Revision history for Bioperl core modules
0.7 Large number of changes, including refactoring of the
Object system, new parsers, new functionality and
all round better system. Highlights are:
o Refactored root of inheritance: moved to a lightweight Bio::Root::RootI;
Bio::Root::IO for I/O and file/handle capabilities.
o Imported BPlite modules from Ian Korf for BLAST
parsing. This is considered the supported BLAST parser;
Bio::Tools::Blast.pm will eventually phase out due to lack of support.
o Improved Sequence Feature model. Added complete location
modelling (with fuzzy and compound locations). See
Bio::LocationI and the modules under Bio/Location. Added
support in Genbank/EMBL format parsing to completely parse
feature tables for complex locations.
o Moved special support for databanks etc to specialized modules under
Bio/Seq/. One of these supports very large sequences through
a temporary file as a backend.
o Explicit Gene, Transcript and Exon SeqFeature objects, supporting
CDS retrieval and exon shuffling.
o More parsers: Sim4, Genscan, MZEF, ESTScan, BPbl2seq, GFF
o Refactored Bio/DB/GenBank+GenPept. There is now also DB/SwissProt and
DB/GDB (the latter has platform-specific limitations).
o New analysis parser framework for HT sequence annotation (see
Bio::SeqAnalysisParserI and Bio::Factory::SeqAnalysisParserFactory)
o New Alignment IO framework
o New Index modules (Swissprot)
o New modules for running Blast within perl
(Bio::Tools::Run::StandAloneBlast). Added modules for running
Multiple Sequence Alignment tools ClustalW and TCoffee
(Bio::Tools::Run::Alignment).
o New Cookbook-style tutorial (see bptutorial.pl). Improved
documentation across the package.
o Much improved cross platform support. Many known incompatibilities
have been fixed; however, NT and Mac do not work across the entire
setup (see PLATFORMS).
o Many bug fixes, code restructuring, etc. Overall stability and
maintainability benefit a lot.
o A total of 957 automatic tests
0.6.2
There are very few functionality changes but a large
number of software improvements/bug fixes across the package.
o The EMBL/GenBank parsing are improved.
o The Swissprot reading is improved. Swissprot writing
is disabled as it doesn't work at all. This needs to
wait for 0.7 release
o BLAST reports with no hits are correctly parsed.
o Several other bugs of the BLAST parser (regular expressions, ...)
fixed.
o Old syntax calls have been replaced with more modern syntax
o Modules that did not work at all, in particular the Sim4
set have been removed
o Bio::SeqFeature::Generic and Bio::SeqFeature::FeaturePair
have improved compliance with interface specs and documentation
o Mailing list documentation updated throughout the distribution
o Most minor bug fixes have happened.
o The scripts in /examples now work and have the modern syntax
rather than the deprecated syntax
0.6.1 Sun April 2 2000
o Sequences can have Sequence Features attached to them
- The sequence features can be read from or written to
EMBL and GenBank style flat files
o Objects for Annotation, including References (but not
full medline abstracts), Database links and Comments are
provided
o A Species object to represent nodes on a taxonomy tree
is provided
o The ability to parse HMMER and Sim4 output has been added
o The Blast parsing has been improved, with better PSI-BLAST
support and better overall behaviour.
o Flat file indexed databases provide both random access
and sequential access to their component sequences.
o A CodonTable object has been written with all known
CodonTables accessible.
o A number of new lightweight analysis tools have been
added, such as molecular weight determination.
The 0.6 release also has improved software engineering
o The sequence objects have been rewritten, providing more
maintainable and easier to implement objects. These
objects are backwardly compatible with the 0.05.1 objects
o Many objects are defined in terms of interfaces and then
a Perl implementation has been provided. The interfaces
are found in the 'I' files (module names ending in 'I').
This means that it is possible to wrap C/CORBA/SQL access
as true "bioperl" objects, compatible with the rest of
bioperl.
o The SeqIO system has been overhauled to provide better
processing and perl-like automatic interpretation of <>
over arguments.
o Many more tests have been added (a total of 172 automatic
tests are now run before release).
0.05.1 Tue Jun 29 05:30:44 1999
- Central distribution now requires Perl 5.004. This was
done to get around 5.003-based problems in Bio/Index/*
and SimpleAlign.
