search

treeppl / treeppl.github.io

treeppl.org website
MIT License
1 stars 4 forks source link

Very good installation guide on the website #4

Open vsenderov opened 10 months ago

vsenderov commented 10 months ago

@kudlicka is working on a principled installation pipeline; however the README.md in the other repo already gives some hints how to install from source.

Eventually the whole installation can be described in detail.

vsenderov commented 1 month ago

I also have some local notes. They are probably outdated, but we can start by testing them and compiling the correct notes to the website.

TreePPL - Installation Instructions

Step 1: Install Miking

  1. Clone the Miking repository: 
git clone [https://github.com/miking-lang/miking](https://github.com/miking-lang/miking)
cd miking
  1. Install Miking using the following command:

make install

Step 2: Install Miking DPPL

  1. Clone the Miking DPPL repository: 
git clone [https://github.com/miking-lang/miking-dppl](https://github.com/miking-lang/miking-dppl)
cd miking-dppl
  1. Install Miking DPPL using the following command:

make install During the installation process, pay attention to the information about the environment variables (shown in red).

Step 3: Install TreePPL1. Clone the TreePPL repository:

git clone [https://github.com/vsenderov/treeppl](https://github.com/vsenderov/treeppl)
cd treeppl
  1. Switch to the json-dlunde branch:

git checkout json-dlunde

  1. Install TreePPL using the following command:

make install

Again, during the installation process, pay attention to the information about the environment variables.## Running TreePPLHere’s how you can run TreePPL and work with a sample model from the TreePPL repository, which simulates biased coin flips:1. While in the root directory of the TreePPL repository, run the following command to compile the example model:

tpplc models/lang/coin.tppl --output coin

This command generates an executable file named coin (you can specify the name of the executable using the --output flag).2. Run the executable, providing the path to a JSON data file containing values for the model parameters as an argument:

./coin models/data/coin.json

In this case, the model coinModel (defined in models/lang/coin.tppl) has a single parameter: coinflips: Bool[]. The coin.json file contains the following JSON object:

{"coinflips":[true, true, true, false, true, false, false, true, true, false, false, false, true, false, true, false, false, true, false, false]}

Running the executable will output a JSON object containing samples, weights and the normalization constant to the standard output.

TreePPL - Installation Procedure under WSL

Pre-prerequisites

Prerequisites

sudo apt update
sudo apt upgrade
sudo apt install unzip
sudo apt install bubblewrap
sudo apt install bzip2
sudo apt install python3-pip
sudo apt install gnome-text-editor -y

Opam and OCaml

You can save the following to a script, make it executable and run it.

```bash
#!/usr/bin/env bash
# Tested on Ubuntu 20.04 and 22.04.1 LTS WSL

####################
## Opam and OCaml ##
####################
read -p "Install/reinstall Opam and OCaml? " -n 1 -r; echo
if [[ $REPLY =~ ^[Yy]$ ]]; then
  ## Clean up
  rm -rf ~/.opam

  ## Install opam
  sh <(curl -sL https://raw.githubusercontent.com/ocaml/opam/master/shell/install.sh)

  ## Install required packages
  sudo apt install m4 make gcc

  ## Install Ocaml 5 and standard packages
  opam init -y
  opam update
  opam switch create 5.0.0
  eval $(opam env)
  opam install -y dune linenoise utop ocp-indent merlin

  ## Install packages required in Miking
  opam pin ocamlformat 0.24.1
  opam install -y pyml toml owl
fi

Miking and TreePPL

eval $(opam env)
gh repo clone miking-lang/miking
cd miking
make install
cd ..
gh repo clone miking-lang/miking-dppl
cd miking-dppl
export PATH="$HOME/.local/bin:$PATH"
make install
export MCORE_LIBS=coreppl=$HOME/.local/src/coreppl/
cd ..
gh repo clone https://github.com/treeppl/treeppl.git
cd treeppl
export MCORE_LIBS=$MCORE_LIBS:treeppl=$HOME/.local/src/treeppl/
make install
cd ..
pip install "git+https://github.com/treeppl/treeppl-python#egg=treeppl"

You can put the environment stuff in your .bashrc. Fire up gnome-text-editor ~/.bashrc and add

## Customizations for TreePPL
eval $(opam env)
export PATH="$HOME/.local/bin:$PATH"
export MCORE_LIBS="coreppl=$HOME/.local/src/coreppl/"
export MCORE_LIBS="$MCORE_LIBS:treeppl=$HOME/.local/src/treeppl/"

Running an example

pip install matplotlib
pip install seaborn
sudo apt-get install python3-tk

gh repo clone treeppl/treeppl-python
cd treeppl-python/examples
python3 coin.py

Setup dev under vscode

Install WSL Extension and/or Remote-extension-pack

https://code.visualstudio.com/docs/remote/wsl https://marketplace.visualstudio.com/items?itemName=ms-vscode-remote.vscode-remote-extensionpack

from a directory that you want to develop from do code .

vsenderov commented 1 month ago

Installation guide is now live. Asked @vresbok to take a look.