Update README with detailed instructions

Trusted-AI · Sep 11, 2020 · af4688c · af4688c
1 parent e267a8d
commit af4688c
Show file tree

Hide file tree

Showing 2 changed files with 147 additions and 4 deletions.
diff --git a/aif360/aif360-r/README.Rmd b/aif360/aif360-r/README.Rmd
@@ -47,6 +47,69 @@ library(aif360)
 install_aif360()
 ``` 
 
+## Installation methods
+
+AIF360 is distributed as a Python package and so needs to be installed within a Python environment on your system. By default, the install_aif360() function attempts to install AIF360 within an isolated Python environment (“r-reticulate”).
+
+You can check using `reticulate::conda_python()` and `reticulate::py_config()`
+
+### Suggested steps
+
+ 1)  Install reticulate  and check if you have miniconda installed. If you do, go to step 2.
+
+ ```r
+ install.packages("reticulate")
+ reticulate::conda_list()
+ ```
+ If you get an error: `Error: Unable to find conda binary. Is Anaconda installed?`, please install
+ miniconda
+
+```r
+reticulate::install_miniconda()
+```
+If everything worked, you should get the message: 
+
+`* Miniconda has been successfully installed at '/home/rstudio/.local/share/r-miniconda'.` 
+
+You can double check:
+
+```
+reticulate::conda_list()
+```
+You will get something like this:
+
+```
+          name                                                              python
+1  r-miniconda                   /home/rstudio/.local/share/r-miniconda/bin/python
+2 r-reticulate /home/rstudio/.local/share/r-miniconda/envs/r-reticulate/bin/python
+```
+
+
+2) You can create a new conda env and then configure which version of Python to use:
+
+```r
+reticulate::conda_create(envname = "r-test")
+reticulate::use_miniconda(condaenv = "r-test", required = TRUE)
+```
+Check that everything is working `reticulate::py_config()`.
+
+3) If you haven't yet, please install the aif360 package `install.packages("aif360")` and then
+install aif360 dependencies
+
+```r
+aif360::install_aif360(envname = "r-test")
+```
+Note that this step should take a few minutes and the R session will restart.
+
+4) Finally, load the aif360 functions
+
+```
+library(aif360)
+reticulate::use_miniconda(condaenv = "r-test", required = TRUE)
+load_aif360_lib()
+```
+
+
 ## Getting Started
 
 ``` r
@@ -56,8 +119,9 @@ load_aif360_lib()
 ``` r
 # load a toy dataset
 data <- data.frame("feature1" = c(0,0,1,1,1,1,0,1,1,0), 
-                   "feature2" = c(0,1,0,1,1,0,0,0,0,1),
-+                  "label" = c(1,0,0,1,0,0,1,0,1,1))
+                   "feature2" = c(0,1,0,1,1,0,0,0,0,1), 
+                   "label" = c(1,0,0,1,0,0,1,0,1,1))
+
 # format the dataset
 formatted_dataset <- aif360::aif_dataset(data_path = data, 
                                           favor_label = 0, 

diff --git a/aif360/aif360-r/README.md b/aif360/aif360-r/README.md
@@ -38,6 +38,84 @@ library(aif360)
 install_aif360()
 ```
 
+## Installation methods
+
+AIF360 is distributed as a Python package and so needs to be installed
+within a Python environment on your system. By default, the
+install\_aif360() function attempts to install AIF360 within an isolated
+Python environment (“r-reticulate”).
+
+You can check using `reticulate::conda_python()` and
+`reticulate::py_config()`
+
+### Suggested steps
+
+1)  Install reticulate and check if you have miniconda installed. If you
+    do, go to step 2.
+
+<!-- end list -->
+
+``` r
+install.packages("reticulate")
+reticulate::conda_list()
+```
+
+If you get an error: `Error: Unable to find conda binary. Is Anaconda
+installed?`, please install miniconda
+
+``` r
+reticulate::install_miniconda()
+```
+
+If everything worked, you should get the message:
+
+`* Miniconda has been successfully installed at
+'/home/rstudio/.local/share/r-miniconda'.`
+
+You can double check:
+
+    reticulate::conda_list()
+
+You will get something like this:
+
+``` 
+          name                                                              python
+1  r-miniconda                   /home/rstudio/.local/share/r-miniconda/bin/python
+2 r-reticulate /home/rstudio/.local/share/r-miniconda/envs/r-reticulate/bin/python
+```
+
+2)  You can create a new conda env and then configure which version of
+    Python to use:
+
+<!-- end list -->
+
+``` r
+reticulate::conda_create(envname = "r-test")
+reticulate::use_miniconda(condaenv = "r-test", required = TRUE)
+```
+
+Check that everything is working `reticulate::py_config()`.
+
+3)  If you haven’t yet, please install the aif360 package
+    `install.packages("aif360")` and then install aif360 dependencies
+
+<!-- end list -->
+
+``` r
+aif360::install_aif360(envname = "r-test")
+```
+
+Note that this step should take a few minutes and the R session will
+restart.
+
+4)  Finally, load the aif360 functions
+
+<!-- end list -->
+
+    library(aif360)
+    reticulate::use_miniconda(condaenv = "r-test", required = TRUE)
+    load_aif360_lib()
+
 ## Getting Started
 
 ``` r
@@ -47,8 +125,9 @@ load_aif360_lib()
 ``` r
 # load a toy dataset
 data <- data.frame("feature1" = c(0,0,1,1,1,1,0,1,1,0), 
-                   "feature2" = c(0,1,0,1,1,0,0,0,0,1),
-+                  "label" = c(1,0,0,1,0,0,1,0,1,1))
+                   "feature2" = c(0,1,0,1,1,0,0,0,0,1), 
+                   "label" = c(1,0,0,1,0,0,1,0,1,1))
+
 # format the dataset
 formatted_dataset <- aif360::aif_dataset(data_path = data, 
                                           favor_label = 0,