), GARLI(recidcm3-garli-gtr-g.xml) based methods:
change SetMaxTaxaSubsetSize=20 to SetMaxTaxaSubsetPercent=(12.5, 25, 50)
rid3-paup-12.5-pct.xml
rid3-paup-25-pct.xml
rid3-paup-50-pct.xml
rid3-garli-gtr-g-12.5-pct.xml
rid3-garli-gtr-g-25-pct.xml
rid3-garli-gtr-g-50-pct.xml
8. make email field a required field
9. remove "-i 10" parameter value (delete -i 10 from
both Large Tree Inference and Small Tree Infererence for ridcm3-raxml:
recidcm3_raxml.xml: -i 10
recidcm3_raxml.xml: -i 10
rid3-rax-12.5-pct.xml: -i 10
rid3-rax-12.5-pct.xml: -i 10
rid3-rax-25-pct.xml: -i 10
rid3-rax-25-pct.xml: -i 10
rid3-rax-50-pct.xml: -i 10
rid3-rax-50-pct.xml: -i 10
0. provide viewJobs.jsp link under "Admin" to show list of portal jobs status
1. remove breadcrumb trails when viewing "Admin" pages (decorators/main.jsp)
2. fix problem of using symbolic link to view NexusFiles.xml file.
Replace it with PerformanceAction and performance.jsp files
April 20, 2007 (Friday)~
---------------------------------------------------------------------------------
<<<<<<< .mine
1. remove the "report an issue" and "request a feature" button and replace with
Ashton's links instead per Mark Miller's request;
=======
0. change availablememory = 2048 to
availablememory = 2056
1. remove the "report an issue" and "request a feature" button and replace with Ashton's
links instead per Mark Miller's request; Move links to the left of page instead of
center of page.
>>>>>>> .r3607
2. Add Web Count which is a global variable for the web portal in -DgroupLabel
property when calling RunServiceApp() so backend can keep track of the job;
3. display intermediate results files for user to download.
4. send email to user of URL so they can view the intermediate results
April 23, 2007 (Monday)~
--------------------------------------------------------------------------------
0. build and install cipres on cipres14;
1. implemented sychronized method for webcount;
2. move admin* pages to /admin directory so secure page can be managed easily;
3. FileAction, ProcessRequestAction, add a "DEBUG" flag
set environment variable (DEBUG) in Eclipse; value can be anything (true, TRUE);
as long as this "DEBUG != null", FileAction and ProcessRequestAction will use
my cipres-framework-stub code for testing instead of the actual code;
TO DO:
+. webinfo.properties: webCount -> WebCount; reset value to "1" when reachese
Integer.MAX;
+. make job status table columns sortable and searchable; (maybe replace using mysql)
---------------------------------------------------------------------------------
April 24, 2007 (Tuesday), April 25, 2006 (Wednesday)
--------------------------------------------------------------------------------
1. separate NexusFiles.xml into NexusFiles.xml.(paup/raxml/garli) to make it
more readable;
2. add "DATATYPE" to mapping.xml for "DATATYPE" (DNA vs Amino Acid):
list of analysis depends on the uploaded nexus files's datatype;
list of tools for a selected analysis depends on the "tool" attribute
if RunNexusReader return datatype of:
NULL, or
anything other than "DNA_DATATYPE" or "AA_DATATYPE", set datatype= ANY_DATATYPE;
3. add "treeinfer" to mapping.xml to indicate which paup infer xml to use
(the web will pass paup_infer.xml if paup if chose, otherwise, quick_paup_infer.xml)
RunServiceApp() will ignore this parameter if the nexus file already has a
starting tree
4. add "timeout" to mapping.xml for each tool (in minutes);
Terri removed in all command files
5. update mapping.xsd for the abvove changes (add 3 required attributes)
6. Mark holder make the following changes to paup.xml:
I have made the PAUP command in the *paup*xml files in
cipres-web/src/main/resources/conf more thorough.
Specifically it is now:
Set maxtrees = 1000;
and in the context of ab initio tree searching (rather than improving
a current tree), the starting tree is coming from
Hsearch start=stepwise addseq=random nreps=200
I think that in the context of getting a starting tree for an ML run (e.g. using GARLI when the user has uploaded a file without a tree in it), then we shouldn't spend too much time getting a good parsimony tree. Thus, I have added a quick_paup_infer.xml file. This has the old, quick and dirty search settings and should be used when we are just getting a tree so that we can use GARLI.
April 30, 2007 (Monday)
----------------------------------------------------------------------------
+ update raxml to 2.2.3 from 2.2.1 (see Holder's email dated 04/24/2007):
/projects/cipres/bin/RaXMLHPC2.2.3
May 3, 2007 (Thursday)
-----------------------------------------------------------------------------
+ change the return MIME-type in DownFileAction to "text/plain" so end users won't be told they are
download a binary file;
+ change package name (from org.cipres.helpers to org.cipres.portal) for RunServiceApp and RunNexusReader
+ also change tryit script and Performance;
+ eanble recidcm3-raxml using for nexus files with at least 500 taxa:
+ 500 - 1000: 50% subproblem size
+ 1001 - 2000: 25% subproblem size
+ 2001 - 3000: 16% subproblem size
+ 3001 - 5000: 6.25% subproblem size
? replace to_nexus.py to philip_to_nexus to convert philip to nexus files;
+ 5001 + : 6.25% subproblem size