Merge pull request #21 from bnicenboim/dev

Dev

DESCRIPTION

@@ -1,7 +1,7 @@
 Package: pangoling
 Type: Package
 Title: Access to Large Language Model Predictions
-Version: 0.0.0.9009
+Version: 0.0.0.9010
 Authors@R: c(
     person("Bruno", "Nicenboim",
     email = "[email protected]",
@@ -18,14 +18,7 @@ BugReports: https://github.com/bnicenboim/pangoling/issues
 License: MIT + file LICENSE
 Encoding: UTF-8
 LazyData: false
-Config/reticulate/autoconfigure:
-  list(
-    packages = list(
-      list(package = "torch"),
-      list(package = "transformers")
-    )
-  )
-Imports: 
+Imports:
     data.table,
     memoise,
     reticulate,
@@ -39,9 +32,10 @@ Suggests:
     testthat (>= 3.0.0),
     tictoc,
     covr,
-    spelling
+    spelling,
+    rstudioapi
 Config/testthat/edition: 3
-RoxygenNote: 7.2.3
+RoxygenNote: 7.3.1
 Roxygen: list(markdown = TRUE)
 Depends: 
     R (>= 4.1.0)

NAMESPACE

@@ -8,6 +8,7 @@ export(causal_lp_mats)
 export(causal_next_tokens_tbl)
 export(causal_preload)
 export(causal_tokens_lp_tbl)
+export(install_py_pangoling)
 export(masked_config)
 export(masked_lp)
 export(masked_preload)

R/utils.R

@@ -1,23 +1,133 @@
+#' Install the Python packages needed for `pangoling`
+#'
+#' @description
+#' `install_py_pangoling` function facilitates the installation of Python packages needed for using `pangoling` within an R environment,
+#' utilizing the `reticulate` package for managing Python environments. It supports various installation methods,
+#' environment settings, and Python versions.
+#'
+#' @usage
+#' install_py_pangoling(method = c("auto", "virtualenv", "conda"),
+#'                      conda = "auto",
+#'                      version = "default",
+#'                      envname = "r-pangoling",
+#'                      restart_session = TRUE,
+#'                      conda_python_version = NULL,
+#'                      ...,
+#'                      pip_ignore_installed = FALSE,
+#'                      new_env = identical(envname, "r-pangoling"),
+#'                      python_version = NULL)
+#'
+#' @param method A character vector specifying the environment management method. 
+#'               Options are 'auto', 'virtualenv', and 'conda'. Default is 'auto'.
+#' @param conda Specifies the conda binary to use. Default is 'auto'.
+#' @param version The Python version to use. Default is 'default', automatically selected.
+#' @param envname Name of the virtual environment. Default is 'r-pangoling'.
+#' @param restart_session Logical, whether to restart the R session after installation. 
+#'                        Default is TRUE.
+#' @param conda_python_version Python version for conda environments.
+#' @param ... Additional arguments passed to `reticulate::py_install`.
+#' @param pip_ignore_installed Logical, whether to ignore already installed packages. 
+#'                             Default is FALSE.
+#' @param new_env Logical, whether to create a new environment if `envname` is 'r-pangoling'.
+#'               Default is the identity of `envname`.
+#' @param python_version Specifies the Python version for the environment.
+#'
+#' @details
+#' This function automatically selects the appropriate method for environment management and Python installation,
+#' with a focus on virtual and conda environments. It ensures flexibility in dependency management and Python version control.
+#' If a new environment is created, existing environments with the same name are removed.
+#'
+#' @return
+#' The function returns `NULL` invisibly, but outputs a message on successful installation.
+#' @export
+install_py_pangoling <- function(method = c("auto", "virtualenv", "conda"),
+                                 conda = "auto",
+                                 version = "default", 
+                                 envname = "r-pangoling", 
+                                 restart_session = TRUE,
+                                 conda_python_version = NULL,
+                                 ...,
+                                 pip_ignore_installed = FALSE,
+                                 new_env = identical(envname, "r-pangoling"),
+                                 python_version = NULL
+                                 ){
+
+  method <- match.arg(method)
+
+  python_version <- python_version %||% conda_python_version
+  if(method %in% c("auto", "virtualenv") &&
+     is.null(python_version)) {
+
+    # virtualenv_starter() picks the most recent version available, but older
+    # versions of tensorflow typically don't work with the latest Python
+    # release. In general, we're better off picking the oldest Python version available
+    # that works with the current release of tensorflow.
+
+    available <- reticulate::virtualenv_starter(version = ">=3.9", all = TRUE)
+    # pick the smallest minor version, ignoring patchlevel
+    if(nrow(available))
+      python_version <- min(available$version[, 1:2])
+  }
+
+  if (isTRUE(new_env)) {
+
+    if (method %in% c("auto", "virtualenv") &&
+        reticulate::virtualenv_exists(envname))
+      reticulate::virtualenv_remove(envname = envname, confirm = FALSE)
+
+    if (method %in% c("auto", "conda")) {
+      if (!is.null(tryCatch(reticulate::conda_python(envname, conda = conda),
+                            error = function(e) NULL)))
+        reticulate::conda_remove(envname, conda = conda)
+    }
+
+  }
+  packages <- c("transformers", "torch")
+  py_install_args <- list(
+    packages       = packages,
+    envname        = envname,
+    method         = method,
+    conda          = conda,
+    python_version = python_version,
+    pip            = TRUE,
+    pip_ignore_installed = pip_ignore_installed,
+    ...
+  )
+
+  do.call(reticulate::py_install, py_install_args)
+  cat("\nInstallation complete.\n\n")
+
+  if (restart_session &&
+      requireNamespace("rstudioapi", quietly = TRUE) &&
+      rstudioapi::hasFun("restartSession"))
+    rstudioapi::restartSession()
+
+  invisible(NULL)
+
+}
+
 
