October 2006 Commit Log

Number of Commits:
37
Number of Active Developers:
3
rambaut 2006-10-31 03:44 Rev.: 63

Fixed problem with LogCombiner not working with newer BEAST output (which stores the lnP as a comment next to each tree).

11 lines of code changed in 1 file:

rambaut 2006-10-30 09:46 Rev.: 62

Fixes to BEAUti - loads and saves taxon sets in template files.

527 lines of code changed in 5 files:

rambaut 2006-10-27 03:27 Rev.: 61

Some final fixes for BEAUti. BeautiTester generates over 2700 combinations of models using the same generating code as BEAUti. All of these now start in BEAST successfully (can't promise they will produce good results).

538 lines of code changed in 3 files:

rambaut 2006-10-25 02:55 Rev.: 60

Some final fixes for BEAUti. BeautiTester generates over 2700 combinations of models using the same generating code as BEAUti. All of these now start in BEAST successfully (can't promise they will produce good results).

163 lines of code changed in 6 files:

rambaut 2006-10-25 01:26 Rev.: 59

The RuntimeException thrown when there was a zero likelihood as the first state was not being caught because it was in a different thread. The initial likelihood is now evaluated when the MCMC class is created in the MCMC parser and the exception is thrown then. This results in nicer error messages.

1 lines of code changed in 1 file:

rambaut 2006-10-25 01:12 Rev.: 58

The RuntimeException thrown when there was a zero likelihood as the first state was not being caught because it was in a different thread. The initial likelihood is now evaluated when the MCMC class is created in the MCMC parser and the exception is thrown then. This results in nicer error messages.

18 lines of code changed in 2 files:

rambaut 2006-10-24 23:23 Rev.: 57

Solved the problem of the missing integrator by copying out of the relevant source code. The numericalMethods library is GPL'ed.

72 lines of code changed in 1 file:

rambaut 2006-10-24 20:56

Fixes to BEAUti - loads and saves taxon sets in template files.

475 lines of code changed in 8 files:

alexei.drummond 2006-10-24 11:07 Rev.: 54

Removed line based on this commentary from Ben Redelings:
The "if (...) break;" statement should not be there, I think. In some of my examples, the autocorrelation generally decreases as a function of the lag, but not completely monotonically. Stopping as soon as you see an increase can result in ignoring a significant portion of the positive autocorrelation. This is probably due to estimation error for gammaStat[lag] but this can happen when gammaStat[lag] is still pretty high, so that rounding to zero (e.g. 'break') seems like it would not give very conservative (i.e. low enough) estimates of the ESS.

6 lines of code changed in 1 file:

rambaut 2006-10-24 03:35 Rev.: 53

BEAUti fixes and added the SRD06 model as a one-click setting.

643 lines of code changed in 8 files:

rambaut 2006-10-24 00:52 Rev.: 52

Fix for Issue 3. After some further static initialization issues, I returned the initial registering of the main DataType subclasses to DataType.java but introduced a lazy initialization.

34 lines of code changed in 7 files:

rambaut 2006-10-24 00:30 Rev.: 51

Fix for Issue 3.

34 lines of code changed in 7 files:

rambaut 2006-10-23 21:22 Rev.: 50

Fixed a small problem loading templates with the guess dates option on.

8 lines of code changed in 1 file:

rambaut 2006-10-21 01:10 Rev.: 49

Although the code to generate the birth-death prior is there, I have hidden the option from the drop-down tree prior menu until we can be sure it works.

323 lines of code changed in 1 file:

rambaut 2006-10-21 01:09 Rev.: 48

I think Yang & Rannala Birth Death priors are working now. testBirthDeathPrior.xml samples birth and death rates on a fixed tree simulated under b=1.0 d=0.5. When a boolean test is applied to keep b < d then it more or less samples the correct values. Without that addition prior, it samples b=d=0.75 approximately.

201 lines of code changed in 2 files:

rambaut 2006-10-21 00:57

I think Yang & Rannala Birth Death priors are working now. testBirthDeathPrior.xml samples birth and death rates on a fixed tree simulated under b=1.0 d=0.5. When a boolean test is applied to keep b < d then it more or less samples the correct values. Without that addition prior, it samples b=d=0.75 approximately.

461 lines of code changed in 7 files:

rambaut 2006-10-21 00:54 Rev.: 44

Fixes to BEAUti - loads and saves taxon sets in template files.

