Activity for diCal2
-
Matthias committed [r81]
Minor clarification of output.
-
now throws a better error message when number of haplotypes in vcf is larger than that given in config-file.
-
diCal2 released /release/diCal2_2_0_5.tar.gz
-
Version update into readme.
-
Added an option to ignore double entries in the vcf.
-
Going back to storing fwd/bwd tables as doubles, because the loss of precision apparently breaks some cases.
-
hidden parameters now handled correctly
-
Apparently, latex needs some convincing to display the correct quotation marks in the verbatim environment.
-
Cosmetic changes for reading in vcfs. If no segregating sites left after filtering, an error is thrown. Minor bugfixes.
-
fwd & bwd table now TwoDimFloatArray instead of ArrayList(ArrayList(double[]))
-
fwd & bwd table now TwoDimDoubleArray instead of ArrayList(ArrayList(double[]))
-
HmmStepHandler implements the missing sites more memory efficient
-
reading vcf more memory efficent (not yet fully tested)
-
-
changed some command line parameters
-
manual is approaching a presentable state
-
deleted some files
-
updated manual & renamed example files
-
changed examples from readme & added new examples for manual
-
smole bugfixes & more manual
-
more manual
-
more manual
-
more manual
-
more manual
-
some stuff
-
some more manual
-
tiny bit more manual
-
more manual writing
-
started revamping the manual
-
small change of constants, but still doing weird stuff
-
Running the ODEs only until maxTime and taking the values did mess up the EM. Now, we adjusted the epsilons a bit and just throw errors, if it does not work.
-
ODE solver could, and in fringe cases did, run forever. Now there is a max time.
-
minor bug in coordinateOrder fixed
-
optimized M-steps (memoiziation)
-
ODE Core does not work properly with admixutre pulses yet.
-
-
Should now be able to read haploid vcf files as well.
-
before starting to implement haploid vcf reading
-
finished deleting old sequence format
-
removed old sequence format from examples
-
just some restructuring
-
-
The default behaviour is now to not accept unphased genotypes. If they should be accepted, you have to explicitly specify a flag.
-
if a vcf file is read, we now explicitly print out how many segregating sites have at least one missing allele. and warn about them being ignored in the analysis
-
fixed a bug where no new meta points where chossen, if the previous best one didn't have variation between them
-
vcfDiAllelic and TriAllelic removed, because they don't do anything anyways
-
minor bugfix
-
implemented option to given vcf-reference on command line. Flag for missing,fixed,random reference abandoned, because numLoci is kind of a problem here.
-
Implemented a flag to also show the hidden parameters.
-
no seed needed if no randomness involved seems to work
-
now we have a switch to chose between the cores
-
why do you always ignore the manual?
-
rearranged the parameters a bit more. hopefully nothing broken yet
-
apparently I forgot this one
-
reorganizing command line parameters
-
last commit before I break everything
-
start of a manual
-
-
multiple param files and bed files for the branch as well
-
now we can read multiple parameter files and bed files
-
hacked to have modified parameters
-
putting in some dependent doubles
-
hacking
-
bugfix
-
changed the reco and mut parameter objects to h...
-
some small bugfixes
-
commented the examples
-
-
the dot though
-
inserted preprint url
-
changed readme a bit again
-
new readme
-
no more build script
-
-
minor typo
-
I think something like the first release (finally)
-
fixed some unchecked conversions
-
fixed small issue with imports
-
modified scripts a bit
-
some building
-
rename
-
added license, build script, readme, and some e...
-
I think this is the working first release
-
whatevs eclipse
-
-
-
-
1