- Various bug fixes in the Bio::Tools::Blast modules
including better exception handling and PSI-Blast
support. See Bio/Tools/Blast/CHANGES for more.
- Fixed the Parse mechanism in Seq.pm to use readseq.
Follow the instructions in README for how to install
it (basically, you have to edit Parse.pm).
- Improved documentation of Seq.pm, indicating where
objects are returned and where strings are returned.
- Fixed uninitialized warnings in Bio::Root::Object.pm
and Bio::Tools::SeqPattern.pm.
- Bug fixes for PR#s: 30,31,33-35,41,42,44,45,47-50,52.
0.05 Sun Apr 25 01:14:11 1999
- Bio::Tools::Blast modules have less memory problems
and faster parsing. Webblast uses LWP and supports
more functionality. See Bio/Tools/Blast/CHANGES for more.
- The Bio::SeqIO system has been started, moving the
sequence reformatting code out of the sequence object
- The Bio::Index:: system has been started, providing
generic index capabilities and specifically works for
Fasta formatted databases and EMBL .dat formatted
databases
- The Bio::DB:: system started, providing access to
databases, both via flat file + index (see above) and
via http to NCBI
- The scripts/ directory, where industrial strength scripts
are put has been started.
- Many changes - a better distribution all round.
1.3a1 Fixed and restored XBM export. Added a (almost empty) template menu
(the beginnings of which were already hidden in 1.2a8). Fixed kerning
of mixed sub- and superscripts (as in SO_4^2^-). Changed allocation of
initial drawing area, calculate actual size before saving.
1.3a2 Added option menu for default bond type , added label support to
template function
1.3a3 Reduced preallocated drawing area to 'windowsize+100', as the increase
in 1.3a1 (which used window+1000) caused a massive slowdown at startup.
Freed unneeded pixmaps in the button setup code.
Fragment rotation code now stores the initial coordinates and uses them
as basis for the new positions - the previous, incremental calculation
was severely plagued by accumulation of errors.
1.3a4 Changed bond option menu to use small icons instead of longer and less
clear text labels. Removed check for 'cursor inside marked region' in
'Move' mode to allow smooth dragging. Started populating the O and N
heterocycle panels of the template menu. Fixed a duplicate bond in the
tcdd example, and removed those examples that have become templates.
1.3a5 Added shortcut for some labels - pressing c,n,o,s,p,r,i,h,d or b
inserts the corresponding centered element symbol at the last
drawing position, 'l','1','2','3' insert the left justified labels
'Cl','CH','CH_2','CH_3'. Fixed a GC leak in export_bitmap.
1.3a6 Moved the nucleoside examples to the template system. Made marker
position default to last drawing position, so that the 'Add' function
is now always active. Beginnings of internationalization (.mo file
for German locale added). Fixed naphthalene template, added dicyclo-
pentadiene. Changed Add_atom to automatically convert blanks in label
input to backslashes for storage. Prepared a few drawings for later
inclusion in the template system (camphor,pteridine,caffeine,glucose,
fructose,mannose,galactose,neuraminic acid, spiro[4.5]decane).
1.3a7 Template cleanup: porphine moved to heterocycle panel, spirodecane
added in its place. Glc,Fru,Man,Gal transfered to carbohydrate
templates, caffeine to heterocycles. Czech translation (by Radek
Liboska) added. Improved bounding box coordinates for xfig compound
object (used for sub/superscripted labels) Changed export functions
to remove the intermediate '.f2l' files. Changed orientation in xfig
header to Portrait. Adapted xfig (and xfig-based) export modes to
the new XFig default depth of 100. Changed depth of xfig text (labels)
to 90 and added filled white polygons at depth 95 to hide overlapping
line segments (experimental, used only for sub/superscripted text at
the moment). Reset default filename to 'unnamed' when the current
molecule is deleted.
1.3a8 Changed buttons mark/move/rotate/flip/delete/copy from text to icon
to reclaim space for the text entry box (icons created by Radek
Liboska). Added white polygon to clear area under normal text as well.
1.3a9 Added program icon. Added an interface for Radek Liboskas standalone
program CHT, which calculates sum formula and molecular weight from
a chemtool sketch. Corrected example drawing of brevetoxin.
1.3a10 Really reset mark flag when nothing was within the marker box.
Allow saving only the marked region (formula weight computation only).
Helper program CHT now parses those labels that are not in its
internal table of common substituents; exact mass is now returned
automatically.Added status line with message history.
