EMBOSS: Project Meeting (Mon 12th April 10)
1. Minutes of the last meeting
Minutes of the meeting of 29th March 2010 are
2. Maintenance etc.
Peter will look into a user request for polydot to
display matching in both orientations as lines in opposite
directions. It is not clear whether this should be on by default as
short matches may appear as points.
Alan raised the issue of changes in MIRA 3.0 where options
change according to the specified sequencing technology. Applications
can be created for single technologies, but it is more complicated to
cover cases of mixed input data. Peter will check through and consider
the ACD implications.
New patches of EMBASSY packages are needed where the documentation
fails to build, a problem reported for MIRA. Peter will check
through the builds to identify any others.
Mahmut will try the word matching performance improvements
(using Rabin-Karp) on supermatcher.
Alan has added Microsoft versions for the new ajsys.c
functions that replace bare system calls. This required an alternative
structure for the AjPFile object as the Unix pid does not exist on
Windows, and the Microsoft handles do not exist on
Unix. The embossdatadoc script will need extending to allow
this kind of ifdef and to document the pid and handles as being system
-specific. This is not urgent as the Unix documentation is correct as is.
Alan pointed out that there are duplicate string (suffix S) and
char* (suffix C) functions. Peter will check and make the
string function call the char* version so we have one copy of each
function. The string version was written first and should have been
converted rather than duplicated.
A few system calls remain, for example in ajfile.c where there
are dependencies on global variables for data files and a waitpid call
on closing a pipe file. In ajseqdb.c there are signals and
timeouts for Microsoft where the signal is set in one function and
released in another.
Alan has added a new version of ajFileNewInPipe for
Alan will check on home directory functions on Microsoft
systems to provide the equivalent of parsing ~username on Unix.
Peter will look into scalable graphics options. A user has
requested PDF output. SVG is another possibility which plplot
may support. There was a second user request for scalable PNG files
which may be possible now that the graphics scaling code has been
cleaned up. The default size PNG files have too few pixels for fine
detail. A general graphics options could provide a scale factor for
fixed-size graphics outputs. The graphics output format definitions
could include a flag to determine whether this option applies to a
particular output device.
Client libraries using gsoap and axis1 have been tested.
Mahmut reported on tests by External Services when defining an
extra type in the SoapLab typed interface. In C++ clients the
behaviour of gsoap causes problems where the minoccurs and maxoccurs
XSD attributes are defined but have their default value of "1".
Mahmut discussed problems with the icon size generated for
Jemboss by java webstart, The rectangular icon is reduced to an
unrecognizable 32 by 32 square. Tim will be consulted on adding a new
32 by 32 icon for webstart. The current icon will be retained for web
pages and other uses.
Peter has looked at Galaxy definitions for EMBOSS
applications, It seems to be simple to generate. He will try building
an automatic version for the CVS code and comparing to the EMBOSS
5.0.0 definitions in the Galaxy distribution.
3. New developments
3.1 BioMart access
Alan is waiting for stub code to include BioMart as a database
access method. Peter will add BioMart as a query and entry
method using the example code from martquery.
3.1 DAS access
Peter attended the DAS workshop last week on campus. The new
DAS 1.6 specification is not yet finalized. We can look into support
for DAS 1.53 as well, in case DAS servers are slow to move to the new
standard. DAS 1.6 included a query for the version number of the
server so we can identify 1.6 and assume earlier versions are 1.53.
Mahmut has looked briefly into the coding requirements for a C
DAS client. There were no other developers at the workshop with any
code to offer.
Mahmut reported that the BioCatalogue team inquired about
possible support for SAWSDL annotations using EDAM. We can make
SAWSDL-annotated SoapLab services under new URLs for testing. The CVS
ACD files have been parsed to test extraction of annotation for
Alan will remind the systems group that we are still waiting
for a test server configuration.
5. Documentation and Training
Jon is away this week.
Peter plans to test all URLs in the book text and report any
that are not yet available. We can populate these before publication.
Peter has updated the embossdoc and embossdatadoc
scripts to enforce 4 spaces between functions, and added a script to
reset to 4 lines between functions and purge any trailing white space
in the source code.
Peter has started on generating XML documentation for library
source code. This can be completed when Jon is back.
Alan reported on a request from Dave Judge who had issues with
a web proxy that required a password. He found a workaround this
time. We should implement passwords in the proxy handling. Alan
will set up an internal proxy for testing.
Mahmut has been invited to teach on a course in Turkey in the
summer. No further details as yet.
6. User queries and answers
All outstanding queries put on the Sourceforge tracker.
8. Date Of Next Meeting
The next meeting will be on Monday 19th April.