* update README with latest installation instructions

* add r-universe badge
* fix cran badge
* update greta_python_deps() and install_greta_deps() docs

Author: njtierney

R/install_greta_deps.R

@@ -201,17 +201,44 @@ restart_or_not <- function(restart){
 
 #' Specify python dependencies for greta
 #'
-#' A helper function for specifying versions of Tensorflow (TF), Tensorflow Probability (TFP), and Python.
+#' A helper function for specifying versions of Tensorflow (TF), Tensorflow
+#'   Probability (TFP), and Python. Defaulting to 2.15.0, 0.23.0, and 3.10.
+#'   You can specify the version that you want to install, but it will check
+#'   if these are compatible. That is, if you specify versions of TF/TFP/Python
+#'   which do not work with each other, it will error and give a suggested
+#'   version to install. It does this by using a dataset, `greta_deps_tf_tfp`,
+#'   to check if the versions of TF, TFP, and Python specified are compatible
+#'   on your operating system. You can inspect this  dataset with
+#'   `View(greta_deps_tf_tfp)`.
 #'
-#' @param tf_version Character. Tensorflow (TF) version in format major.minor.patch. Default is "2.15.0".
-#' @param tfp_version Character.Tensorflow probability (TFP) version major.minor.patch. Default is "0.23.0".
-#' @param python_version Character. Python version in format major.minor.patch. Default is "3.10".
+#' @param tf_version Character. Tensorflow (TF) version in format
+#'   major.minor.patch. Default is "2.15.0".
+#' @param tfp_version Character.Tensorflow probability (TFP) version
+#'   major.minor.patch. Default is "0.23.0".
+#' @param python_version Character. Python version in format major.minor.patch.
+#'   Default is "3.10".
 #'
-#' @return list of dependencies
+#' @return data frame of valid dependencies
 #' @export
 #'
 #' @examples
 #' greta_python_deps()
+#' greta_python_deps(tf_version = "2.15.0")
+#' greta_python_deps(tf_version = "2.15.0", tfp_version = "0.23.0")
+#' greta_python_deps(tf_version = "2.15.0", python_version = "3.10")
+#' greta_python_deps(
+#'   tf_version = "2.15.0",
+#'   tfp_version = "0.23.0",
+#'   python_version = "3.10"
+#'   )
+#' # this will fail
+#' \dontrun{
+#' greta_python_deps(
+#'   tf_version = "2.11.0",
+#'   tfp_version = "0.23.0",
+#'   python_version = "3.10"
+#'   )
+#'   }
 greta_python_deps <- function(tf_version = "2.15.0",
                               tfp_version = "0.23.0",
                               python_version = "3.10"){

README.md

@@ -40,19 +40,21 @@ devtools::install_github("greta-dev/greta")
 
 # Installing Python Dependencies
 
-The `install_greta_deps()` function helps install the Python dependencies (Google's [TensorFlow](https://www.tensorflow.org/) version 1.14.0, and  [tensorflow-probability](https://github.com/tensorflow/probability) version 0.7.0).
+The `install_greta_deps()` function helps install the Python dependencies (Google's [TensorFlow](https://www.tensorflow.org/) and [tensorflow-probability](https://github.com/tensorflow/probability)). 
 
-This helper function, `install_greta_deps()`, installs the exact pythons package versions needed. It also places these inside a "greta-env" conda environment. This isolates these exact python modules from other python installations, so that only `greta` will see them. This helps avoids installation issues, where previously you might update tensorflow on your computer and overwrite the current version needed by `greta`. Using this "greta-env" conda environment means installing other python packages should not be impact the Python packages needed by `greta`.
+By default, `install_greta_deps()` installs versions TF 2.15.0, and TFP version 0.23.0, using python 3.10. To change the versions of TF, TFP, or python that you want to use, you specify the `python_deps` argument of `install_greta_deps()`, which used `greta_python_deps()`. See `?install_greta_deps()` or `?greta_python_deps()` for more information.
+
+This helper function, `install_greta_deps()`, installs the exact pythons package versions needed. It also places these inside a conda environment, "greta-env-tf2". This isolates these exact python modules from other python installations, so that only `greta` will see them. This helps avoids installation issues, where previously you might update tensorflow on your computer and overwrite the current version needed by `greta`. Using this "greta-env-tf2" conda environment means installing other python packages should not be impact the Python packages needed by `greta`.
 
 If these python modules aren't yet installed, when `greta` is used, it provides instructions on how to install them for your system. If in doubt follow those. 
 
 <!-- badges: start -->
 [![Codecov test coverage](https://codecov.io/gh/greta-dev/greta/branch/master/graph/badge.svg)](https://app.codecov.io/gh/greta-dev/greta?branch=master)
 [![R-CMD-check](https://github.com/greta-dev/greta/workflows/R-CMD-check/badge.svg)](https://github.com/greta-dev/greta/actions)
-[![cran
-version](http://www.r-pkg.org/badges/version/greta)](https://CRAN.R-project.org/package=greta)
+[![cran version](http://www.r-pkg.org/badges/version/greta)](https://CRAN.R-project.org/package=greta)
 [![license](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](https://opensource.org/license/apache-2-0)
 [![doi](https://zenodo.org/badge/73758247.svg)](https://zenodo.org/badge/latestdoi/73758247)
 [![joss](https://joss.theoj.org/papers/10.21105/joss.01601/status.svg)](https://joss.theoj.org/papers/10.21105/joss.01601)
 ![](logos/bottom_banner.png)
+[![greta status badge](https://greta-dev.r-universe.dev/badges/greta)](https://greta-dev.r-universe.dev/greta)
 <!-- badges: end -->