R package to generate the same BEAST2 XML parameter files as generated by BEAUti 2
Home Page: https://docs.ropensci.org/beautier
License: GNU General Public License v3.0
This repository has been archived. The former README is now in README-NOT.md.
When using 2 MRCA priors (#25):
File: bad.xml seed: 1520947802409 threads: 1
Loading package BEAST v2.4.8
Loading package BEAST v2.4.8
Error 104 parsing the xml input file
IDs should be unique. Duplicate id 'clockRate.c:anthus_aco' found
Error detected about here:
<beast>
<run id='mcmc' spec='MCMC'>
<distribution id='posterior' spec='util.CompoundDistribution'>
<distribution id='likelihood' spec='util.CompoundDistribution'>
<distribution id='treeLikelihood.anthus_aco' spec='ThreadedTreeLikelihood'>
<branchRateModel id='StrictClock.c:anthus_aco' spec='beast.evolution.branchratemodel.StrictClockModel'>
<parameter id='clockRate.c:anthus_aco' name='clock.rate'>
Rename initial_phylogenies to initial_species_tree
That way, they are not code-cov'd (thus run twice)
Feedback bjoelle:
mrca_prior_to_xml_tracelog,mrca_prior_to_xml_operators,mrca_prior_to_xml_prior_distr,mrca_prior_to_xml_state: these functions only add the parts relative to the strict clock if the MRCA prior is monophyletic, whereas in BEAUTi these parts are added if the MRCA prior has a distribution associated with it, whether it is monophyletic or not.
Now all is hard-coded.
Helps fixing #23
Will result in v1.12 according to the road map
This package attempts to cover the very large set of parameter values (and their permutations) that can be set for a BEAST2 run. As one might expect, it is not capable of specifying all possible BEAST2 runs, and I appreciate the clear statement about functionality that is currently missing in the package's README. I did not a number of missing use-cases that are not currently mentioned in the README.
Moved to #69.
Moved to #70.
Both the user-interface and the code for specifying distributions is very complex, and I am concerned about how easily this package will be to maintain. @bjoelle has made some suggestions about a more modular approach to the internal code and possibly grouping like-functions into files. I think this would be a good idea, I also wonder if you could also consider a more high-level approach to specifying these models in user-exposed code. For instance, how much power/flexibility would be lost by losing the create_*_param() functions and instead doing something like this:
node_calibration <- create_gamma_distr(alpha=1, beta=10, estimate=FALSE)
I appreciate there may be reasons for taking the approach you have, but the current does make for a complex API. Whatever the API, some narrative examples
of setting up these paramaters in the documentation would be helpful.
Moved to #71.
The function create_beast2_input sets the user's global scipen option. It's not clear from the code exactly where this value is needed, but it would seem a better idea to either store the existing value (old_scipen <- getOption("scipen")) then ensure the value is reset using on.exit.
Moved to #72
Made an issue at rOpenSci
For starters...
setwd(paste(getwd(), "vignettes", sep = "/"))
knit("demo.Rmd")
markdownToHTML('demo.md', 'demo.html', options=c("use_xhml"))
system("pandoc -s demo.html -o demo.pdf")
Feedback bjoelle:
is_distr: this function just checks if the input has a name attribute, which doesn't match with the name or documentation of the function
Feedback bjoelle:
creating a TN93 model with the default settings creates two log-normal distributions as priors for kappa1 and kappa2 where the S parameter has
lower=0,upper=0andvalue=1.25. This works if S is not estimated, because BEAST2 only checks upper and lower when modifying the value (not when initialising it) but it will break if S is estimated.
The structure is my main concern with this package, as it contains a lot of code duplication and many functions need to run through all possible types of a certain object to figure out what to do with it. I think this makes the package hard to maintain and to extend to additional models.
The main issue in my opinion is that the objects (site models, clock models, etc) do not store their dependencies in a generic way, so that there is no way to know what objects a model depends on unless you know which specific model it is. My suggestion would be to add a list of components to all objects - for instance a TN93 site model would have components = c("kappa_1", "kappa_2"), a HKY would have components = "kappa" and so on. This would allow you to save a lot of code in functions like init_site_models, which could do things such as for(x in model$components) model[[paste0(x,"_param")]] = init_param(model[[paste0(x,"_param")]]) instead of needing one subfunction per type of model.