 #' @noRd
 message_verbose <- function(...) {
   if (options()$pangoling.verbose > 0) message(...)
 }
 
+
+#' @noRd
 stop2 <- function(...) {
   stop(..., call. = FALSE)
 }
 
-#' #' Replacement of str_match
-#' #' @noRd
-#' chr_match <- function(string, pattern) {
-#'   matches <- regexec(pattern = pattern, text = string)
-#'   list_matches <- lapply(
-#'     regmatches(x = string, m = matches),
-#'     function(x) if (length(x) == 0) NA else x
-#'   )
-#'   do.call("rbind", list_matches)
-#' }
+# #' Replacement of str_match
+# #' @noRd
+# chr_match <- function(string, pattern) {
+#   matches <- regexec(pattern = pattern, text = string)
+#   list_matches <- lapply(
+#     regmatches(x = string, m = matches),
+#     function(x) if (length(x) == 0) NA else x
+#   )
+#   do.call("rbind", list_matches)
+# }
 
 
 #' Replacement of str_detect
@@ -27,7 +137,10 @@ chr_detect <- function(string, pattern, ignore.case = FALSE) {
 }
 
 
-#' #' @noRd
-#' message_debug <- function(...) {
-#'   if (options()$pangoling.verbose > 1) message(...)
-#' }
+# 
+# message_debug <- function(...) {
+#   if (options()$pangoling.verbose > 1) message(...)
+# }
+
+#' @noRd
+"%||%" <- function(x, y) if (is.null(x)) y else x

R/zzz.R

@@ -7,17 +7,16 @@ torch <- NULL
 #' @noRd
 .onLoad <- function(libname, pkgname) {
 
-  # This will instruct reticulate to immediately try to configure the
-  # active Python environment, installing any required Python packages
-  # as necessary.
-  reticulate::configure_environment(pkgname)
+  reticulate::use_virtualenv("r-pangoling", required = FALSE)
+
   # use superassignment to update global reference
   transformers <<- reticulate::import("transformers",
     delay_load = TRUE,
     convert = FALSE
   )
   torch <<- reticulate::import("torch", delay_load = TRUE, convert = FALSE)
   # TODO message or something if it's not installed
+  # ask about the env
   op <- options()
   op.pangoling <- list(
     pangoling.debug = FALSE,
@@ -47,5 +46,3 @@ torch <- NULL
     "\nAn introduction to the package can be found in https://bruno.nicenboim.me/pangoling/articles/\n Notice that pretrained models and tokenizers are downloaded from https://huggingface.co/ the first time they are used.\n For changing the cache folder use:\n
 set_cache_folder(my_new_path)")
 }
-
-

README.Rmd

@@ -41,6 +41,12 @@ There is still no released version of `pangoling`. The package is in the ** earl
 remotes::install_github("bnicenboim/pangoling")
 ```
 
+`install_py_pangoling` function facilitates the installation of Python packages needed for using pangoling within an R environment, using the `reticulate` package for managing Python environments. This needs to be done once.
+
+```{r, eval = FALSE}
+install_py_pangoling()
+```
+
 ## Example
 
 This is a basic example which shows you how to get log-probabilities of words in a dataset:

README.md

@@ -50,6 +50,15 @@ changes. To install the latest version from github use:
 remotes::install_github("bnicenboim/pangoling")
 ```
 
+`install_py_pangoling` function facilitates the installation of Python
+packages needed for using pangoling within an R environment, using the
+`reticulate` package for managing Python environments. This needs to be
+done once.
+
+``` r
+install_py_pangoling()
+```
+
 ## Example
 
 This is a basic example which shows you how to get log-probabilities of
@@ -101,6 +110,9 @@ df_sent <- df_sent |>
 #> `The apple doesn't fall far from the tree.`
 #> Text id: 2
 #> `Don't judge a book by its cover.`
+```
+
+``` r
 df_sent
 #> # A tidytable: 15 × 3
 #>    sent_n word         lp
@@ -125,7 +137,7 @@ df_sent
 ## How to cite
 
 > Nicenboim B (2023). *pangoling: Access to language model predictions
-> in R*. R package version 0.0.0.9008, DOI:
+> in R*. R package version 0.0.0.9010, DOI:
 > [10.5281/zenodo.7637526](https://zenodo.org/badge/latestdoi/497831295),
 > <https://github.com/bnicenboim/pangoling>.