FrequencyModel parser now writes out what it is doing
114 lines of code changed in 1 file:
fix bug (primates_twoFossil.xml)
3 lines of code changed in 1 file:
Added native (fortran) routines to approximate the exponential of huge sparse matrices. These routines will be used in random-walk on maps models, where the order of a 100 x 100 map = 10,000 and it is impractical to exponentiate the 10,000 x 10,000 rate matrix exactly. These approximations come from ExpoKit (ok to modify/distribute for non-commercial purposes) and use a Krylov-subspace approach. Also included is a good Pade approximator (which we may want to port to Java/C and use for the character substitution rate matrices). I plan to commit a working random-walk on the San Francisco Bay Area map example soon.
4887 lines of code changed in 8 files:
sortTranslationTable now defaults to true.
1 lines of code changed in 1 file:
Fixed problem logging trees MCMCMC (was giving null taxon labels).
25 lines of code changed in 3 files:
Added UI for doing marginal likelihood analyses. Actual call to analysis code is not implemented yet. May generalize this for calculating BFs between traces?
3 lines of code changed in 1 file:
Allow # comments at start of log files...
18 lines of code changed in 2 files:
tweaks
26 lines of code changed in 7 files:
[FIX] Allow 'ant' to build beast on linux.
* build.xml: add lib/mtj.jar to beast classpath
* build_development.xml: ditto.
95 lines of code changed in 2 files:
Added UI for doing marginal likelihood analyses. Actual call to analysis code is not implemented yet. May generalize this for calculating BFs between traces?
211 lines of code changed in 4 files:
Many fixes to Tracer - progressing to v1.4 release
396 lines of code changed in 15 files:
8 lines of code changed in 4 files:
FigTree tree now has mean heights
0 lines of code changed in 1 file:
More tweaking...
122 lines of code changed in 2 files:
Fixed up the practice section according to Philippe's suggestions
190 lines of code changed in 23 files:
Possibly fixed problem with swapping loggers during MCMCMC. This compiles but is not tested - have run out of time.
115 lines of code changed in 2 files:
added new move to help mixing in RLC models... might have to use MCMCMC though...
212 lines of code changed in 3 files:
0 lines of code changed in 1 file:
FIrst-pass hack at implementing a weighted random walk on a d-2 Markov lattice. Allows for discretized diffusion on a terrain (topographic) map for use with the "coyote" (errr cougar) data. The MapMaker class currently requires mtj.jar (optimized sparse matrix operations) to construct the eigensystem of the weighted lattice.
1046 lines of code changed in 9 files:
Cleaned up output of XML documentation generation for MediaWiki. This now generates a file which can be pasted into the BEAST wiki. See http://beast.bio.ed.ac.uk/BEAST_XML_Reference for current version
28 lines of code changed in 4 files:
Added F81. Fixed some references to existing papers.
518 lines of code changed in 11 files:
Adding a new figure
0 lines of code changed in 6 files:
More edits to BookSection
130 lines of code changed in 4 files:
Added the ability to create MediaWiki output to BEASTDoc - not beautiful yet.
326 lines of code changed in 11 files:
many small grammatical fixes -- there was a big merge conflict because of different formats for software names -- I thought the bold was a bit intrusive.
212 lines of code changed in 1 file:
Cleaned up the BookSection a bit to make the formatting more consistent, use the more correct term 'dengue virus serotype 4', 'E' (not env) gene. Teletype style for UI components. Bold style for programs. Resized figures. Few typos.
78 lines of code changed in 1 file:
I added the basic guff about Bayesian MCMC and what it is all about. The only thing missing is a proper discussion of the important role of priors...
96 lines of code changed in 1 file:
Finished up the text of the practical section. My latex is still rusty so there may be glitches.
82 lines of code changed in 1 file:
Added a few more synchronisations to allow multithreading.
16 lines of code changed in 1 file:
1 lines of code changed in 3 files:
Why the coalescent likelihood we currently have is right :-) (of course!)
118 lines of code changed in 2 files:
Added a few more synchronisations to allow multithreading. This includes GammaSiteModel & HKY. Basically, any lazy calculations for instance variables in functions called by TreeLikelihood need to be synchronized.
771 lines of code changed in 6 files:
implemented multi-threading support for AbstractSubstitutionModel in aid of AR's work on multithreading likelihood calculations.
1050 lines of code changed in 2 files:
An experiment: Implemented multithreading in CompoundLikelihood. The idea is that all the contained likelihoods are calculated concurrently on multicore machines. This may work with multiple treelikelihoods (i.e., data partitions). To turn on, add the attribute threads="x" to the compoundLikelihood (including "posterior", "prior" and "likelihood" aliases) where x is the number of threads to use (match number of cores or number of likelihoods?) omit for no threads. This uses a thread pool implemented in ExecutorService to run the threads but spawning that many threads (i.e., each state) may have too much overhead unless there are a lot of patterns.