Moved to #38
is_%sth%_param(x) can be replaced by one function is_named_param(x, name) and the parameter_to_xml functions also have a lot of redundancy.Moved to #39
mrca_prior_to_xml functions.Moved to #40
get_id, get_ids: These functions appear to be duplicated by get_alignment_id and get_alignment_ids.Moved to #41
are_site_models, is_site_model, is_site_model_name and get_site_model_names) in one file would make the package easier to read and search in my opinion.Moved to #42
clock_models_to_xml_state l24-25: the comment is unclear as to whether it is supposed to be checking for any or no MRCA priors, and no check is performed here for either.Moved to #43
mrca_prior_to_xml_tracelog, mrca_prior_to_xml_operators, mrca_prior_to_xml_prior_distr, mrca_prior_to_xml_state: these functions only add the parts relative to the strict clock if the MRCA prior is monophyletic, whereas in BEAUTi these parts are added if the MRCA prior has a distribution associated with it, whether it is monophyletic or not.Moved to #44
estimate = TRUE, but the xml will not be created correctly with those settings (missing distribution, operators and state sections). I think the code should not allow this setting until this functionality is fully implemented.Moved to #45
lower=0, upper=0 and value=1.25. This works if S is not estimated, because BEAST2 only checks upper and lower when modifying the value (not when initialising it) but it will break if S is estimated.Moved to #46.
get_phylo_crown_age will not work with sampling through time/non-ultrametric trees. I think these are not supported at the moment anyway, but it might be good to keep it in mind.Moved to #47.
is_index_of_first_shared_clock_model: this function will return TRUE if the tested model is the 2nd of 3 shared clock models, in contradiction with the doc. It will also always return TRUE if the clock model at index i has id=NA.Moved to #48.
is_distr: this function just checks if the input has a name attribute, which doesn't match with the name or documentation of the functionMoved to #49.
get_site_model_n_params and get_site_model_n_distrs do not check if the site model has gamma rate categories, they just assume it doesMoved to #50.
kappa_1, kappa_2 and the tree prior parameters are always estimated and this cannot be changed - I think this should be mentioned in the corresponding doc.Moved to #51.
Moved to #52.
Moved to #53.
Moved to #54.
Moved to #55.
Moved to #56.
@return say this creates a random phylogeny instead of one or more depending on the inputMoved to #57.
Moved to #58.
Moved to #59.
Moved to #60.
Moved to #61.
Moved to #62.
Moved to #63.
Moved to #64.
Moved to #65.
Moved to #66.
Moved to #67.
Moved to #68.
Search for:
# Irregularity: WIP, TODO, to be moved to someplace else
Move someplace better.
Feedback bjoelle:
the relaxed clock model doesn't work with MRCA priors that have a time calibration distribution: the state, operators, distribution and tracelog sections are all missing elements compared to an XML produced with BEAUti. This issue appears to be linked to
clock_models_to_xml_statel24-25: the comment is unclear as to whether it is supposed to be checking for any or no MRCA priors, and no check is performed here for either.
At, for example, a1d8693, there was an 'Examples' vignette, close to identical to the Examples markdown file.
It is removed as it gives error when building:
0.34schecking re-building of vignette outputs ... WARNING
Error in re-building vignettes:
...
pandoc: Could not fetch all_default.png
all_default.png: openBinaryFile: does not exist (No such file or directory)
Error: processing vignette 'examples.Rmd' failed with diagnostics:
pandoc document conversion failed with error 67
all_default.png does exists. The vignette does build locally with CTRL-K. The vignette also fails when testing the package with CTRL-SHIFT-E.
No idea how to fix this, the minimal example, travis_r_vignette_with_pics, works as expected.
Feedback bjoelle:
is_index_of_first_shared_clock_model: this function will return TRUE if the tested model is the 2nd of 3 shared clock models, in contradiction with the doc. It will also always return TRUE if the clock model at index i has id=NA.
Feedback bjoelle:
get_phylo_crown_agewill not work with sampling through time/non-ultrametric trees. I think these are not supported at the moment anyway, but it might be good to keep it in mind.