292 lines of code changed in 1 file:

rambaut 2006-10-20 21:57 Rev.: 43

Fixes to BEAUti - loads and saves taxon sets in template files.

2 lines of code changed in 1 file:

rambaut 2006-10-20 21:52 Rev.: 42

Fixes to BEAUti - loads and saves taxon sets in template files.

236 lines of code changed in 1 file:

alexei.drummond 2006-10-20 19:32 Rev.: 41

Somehow this got missed out in the last commit...

0 lines of code changed in 1 file:

alexei.drummond 2006-10-20 19:28 Rev.: 40

Added binary covarion model (two visible state (0, 1) and two hidden states with different rates). Also added some static code in DataType to handle the registration of datatypes. This simplifies and consolidates the boilerplate parsing code in AlignmentParser and FrequencyModel into a method in the newly added class DataTypeUtils. I also fixed StringAttributeRule to report the valid values in the case of a string attribute that has a set of valid values.

758 lines of code changed in 17 files:

rambaut 2006-10-19 03:34 Rev.: 39

I have added the ability to turn on the Yang & Rannala Birth-Death model but I am not convinced it is working. Setting deathrate=0 and samplingProportion=1 and it crashes (should get the Yule model). Have added testBirthDeath.xml & testYule.xml to the examples folder. The UI code to generate the model works but I will hide it if we can't get this working.

3 lines of code changed in 2 files:

rambaut 2006-10-19 03:19 Rev.: 38

I have added the ability to turn on the Yang & Rannala Birth-Death model but I am not convinced it is working. Setting deathrate=0 and samplingProportion=1 and it crashes (should get the Yule model). Have added testBirthDeath.xml & testYule.xml to the examples folder. The UI code to generate the model works but I will hide it if we can't get this working.

2344 lines of code changed in 6 files:

rambaut 2006-10-18 23:56 Rev.: 37

Updated some benchmarks in the benchmark.xml comments and fixed the Mac OS X native compile script

7 lines of code changed in 2 files:

rambaut 2006-10-17 02:59 Rev.: 36

Fixed up a few BEAUti issues and updated some benchmarks in the benchmark.xml comments.

52 lines of code changed in 4 files:

rambaut 2006-10-16 21:40 Rev.: 35

Fixed up example & test files

204 lines of code changed in 6 files:

rambaut 2006-10-14 02:34 Rev.: 34

Added appropriate bylines to all the programs and updated the test and example files

673 lines of code changed in 13 files:

rambaut 2006-10-14 02:16

AboutBox can display HTML. Fixed problem with quit/save confirmation being displayed when no windows are open on a Mac.

47 lines of code changed in 7 files:

sidneym 2006-10-14 00:56 Rev.: 31

remove some compiler warnings introduced in last few checkins

10 lines of code changed in 3 files:

rambaut 2006-10-13 21:32 Rev.: 30

9 lines of code changed in 2 files:

rambaut 2006-10-13 21:10 Rev.: 29

Fixes to BEAUti - loads and saves taxon sets in template files.

23 lines of code changed in 2 files:

rambaut 2006-10-13 10:39

Fixes to TreeAnnotator. Now summarizes rates.

495 lines of code changed in 3 files:

rambaut 2006-10-13 03:22 Rev.: 26

Fixed tree logging to incorporate the rate as a metacomment.

3 lines of code changed in 2 files:

rambaut 2006-10-13 03:19

More cleaning up and testing in BEAUti 1.4.

62 lines of code changed in 10 files:

rambaut 2006-10-13 00:03 Rev.: 23

Changed the mode of operation of BEAUti. It is now a MultiDocApplication which means you can have multiple BEAUti windows open. The general course of operation is New, Import NEXUS, optional Apply Template File, optional Save Template File, Generate BEAST File. The file menus have been customized to reflect this.

481 lines of code changed in 7 files:

rambaut 2006-10-12 22:53 Rev.: 22

Updated the "compileNativeMacOSX.sh" script to create a fat dylib that runs on both intel & PPC macs.

12 lines of code changed in 1 file:

rambaut 2006-10-12 21:49 Rev.: 21

This is our native library that helps speed up likelihood calculations for nucleotide sequences. If the compiled JNI dynamic library is found on BEAST start up then it will be used otherwise a pure Java version will be used. The scripts contain the commands to compile the library on Mac & Linux (I have no idea how to do it on Windows). Ideally we would produce a "configure" script so that users can compile a version optimized for their machine.

397 lines of code changed in 4 files:

Generated by StatSVN 0.7.0