1.3a11 Internationalization support made optional (define DISABLE_NLS in
the Imakefile if you do not want it) to allow compiling on systems
without GNU-style locale support again.
Improved ring size and position in the benzene and cyclopentadiene
templates (Michael Banck). Reset all marks before loading a template.
If the template menu is already open, pressing the Template button now
brings it to the front. Added templates (inden,biphenyl), removed the
ugly question mark placeholders.
Changed handling of windowmanager events, so that using the window
decorations' CLOSE button brings up the Yes/No dialog while the
chemtool window is still alive.
Changes in helper program CHT (cht011a): now recognizes Ac in complex
substituents; correctly handles aromatic 'ring' in formula input;
checks input file for identifier "Chemtool"; reads number of bonds
from "bonds" line (to handle label-only files gracefully); exits on
unattached labels instead of silently miscalculating formula and
mass. Subsequently updated to Radek Liboskas current version 0.19,
which already includes equivalent changes. Added list of abbreviations
supported by cht.
1.3a12 Reset all marks before 'adding' a sketch. Adding a label on top of an
existing one now replaces the old label as it should. Added a function
for rescaling objects (useful for drawing reagents above the reaction
arrow - but labels are not rescaled yet). Added linetype 13, a single
bond with opaque background (for crossing bonds - see the bicyclooctane
template). Updated the German translation.
This - plus or minus some templates - is my release candidate for 1.3.
1.3a13 Dropped the internal icon, as it was only causing trouble, especially
with KDE. Most window managers allow specification of an external icon
anyway (added chemtool.xpm for this purpose). Moved 'delete fragment'
button to the end of the button row. Updated Czech translation by
Radek Liboska. Added linetype 13 to the helper program CHT, made it
accept any Chemtool 1.x file, not only version 1.2. Updated the
'documentation', i.e. the README file, split out an INSTALL file. Set
version number in chemtool files to 1.3. Fixed bug that left a newly
'added' structure active but unable to move. Rewrote positioning
logic for subscripts and right-justified text in export modes again.
Changed screen display of molecule from 'fixed' fonts to helvetica
as used in the export modes - previously, labels that appeared fine
on screen could overlap in the final (postscript) output. Fixed
'mannose' template that showed exactly this. Added formatting option
'|' for slanted characters (as in iPr,tBu).
1.3a14 Fixed label shortcuts for CH_2 and CH_3 that could crash chemtool
(buffer overrun). Fixed scaling in export function , which had been
applied to labels only. Improved label size at smallest scale. Added
correction factor for sub/superscript after certain characters (the joys
of proportionally-spaced fonts :-( ). Handled reopening of template menu
after a close via the window manager. Updated 'About' menu (mention Radek
Liboska as developer, add the tu-darmstadt webpage).
1.3a15 Changed canvas bitmap allocation to avoid uncovering garbled portions
when the sliders are used. Fixed fragment deletion code (deletion of
the marked fragment and redisplay of the modified drawing are now
performed immediately). This hopefully fixes the crashes some people
have been seeing with this function.
1.3 Added Check for no (or no selected) atoms before trying to delete a
fragment, fixed bug that could deletion of a fragment to fail.
Changed functions for horizontal and vertical flip to transform the
atom(s) directly bonded to the marked fragment as well. Added current
filename (if any) to window title. Use xfig's own arrow type in export
of regular arrows (Michael Banck).
Some code cleanup: removed dead code, fixed unnecessary parameter
passing between functions, tidied formatting with GNU indent. Removed
Imakefile and provided regular Makefile (the build process relies on
gtk-config, not xmkmf, since 1.2), added install targets for the
localization files and manpage. Updated cht to version 1.3.
ones to do, and each compiled and installed/de-installed apparently
correctly.
As a side effect of the dynamic PLIST, we no longer need to have separate
-static and -shared PLISTs. It's now easier than ever to make a perl5
package for NetBSD :)
proteins, nucleic acids and small molecules. The program is aimed at
display, teaching and generation of publication quality images.
Submitted by Marc Bauoin <babafou@babafou.eu.org> in pkg/8199, slightly
enhanced.
for DNA or proteins. This is a merge of the packages submitted in PRs
pkg/7075 by Brook Milligan and pkg/8094 by Marc Baudoin (with some minor
modifications), thanks!