print(Sys.time())
Rprof("~/profile.txt")
input_fasta_filenames <- beautier:::get_paths(
c("anthus_aco.fas", "anthus_nd2.fas"))
n_fail <- 0
for (site_model_1 in beautier::create_site_models()) {
for (site_model_2 in beautier::create_site_models()) {
for (clock_model_1 in beautier::create_clock_models()) {
for (clock_model_2 in beautier::create_clock_models()) {
for (tree_prior in beautier::create_tree_priors()) {
if (runif(n = 1) < 0.9) next
output_xml_filename <- "~/invalid.xml"
create_beast2_input_file(
input_fasta_filenames = input_fasta_filenames,
site_models = list(site_model_1, site_model_2),
clock_models = list(clock_model_1, clock_model_2),
tree_priors = list(tree_prior, tree_prior),
output_xml_filename = output_xml_filename
)
is_ok <- beautier::is_beast2_input_file(output_xml_filename)
if (!is_ok) {
print(paste(site_model_1$name, site_model_2$name,
clock_model_1$name, clock_model_2$name, tree_prior$name))
beautier::is_beast2_input_file(output_xml_filename,
verbose = TRUE)
n_fail <- n_fail + 1
}
# testthat::expect_true(is_ok) # nolint one day will be OK
}
}
}
}
}
testthat::expect_equal(n_fail, 0)
Rprof()
print(Sys.time())
summaryRprof("~/profile.txt")
results in
print(Sys.time())
[1] "2017-12-19 11:04:45 CET"
[code]
> print(Sys.time())
[1] "2017-12-19 11:06:00 CET"
> summaryRprof("~/profile.txt")
$by.self
self.time self.pct total.time total.pct
"cat" 1.20 24.90 1.20 24.90
"getExportedValue" 0.36 7.47 1.02 21.16
"%in%" 0.28 5.81 0.70 14.52
"asNamespace" 0.20 4.15 0.34 7.05
"system" 0.18 3.73 0.18 3.73
"::" 0.16 3.32 1.20 24.90
"isBaseNamespace" 0.14 2.90 0.14 2.90
"sub" 0.14 2.90 0.14 2.90
"assert2" 0.12 2.49 1.14 23.65
"get0" 0.12 2.49 0.12 2.49
[...]
"writeLines" 0.02 0.41 0.02 0.41
$by.total
total.time total.pct self.time self.pct
"create_beast2_input_file" 4.56 94.61 0.00 0.00
"create_beast2_input" 4.54 94.19 0.00 0.00
"create_beast2_input_beast" 3.14 65.15 0.00 0.00
"eval" 2.28 47.30 0.06 1.24
"create_beast2_input_data" 1.84 38.17 0.00 0.00
"create_beast2_input_data_sequences" 1.82 37.76 0.00 0.00
"beautier::fasta_file_to_sequences" 1.64 34.02 0.00 0.00
"utils::capture.output" 1.44 29.88 0.02 0.41
"testit::assert" 1.32 27.39 0.06 1.24
"withVisible" 1.22 25.31 0.02 0.41
[...]
$sample.interval
[1] 0.02
$sampling.time
[1] 4.82
Thus, out of 75 seconds, only 5 seconds were in the actual code. That means 70 seconds BEAST was called.
GTR relaxed_log_normal birth_death
With scripts/test_random_1 this is all it takes to trigger an error:
set.seed(0)
create_random()
Feedback bjoelle:
get_id,get_ids: These functions appear to be duplicated byget_alignment_idandget_alignment_ids.
Feedback bjoelle:
get_site_model_n_paramsandget_site_model_n_distrsdo not check if the site model has gamma rate categories, they just assume it does
Unknown, possibly mis-spelled, fields in DESCRIPTION:
'Remotes'Please omit this field.
Please single qupte software names such as 'BEAST2' in your Description and give web references in > the form yhttp.....> ifthese are not R packages.
Best,
Uwe Ligges
The main issue [...] (see #38)
Separating the different parameters (alpha, beta, gamma, etc) also leads to a lot of code duplication. It is useful for creating parameters as their default priors are different, but I would remove it for the other functions: the many functions is_%sth%_param(x) can be replaced by one function is_named_param(x, name) and the parameter_to_xml functions also have a lot of redundancy.
Error 110 parsing the xml input file
validate and intialize error: 333: Don't know how to generate a Random Tree for taxon sets that intersect, but are not inclusive. Taxonset auto_name_2.prior and auto_name_1.prior
Error detected about here:
<beast>
<run id='mcmc' spec='MCMC'>
<init id='RandomTree.t:anthus_aco' spec='beast.evolution.tree.RandomTree'>
Feedback bjoelle:
the distribution hyper parameters (alpha, beta, mean etc - all except lambda) can be set with
estimate = TRUE, but the xml will not be created correctly with those settings (missing distribution, operators and state sections). I think the code should not allow this setting until this functionality is fully implemented.
Feedback bjoelle
I found it confusing that the code relative to the interaction of MRCA priors and clock models is in a different place for strict and relaxed clocks, and it seems to have led to different behaviour between the 2 clocks (see comments in Functionality section). I would recommend putting everything in the
mrca_prior_to_xmlfunctions.
