|
EMBOSS: Project Meeting (Mon 26th July 10)
|
Attendees
EBI:
Peter Rice,
Alan Bleasby,
Jon Ison,
Mahmut Uludag
Visitors:
Apologies:
1. Minutes of the last meeting
Minutes of the meeting of 19th July 2010 are
here.
2. Maintenance etc.
2.1 Applications
Peter noted an unreported bu in dbxfasta where the
description index has no test for the length of the term. A test has
been added to trim the word to the maximum index length both in
dbxfasta and in ajindex.c
Mahmut noted that for needle and other alignment
applications the number of gaps and total gap length are not reported.
2.2 Libraries
Peter has committed the changes for embossre.equ to be
stored and accessed in the REBASE data directory.
Peter has checked the image size definitions in
ajgraph.c for PDF and SVG outputs.
2.3 Other
Mahmut has tested a 6.3.0 SoapLab installation with EBI
external services. A few minor SoapLab issues were fixed.
Mahmut looked at Jemboss AJAX utilities and classes. We do not
use JNI for standalone or client Jemboss (JNI is only used by Jemboss
servers using authorisation). We can rename the parent "AjaxUtil"
class to "Util" and move the "Ajax" class to the 'server' package.
3. New developments
3.1 Dbx indexing
Peter has reorganized the internal handling of index fields so
that they can be extended for OBO format file indexing. Changes are in
ajnam.c to allow field-specific resource attributes for the
pagesize and cachesize, in ajindex.c to define the default
index file names and index types for the common field names, and in
embindex.c to use the new defaults when creating new index
caches and files for dbx indexing applications.
Peter will install the latest release of EMBL for testing DBX
indexing.
3.2 Ensembl
Alan has proposed fixes for problems with Ensembl sequence
retrieval. The first is that the regular expression for the Ensembl
software version and database fails for EBI's mysql.ebi.ac.uk server
as it picks the wrong field for the version and species. This can be
corrected with manual parsing, but still fails to find entries. The
naming of entries is species-specific, with an ENSG prefix for human
genes, and SPAC2f7-1.1 for S. pombe (as an SPO 3-letter prefix). Michael
Schuster has been asked about the best way to fix the problem.
The second problem is SQL-related warnings that repository information
has been already read.
When entret is used to return the original text, some output
should be provided for Ensembl data. We need to decide on useful
content to return, perhaps as a FASTA sequence.
4. Administration
Alan has released 6.3.1 with the fixes to file open calls. The
EMBASSY packages have been redone with fixes to configure files for the
qualifier used for detecting the PDF library used by plplot.
On FreeBSD some further configure changes were required to define the
library location.
Alan now has a FreeBSD installation for testing.
5. Documentation and Training
5.1 Books
Jon has sent the final book to the publishers.
Jon has set up new directories and files for the EMBOSS website
based on the books.
This will replace the current CVS directory structure, with separate
trees for HTML and XML and a new web home page. Each directory will
have a readme file explaining the expected contents.
Some additions have been made to the manuals to clean up the develop
and user guide appendix with tables of reference information for
libraries and applications, now also covering the EMBASSY packages.
The GCG/EMBOSS application comparison table has been added to the user
guide appendix
6. User queries and answers
All handled.
7. AOB
None.
8. Date Of Next Meeting
The next meeting will be on Monday 2nd August. Peter will be on vacation.