SETUP.md

In emjun/tisaner: Authoring Statistical Models from Conceptual Models

Updated:

Python

1. brew install python

2. Install virtualenv. In terminal:

pip3 install virtualenv

1. Create virutalenv. In terminal:

virtualenv rTisane-env # rTisane-env is the name of the env

1. Source virtualenv. In terminal:

source rTisane-env/bin/activate

Now you're working in the virtualenv!

1. Install Python dependencies:

pip3 install dash
pip3 install dash_daq
pip3 install dash_bootstrap_components
pip3 install flask
pip3 install plotly
pip3 install tisanecodegenerator
pip3 install numpy

R

1. Install reticulate

install.packages("reticulate")
library(reticulate)
reticulate::use_virtualenv("./rTisane-env")

Check Python installation: reticulate::py_discover_config() python, libpython, etc. should have a path

1. Install more R dependencies:

install.packages("devtools")
install.packages("magrittr")
install.packages("shinyjs")
install.packages("shiny")
install.packages("dagitty")
install.packages("igraph")

library(devtools)
library(magrittr) # for pipe %>%
library(shinyjs)
library(shiny)
library(dagitty)
library(igraph)

Use rTisane!

1. Always show disambiguation interfaces in the browser

source("~/.Rprofile")

1. Load rTisane

# Use local build of rTisane (under development)
devtools::load_all()
# For conceptual model disambiguation
source("conceptualDisambiguation/app.R")

==========

Install rTisane and its dependencies

1. Install python. Note: Reticulate installs and uses miniconda from a path like "~/Library/r-miniconda-arm64"

install.packages("reticulate")
library(reticulate)
reticulate::install_miniconda() 
# Help with install: https://github.com/rstudio/reticulate/issues/637

library(magrittr) # for pipe %>%

# For disambiguation
install.packages("shinyjs")
library(shinyjs)
library(shiny)

# Use local build of rTisane (under development)
library(devtools)
devtools::load_all()
# For conceptual model disambiguation
source("conceptualDisambiguation/app.R")

1. Install Python libraries/dependencies

py_install("dash", pip=TRUE)
py_install("dash_daq", pip=TRUE)
py_install("dash_bootstrap_components", pip=TRUE)
py_install("flask", pip=TRUE)
py_install("plotly", pip=TRUE)
py_install("tisanecodegenerator", pip=TRUE)
py_install("pandas", pip=TRUE)

import("dash")
import("dash_daq")
import("dash_bootstrap_components")
import("flask")
import("plotly")
import("tisanecodegenerator")
import("pandas")

To run all disambiguation processes in the default web browser instead of RStudio's internal browser

1. Add to Rprofile (at "~/.Rprofile")

options(shiny.launch.browser = .rs.invokeShinyWindowExternal)

1. Source at the start of the R session

source("~/.Rprofile")

Should not need: To make sure rTisane finds the conceptual model disambiguation interface

source("conceptualDisambiguation/app.R")

Setup Python virtualenv to use in project:

Follow instructions: https://support.posit.co/hc/en-us/articles/360023654474-Installing-and-Configuring-Python-with-RStudio

To set up

Create python virtualenv

emjun/tisaner documentation built on Nov. 9, 2024, 6:32 p.m.