180 lines of code changed in 3 files:
Finished a basic implementation of the GMRF coalescent smoother. This implementation currently uses a random walk operator on the population sizes and a Gibbs update of the precision parameter. A (much) better/faster approach (to come shortly) is a joint update using Rue's GMRF library.
348 lines of code changed in 4 files:
added a boolean attribute to the tree logger to optionally sort the translation table in BEAST tree files so that AWTY can be more easily used to compare independent runs...
257 lines of code changed in 2 files:
Full evaluation wasn't being set properly.
25 lines of code changed in 2 files:
Have fixed NewCoalescentLikelihood. Have replaced it as the default coalescent likelihood in BeastParser (because changes to CoalescentLikelihood seem to have broken it). Preliminary indications are that it is faster and the code is cleaner.
179 lines of code changed in 6 files:
Have made 'fullEvaluationCount' an attribute of MCMC. If not specified then it has a default of 2000 (this is now an option in MCMCOptions). Set to chain length to fully evaluate the likelihood after every restore.
27 lines of code changed in 4 files:
little utility to order translation table in Nexus files so that AWTY can be used...
131 lines of code changed in 3 files:
Improved the update messages being passed from DiscretizedBranchRates to TreeLikelihood so that only those nodes that need updating do so. This shows a significant improvement in speed and gives the same sequence of likelihoods with the test XML scripts for a given seed.
32 lines of code changed in 2 files:
Added parser error when the rate category parameter is not of dimension nodeCount - 1.
4 lines of code changed in 1 file:
Split TreeModel XML parser into separate class because TreeModel is a monolithic class. Fixed a labeling problem introduced into CoalescentLikelihood. However the CoalescentLikelihood now returns weird probabilities after jheled's last commit -- so this still needs looking into.
337 lines of code changed in 11 files:
200 lines of code changed in 2 files:
Attempted to fix how traits are re-created in TreeModel. testMultivariateTraits.xml is a quick demo of the multivariate Comparative Methods tools.
156 lines of code changed in 3 files:
oh well not quite yet...
5 lines of code changed in 2 files:
speed improvements to relaxed clocks (~20% faster to uncorrelated models) - important changes for porting are DiscretizedBranchRates and TreeLikelihood
93 lines of code changed in 7 files:
I hope I didn't just break TreeModel
1169 lines of code changed in 1 file:
528 lines of code changed in 3 files:
Added the classes below to implement the multivariate Comparative Method. Unfortunately, MultivariateTraitLikelihood no longer works with the current TreeModel with a HashMap of traitParameters.
MultivariateDiffusionModel
MultivariateTraitLikelihood
MultivariateDistributionLikelihood
InternalTraitGibbsOperator
PrecisionMatrixGibbsOperator
MultivariateDistribution
MultivariateNormalDistribution
WishartDistribution
CorrelationStatistic
2278 lines of code changed in 14 files:
84 lines of code changed in 1 file:
tidyup and generifying
64 lines of code changed in 1 file:
tidyup and generifying
231 lines of code changed in 26 files:
Updated practical section
307 lines of code changed in 4 files:
Moved BookSection into a folder with all its figures
782 lines of code changed in 16 files:
Updated BookSection again
244 lines of code changed in 1 file:
tidying
17 lines of code changed in 1 file:
Updated book chapter
88 lines of code changed in 1 file:
fix a bug which caused a crash when two coalecsent events are merged. Also add functionality for obtaining every coalecsent event as a separate event.
29 lines of code changed in 1 file:
export did not work
7 lines of code changed in 1 file:
346 lines of code changed in 3 files:
Random Local Clocks are on there way...
151 lines of code changed in 3 files:
fix calculation of indications/pops number relative to trees
46 lines of code changed in 1 file:
add "method" attribute to xml rules
3 lines of code changed in 1 file:
added ability to specify what posterior probability cutoff to output clades at, and also output mean height of clades in tree trace analysis
230 lines of code changed in 4 files:
fixed url for Tracer, title and acknowledgements
5 lines of code changed in 3 files:
and bib file...
9751 lines of code changed in 1 file:
final draft of manuscript?
30 lines of code changed in 1 file:
Issue a warning for attributes of elements in BEAST xml files which are not declared by the XML rules. This helps both users (like me) not to get stung by silent ignore of typos, and developers (like me) which need to debug rules for new elements.
Those changes involves fixing some elements rules, and additional cleanup.
327 lines of code changed in 27 files:
Updated manuscript and manual and cleaned up a bit of lint...
1926 lines of code changed in 11 files: