From a3436710238a77f27300f1ef09e1163a6cb0365e Mon Sep 17 00:00:00 2001 From: Rich FitzJohn Date: Wed, 6 Nov 2024 07:53:55 +0000 Subject: [PATCH] Add xrefs in docs --- R/sample.R | 24 ++++++++++++------------ man/monty_sample.Rd | 24 ++++++++++++------------ man/monty_sample_manual_prepare.Rd | 24 ++++++++++++------------ 3 files changed, 36 insertions(+), 36 deletions(-) diff --git a/R/sample.R b/R/sample.R index 0945dfa6..431b2e38 100644 --- a/R/sample.R +++ b/R/sample.R @@ -38,23 +38,23 @@ ##' @param burnin Number of steps to discard as burnin. This affects ##' only the recording of steps as your chains run; we don't record ##' the first `burnin` steps. Generally you would want to do this -##' in post-processing as this data is discarded with no chance of -##' getting it back. However, if your observation process creates a -##' large amount of data, then you may prefer to apply a burnin here -##' to reduce how much memory is used. +##' in post-processing with [monty_samples_thin()] as this data is +##' discarded with no chance of getting it back. However, if your +##' observation process creates a large amount of data, then you may +##' prefer to apply a burnin here to reduce how much memory is used. ##' ##' @param thinning_factor A thinning factor to apply while the chain ##' is running. If given, then we save every `thinning_factor`'th ##' step. So if `thinning_factor = 2` we save every second step, ##' and if 10, we'd save every 10th. Like `burnin` above, it is -##' preferable to apply this in post processing. However, for -##' slow-mixing chains that have a large observer output you can use -##' this to reduce the memory usage. Use of `thinning_factor` -##' requires that `n_steps` is an even multiple of -##' `thinning_factor`; so if `thinning_factor` is 10, then `n_steps` -##' must be a multiple of 10. This ensures that the last step is in -##' the sample. The thinning factor cannot be changed when -##' continuing a chain. +##' preferable to apply this in post processing with +##' [monty_samples_thin()]. However, for slow-mixing chains that +##' have a large observer output you can use this to reduce the +##' memory usage. Use of `thinning_factor` requires that `n_steps` +##' is an even multiple of `thinning_factor`; so if +##' `thinning_factor` is 10, then `n_steps` must be a multiple of +##' 10. This ensures that the last step is in the sample. The +##' thinning factor cannot be changed when continuing a chain. ##' ##' @return A list of parameters and densities. We provide conversion ##' to formats used by other packages, notably diff --git a/man/monty_sample.Rd b/man/monty_sample.Rd index 02776879..edc1221b 100644 --- a/man/monty_sample.Rd +++ b/man/monty_sample.Rd @@ -48,23 +48,23 @@ object.} \item{burnin}{Number of steps to discard as burnin. This affects only the recording of steps as your chains run; we don't record the first \code{burnin} steps. Generally you would want to do this -in post-processing as this data is discarded with no chance of -getting it back. However, if your observation process creates a -large amount of data, then you may prefer to apply a burnin here -to reduce how much memory is used.} +in post-processing with \code{\link[=monty_samples_thin]{monty_samples_thin()}} as this data is +discarded with no chance of getting it back. However, if your +observation process creates a large amount of data, then you may +prefer to apply a burnin here to reduce how much memory is used.} \item{thinning_factor}{A thinning factor to apply while the chain is running. If given, then we save every \code{thinning_factor}'th step. So if \code{thinning_factor = 2} we save every second step, and if 10, we'd save every 10th. Like \code{burnin} above, it is -preferable to apply this in post processing. However, for -slow-mixing chains that have a large observer output you can use -this to reduce the memory usage. Use of \code{thinning_factor} -requires that \code{n_steps} is an even multiple of -\code{thinning_factor}; so if \code{thinning_factor} is 10, then \code{n_steps} -must be a multiple of 10. This ensures that the last step is in -the sample. The thinning factor cannot be changed when -continuing a chain.} +preferable to apply this in post processing with +\code{\link[=monty_samples_thin]{monty_samples_thin()}}. However, for slow-mixing chains that +have a large observer output you can use this to reduce the +memory usage. Use of \code{thinning_factor} requires that \code{n_steps} +is an even multiple of \code{thinning_factor}; so if +\code{thinning_factor} is 10, then \code{n_steps} must be a multiple of +10. This ensures that the last step is in the sample. The +thinning factor cannot be changed when continuing a chain.} } \value{ A list of parameters and densities. We provide conversion diff --git a/man/monty_sample_manual_prepare.Rd b/man/monty_sample_manual_prepare.Rd index cb111079..5ff4fe5b 100644 --- a/man/monty_sample_manual_prepare.Rd +++ b/man/monty_sample_manual_prepare.Rd @@ -49,23 +49,23 @@ single chain, but you will likely want to run more.} \item{burnin}{Number of steps to discard as burnin. This affects only the recording of steps as your chains run; we don't record the first \code{burnin} steps. Generally you would want to do this -in post-processing as this data is discarded with no chance of -getting it back. However, if your observation process creates a -large amount of data, then you may prefer to apply a burnin here -to reduce how much memory is used.} +in post-processing with \code{\link[=monty_samples_thin]{monty_samples_thin()}} as this data is +discarded with no chance of getting it back. However, if your +observation process creates a large amount of data, then you may +prefer to apply a burnin here to reduce how much memory is used.} \item{thinning_factor}{A thinning factor to apply while the chain is running. If given, then we save every \code{thinning_factor}'th step. So if \code{thinning_factor = 2} we save every second step, and if 10, we'd save every 10th. Like \code{burnin} above, it is -preferable to apply this in post processing. However, for -slow-mixing chains that have a large observer output you can use -this to reduce the memory usage. Use of \code{thinning_factor} -requires that \code{n_steps} is an even multiple of -\code{thinning_factor}; so if \code{thinning_factor} is 10, then \code{n_steps} -must be a multiple of 10. This ensures that the last step is in -the sample. The thinning factor cannot be changed when -continuing a chain.} +preferable to apply this in post processing with +\code{\link[=monty_samples_thin]{monty_samples_thin()}}. However, for slow-mixing chains that +have a large observer output you can use this to reduce the +memory usage. Use of \code{thinning_factor} requires that \code{n_steps} +is an even multiple of \code{thinning_factor}; so if +\code{thinning_factor} is 10, then \code{n_steps} must be a multiple of +10. This ensures that the last step is in the sample. The +thinning factor cannot be changed when continuing a chain.} } \value{ Invisibly, the path used to store files (the same as the