BIRCH - To Do List (April 7, 2010)
Short term/ high priority
- BIRCH
- put something into local/admin/birch.properties to tell
birchconfig if a birch system is a development platform (eg. psgendb,
albacore etc.) Just as a safeguard against clobbering the master copy
of bin-xxx-xxx_D.tar.gz)
- Need to do a better job at making sure that BIRCH_PLATFORM is
set
properly during an install. A combination of better platform detection,
and some pop-up messages, might help. The simple measure of changing
the platform chooser from a combobox to a scrolling chooser might be
enough.
- We really need to migrate birchconfig to NetBeans.
- Add documentation on upgrading from linux x86 to x86_64.
(Change
platform.source, platform.profile.source, and BIRCH.properties, plus
local bin and lib directories). Need a short statement that linux-intel
can be run on linux-x86_64 system.
- Currently, BIRCH must be completely uninstalled in order
to
update to a new version. Is there a way to update a 'live' BIRCH
installation, so that there is no downtime? This is particularly
critical if people are going to run long-running jobs.
- Change format of BIRCH Administrator's email address so that it
isn't a spam target.
- Replace heapsortHGL with a separate sort program
- Need better documentation for the fact that BIRCH changes the
command prompt. Should we set an environment variable to let the user
turn this off? This may be best to do in a $HOME/.birchrc file.
- Documentation pages could have an icon indicating whether a
program or package is part of the BIRCH core or locally-installed.
- For local.cshrc.source and local.profile.source: option for
shell
to recognize BACKSPACE key.
- On some systems, mv is aliased to mv -i. This will cause some
scripts to fail if they try to overwrite an already existing
file, because the script will try to prompt the user. The
problem has been fixed in dnadist.csh and protdist.csh, but
other scripts, and the .GDEmenus files, need to be checked to
see which ones might be susceptible to this problem.
- get programs to recognize new amino acid symbols: J = Leu/Ile =
Xle; O = Pyl = pyrrolysine. In particular, the following programs
packages need to be looked at:
- readseq
- bioLegato
- FSAP
- XYLEM
- GNOME, KDE menus
- rethink strategy for local copies of PIR/SwissProt
- add newuser and nobirch to birchdb (write man pages)
- miniBIRCH: let's see if we can shrink the size of the .tar.gz
file by reorganization of $doc or $dat directories.
- complete set of pattern discovery/pattern matching programs
- better support for genomic sequences of arbitrary length
- need a more sophisticated usage logging mechanism
- $BIRCH_TMP points to a directory that can be used by scripts
for
temporary files, usually /usr/local/tmp. We probably need a standalone
script that can be called by other scripts to see if this directory is
available, and if so, to use it.
- It probably isn't hard to add code that would let each user
have
their own cshrc.source or profile.source files that could override the
local BIRCH system. But, it there any advantage to doing this, versus
just telling them to modify .cshrc or .profile?
- Forking bioLegato and GDE:
- Move GDE environment variables for helper applications to
BIRCH.properties (how do we handle platform-specific choices?)
- change birchdb documentation links from $tutorials/GDE to
$tutorials/bioLegato
- change birchdb interface items from 'GDE' to 'biolegato' etc.
Note: this may have implications for data items in lbirchdb.
- Need a better way to check for the presence of critical things
like text editors, viewers etc during a new BIRCH installation, and set
those to the defaults. Ideas:
- birchconfig should have a wizard for testing for presence of
these programs and testing them. The final results get stored in the
Birch.properties file.
- Change the admin/profile.source type files to read the value
of
variables such as GDE_TEXTEDIT from Birch.properties, rather than
having them hard-coded. We would have to figure out how to work around
the existing lines in local/admin files.
- copy dat/GDE etc. to dat/bioLegato etc.
- modify scripts
- modify .GDEmenus in order: tbiolegato, mbiolegato, biolegato
- Rewrite biolegato and GDE scripts to reflect the fact that
bioLegato is now the default.
- bioLegato
- Import Free Format (FSAP) - full path is used as name; should
be
base name.
- For programs with no GDE Help file, create a dummy help file
saying that no help is available.
- osx-x86_64
- Test on other Macs
- Are there things that could or should be done to make BIRCH
behave more like something Mac users are used to?
- ACEDB
- Some linux-x86_64 systems don't have a 32-bit libncurses.so.5.
This is required by xace. The quick fix for now is to include the
32-bit libncurses.so.5 in lib-linux-x86_64. The right way to fix this
is to recompile xace as a 64-bit binary.
- FASTA
- FASTA search on GenPept fails to extract names and gi numbers.
- Fasta truncates sequence names if they're too long. May need to
use uniqid.py to encode and decode names.
- add -N # option for fasta fasta/nucleotide. Also -Z
option.
- XYLEM
- Splitdb in biolegato can result in runaway jobs if run without
data.
- getob - DBLINK line overwrites DEFINITION line in message file
- check for new Feature Keys
- splitdb can sometimes enter an infinite loop when
called
from feature
- NCBI
- Update Cn3D and add CDTree
- PHYLIP
- update to most recent version
- add 'gendist' to mbiolegato
- All programs that do resampling should be able to save the raw
tree file
- GETTREES in tree2flat.py needs to be implemented more
efficiently
for translating very large treefiles
- output to TGIF using pstoedit command
- dGDE
- SeqHound at www.blueprint.org is no longer functional. Need to
replace dgde SeqHound calls with calls to Entrez or other data
warehouses.
- expanded functions
- tacg - the
latest version gives compilation errors, and is no longer supported, so
we need a replacement (ApeE?)
- Databases: standard
distribution
- get XYLEM programs to work with remote database services
- Databases: add-ons
- specific documentation for
- local BLAST database installation
- local FASTA dababase installation
GDE legacy issues:
- GDE has several issues on recent Linux releases that prevent it
from running.
- Fedora11 - xset/font problem - won't launch
- Ubuntu 9.04 - launches, but entire X11 desktop freezes
Long term
- BIRCH collaborations?
- EMBOSS - complete biolegato menus
- Eric Linton, MSU: macGDE
- New format to generate bioLegato menu items eg. PISE, EMBOSS ACD?
- Python - The following scripts need to be revised to eliminate
deprecated code that will not be supported in Python 3.0. Running a
script with 'python -3' will generate messages showing which constructs
are deprecated. As well, some stylistic changes should be made, such as
eliminating use of 'string' methods, and switching to 'str', as well as
using a more modern approach of iterating through a file ie. 'for line
in file :'.
- customdoc.py
- htmldoc.py
- BLExtractSubset.py
- csh2sh.py
- dbsout.py
- flat2list.py
- flat2tree.py
- flatcnv.py
- list2flat.py
- tree2flat.py
- uniqid.py
Please send suggestions of comments regarding this
page to psgendb@cc.umanitoba.ca