To know for sure the code is run.
Now it says a hardcoded value like:
id="RealParameter.11
Full error message:
Error 110 parsing the xml input file
validate and intialize error: 333: Don't know how to generate a Random Tree for taxon sets that intersect, but are not inclusive. Taxonset two_and_three.prior and one_and_two.prior
Error detected about here:
<beast>
<run id='mcmc' spec='MCMC'>
<init id='RandomTree.t:test_output_5' spec='beast.evolution.tree.RandomTree'>
Let's break it:
XML file created (that could be created):
When validating file beast2_warning.xml:
WARNING: StateNode (freqParameter.s:anthus_aco) found that has no effect on posterior!
Feedback bjoelle:
I think this package contains too many files, many of which contain only one small function. Gathering functions which are strongly linked to each other (for instance,
are_site_models,is_site_model,is_site_model_nameandget_site_model_names) in one file would make the package easier to read and search in my opinion.
validate and intialize error: There are more groups than coalescent nodes in the tree.
Error detected about here:
<beast>
<run id='mcmc' spec='MCMC'>
<distribution id='posterior' spec='util.CompoundDistribution'>
<distribution id='prior' spec='util.CompoundDistribution'>
<distribution id='BayesianSkyline.t:anthus_aco_sub' spec='BayesianSkyline'>
[1] "ERROR"
[1] ""
[2] " BEAST v2.4.7, 2002-2017"
[3] " Bayesian Evolutionary Analysis Sampling Trees"
[4] " Designed and developed by"
[5] " Remco Bouckaert, Alexei J. Drummond, Andrew Rambaut & Marc A. Suchard"
[6] " "
[7] " Department of Computer Science"
[8] " University of Auckland"
[9] " [email protected]"
[10] " [email protected]"
[11] " "
[12] " Institute of Evolutionary Biology"
[13] " University of Edinburgh"
[14] " [email protected]"
[15] " "
[16] " David Geffen School of Medicine"
[17] " University of California, Los Angeles"
[18] " [email protected]"
[19] " "
[20] " Downloads, Help & Resources:"
[21] " http://beast2.org/"
[22] " "
[23] " Source code distributed under the GNU Lesser General Public License:"
[24] " http://github.com/CompEvol/beast2"
[25] " "
[26] " BEAST developers:"
[27] " Alex Alekseyenko, Trevor Bedford, Erik Bloomquist, Joseph Heled, "
[28] " Sebastian Hoehna, Denise Kuehnert, Philippe Lemey, Wai Lok Sibon Li, "
[29] "Gerton Lunter, Sidney Markowitz, Vladimir Minin, Michael Defoin Platel, "
[30] " Oliver Pybus, Chieh-Hsi Wu, Walter Xie"
[31] " "
[32] " Thanks to:"
[33] " Roald Forsberg, Beth Shapiro and Korbinian Strimmer"
[34] ""
[35] "Random number seed: 1519987266495"
[36] ""
[37] "61430_aco: 960 4"
[38] "626029_aco: 960 4"
[39] "630116_aco: 960 4"
[40] "630210_aco: 960 4"
[41] "AMNH13589_aco: 960 4"
[42] "B25702_aco: 960 4"
[43] "B41613_aco: 960 4"
[44] "B431_aco: 960 4"
[45] "B48218_aco: 960 4"
[46] "B87109_aco: 960 4"
[47] "FALK1_aco: 960 4"
[48] "KU21673_aco: 960 4"
[49] "KU25127_aco: 960 4"
[50] "KU3604_aco: 960 4"
[51] "KU9813_aco: 960 4"
[52] "UWBM54394_aco: 960 4"
[53] "UWBM54511_aco: 960 4"
[54] "UWBM54556_aco: 960 4"
[55] "bas3_aco: 960 4"
[56] "chacoensis_aco: 960 4"
[57] "dabbenei_aco: 960 4"
[58] "meridae_aco: 960 4"
[59] "Alignment(anthus_aco)"
[60] " 22 taxa"
[61] " 960 sites"
[62] " 70 patterns"
[63] ""
[64] ""
[65] "Error 110 parsing the xml input file"
[66] ""
[67] "validate and intialize error: Cannot compute beta density at 1 when beta = %.3g"
[68] ""
[69] "Error detected about here:"
[70] " <beast>"
[71] " <run id='mcmc' spec='MCMC'>"
[72] " <distribution id='posterior' spec='util.CompoundDistribution'>"
[73] " <distribution id='prior' spec='util.CompoundDistribution'>"
[74] " <prior id='YuleBirthRatePrior.t:anthus_aco' name='distribution'>"
[75] ""
attr(,"status")
[1] 1
[1] "site model:"
$name
[1] "JC69"
$id
[1] NA
$gamma_site_model
$gamma_site_model$gamma_cat_count
[1] 0
$gamma_site_model$gamma_shape
[1] 0.8105238
$gamma_site_model$prop_invariant
[1] 0.7663484
$gamma_site_model$gamma_shape_prior_distr
$gamma_site_model$gamma_shape_prior_distr$name
[1] "inv_gamma"
$gamma_site_model$gamma_shape_prior_distr$id
[1] NA
$gamma_site_model$gamma_shape_prior_distr$alpha
$gamma_site_model$gamma_shape_prior_distr$alpha$name
[1] "alpha"
$gamma_site_model$gamma_shape_prior_distr$alpha$id
[1] NA
$gamma_site_model$gamma_shape_prior_distr$alpha$estimate
[1] FALSE
$gamma_site_model$gamma_shape_prior_distr$alpha$value
[1] 92.08046
$gamma_site_model$gamma_shape_prior_distr$beta
$gamma_site_model$gamma_shape_prior_distr$beta$name
[1] "beta"
$gamma_site_model$gamma_shape_prior_distr$beta$id
[1] NA
$gamma_site_model$gamma_shape_prior_distr$beta$estimate
[1] FALSE
$gamma_site_model$gamma_shape_prior_distr$beta$value
[1] -41.83111
[1] "clock model:"
$name
[1] "strict"
$id
[1] NA
$clock_rate_param
$clock_rate_param$name
[1] "clock_rate"
$clock_rate_param$id
[1] NA
$clock_rate_param$estimate
[1] FALSE
$clock_rate_param$value
[1] 66.6412
$clock_rate_distr
$clock_rate_distr$name
[1] "laplace"
$clock_rate_distr$id
[1] NA
$clock_rate_distr$mu
$clock_rate_distr$mu$name
[1] "mu"
$clock_rate_distr$mu$id
[1] NA
$clock_rate_distr$mu$estimate
[1] TRUE
$clock_rate_distr$mu$value
[1] 9.882322
$clock_rate_distr$scale
$clock_rate_distr$scale$name
[1] "scale"
$clock_rate_distr$scale$id
[1] NA
$clock_rate_distr$scale$estimate
[1] TRUE
$clock_rate_distr$scale$value
[1] -98.63774
[1] "tree prior:"
$name
[1] "yule"
$id
[1] NA
$birth_rate_distr
$birth_rate_distr$name
[1] "beta"
$birth_rate_distr$id
[1] NA
$birth_rate_distr$alpha
$birth_rate_distr$alpha$name
[1] "alpha"
$birth_rate_distr$alpha$id
[1] NA
$birth_rate_distr$alpha$estimate
[1] TRUE
$birth_rate_distr$alpha$value
[1] -4.153554
$birth_rate_distr$beta
$birth_rate_distr$beta$name
[1] "beta"
$birth_rate_distr$beta$id
[1] NA
$birth_rate_distr$beta$estimate
[1] FALSE
$birth_rate_distr$beta$value
[1] -93.9874
The structure is my main concern with this package, as it contains a lot of code duplication and many functions need to run through all possible types of a certain object to figure out what to do with it. I think this makes the package hard to maintain and to extend to additional models.
The main issue in my opinion is that the objects (site models, clock models, etc) do not store their dependencies in a generic way, so that there is no way to know what objects a model depends on unless you know which specific model it is. My suggestion would be to add a list of components to all objects - for instance a TN93 site model would have components = c("kappa_1", "kappa_2"), a HKY would have components = "kappa" and so on. This would allow you to save a lot of code in functions like init_site_models, which could do things such as for(x in model$components) model[[paste0(x,"_param")]] = init_param(model[[paste0(x,"_param")]]) instead of needing one subfunction per type of model.
As done by scripts/test_random_nonsense_1.R.
Note that this test for random possible inputs. That is: if a function needs a, say, distribution, it will test with a random, potentially nonsense, input.
In all cases, there should be a clean break.
p230198@peregrine:scripts cat test_random_nonsense_1_698476.log
The following have been reloaded with a version change:
1) ncurses/6.0 => ncurses/6.0-foss-2016a
[1] "seed: 1720902764"
[1] 1
...
[1] 182
[1] 183
[1] 184
[1] "ERROR"
[1] ""
[2] " BEAST v2.4.0, 2002-2016"
[3] " Bayesian Evolutionary Analysis Sampling Trees"
[4] " Designed and developed by"
[5] " Remco Bouckaert, Alexei J. Drummond, Andrew Rambaut & Marc A. Suchard"
[6] " "
[7] " Department of Computer Science"
[8] " University of Auckland"
[9] " [email protected]"
[10] " [email protected]"
[11] " "
[12] " Institute of Evolutionary Biology"
[13] " University of Edinburgh"
[14] " [email protected]"
[15] " "
[16] " David Geffen School of Medicine"
[17] " University of California, Los Angeles"
[18] " [email protected]"
[19] " "
[20] " Downloads, Help & Resources:"
[21] " http://beast2.org/"
[22] " "
[23] " Source code distributed under the GNU Lesser General Public License:"
[24] " http://github.com/CompEvol/beast2"
[25] " "
[26] " BEAST developers:"
[27] " Alex Alekseyenko, Trevor Bedford, Erik Bloomquist, Joseph Heled, "
[28] " Sebastian Hoehna, Denise Kuehnert, Philippe Lemey, Wai Lok Sibon Li, "
[29] "Gerton Lunter, Sidney Markowitz, Vladimir Minin, Michael Defoin Platel, "
[30] " Oliver Pybus, Chieh-Hsi Wu, Walter Xie"
[31] " "
[32] " Thanks to:"
[33] " Roald Forsberg, Beth Shapiro and Korbinian Strimmer"
[34] ""
[35] "Random number seed: 1519993479561"
[36] ""
[37] "Loading package BEAST v2.4.0"
[38] "61430_aco: 960 4"
[39] "626029_aco: 960 4"
[40] "630116_aco: 960 4"
[41] "630210_aco: 960 4"
[42] "AMNH13589_aco: 960 4"
[43] "B25702_aco: 960 4"
[44] "B41613_aco: 960 4"
[45] "B431_aco: 960 4"
[46] "B48218_aco: 960 4"
[47] "B87109_aco: 960 4"
[48] "FALK1_aco: 960 4"
[49] "KU21673_aco: 960 4"
[50] "KU25127_aco: 960 4"
[51] "KU3604_aco: 960 4"
[52] "KU9813_aco: 960 4"
[53] "UWBM54394_aco: 960 4"
[54] "UWBM54511_aco: 960 4"
[55] "UWBM54556_aco: 960 4"
[56] "bas3_aco: 960 4"
[57] "chacoensis_aco: 960 4"
[58] "dabbenei_aco: 960 4"
[59] "meridae_aco: 960 4"
[60] "Alignment(anthus_aco)"
[61] " 22 taxa"
[62] " 960 sites"
[63] " 70 patterns"
[64] ""
[65] ""
[66] "Error 110 parsing the xml input file"
[67] ""
[68] "nullvalidate and intialize error: Cannot compute beta density at 1 when beta = %.3g"
[69] ""
[70] "Error detected about here:"
[71] " <beast>"
[72] " <run id='mcmc' spec='MCMC'>"
[73] " <distribution id='posterior' spec='util.CompoundDistribution'>"
[74] " <distribution id='prior' spec='util.CompoundDistribution'>"
[75] " <prior id='YuleBirthRatePrior.t:anthus_aco' name='distribution'>"
[76] ""
attr(,"status")
[1] 1
[1] "site model:"
Error in print(site_model) : object 'site_model' not found
Calls: source ... withVisible -> eval -> eval -> create_random -> print
In addition: Warning messages:
1: running command 'java -jar ~/Programs/beast/lib/beast.jar -validate /local/698476/RtmpOuJCwU/file210862f948c7 2>/dev/null' had status 1
2: In file.create(to[okay]) :
cannot create file '/home/richel/bad.xml', reason 'No such file or directory'
3: running command 'java -jar ~/Programs/beast/lib/beast.jar -validate /local/698476/RtmpOuJCwU/file210862f948c7 2>/dev/null' had status 1
Execution halted
As done by scripts/test_random_1.R.
Note that this test for random valid inputs. That is: if a function needs a, say, distribution, it will test a random distribution.
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
Django
The Web framework for perfectionists with deadlines.
Laravel
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft
Open source projects and samples from Microsoft.
Google
Google ❤️ Open Source for everyone.
Alibaba
Alibaba Open Source for everyone
D3
Data-Driven Documents codes.
Tencent
China tencent open source team.