machine_initialization.sh

#!/usr/bin/env bash

# This script sets up a new Debian machine to be able to run the bash scripts
# in this repo. Inspiration was taken from here:
# https://hackmd.io/@yisun/BkT0RS87q

# Get the path to the directory this script is in.
INIT_SCRIPT_PATH="$(realpath "${BASH_SOURCE[-1]}")"
INIT_SCRIPT_DIRECTORY="$(dirname "$INIT_SCRIPT_PATH")"

############################################
########### ERROR HANDLING #################
############################################

# Inspiration taken from
# https://unix.stackexchange.com/questions/462156/how-do-i-find-the-line-number-in-bash-when-an-error-occured
# https://www.gnu.org/software/bash/manual/html_node/Bash-Variables.html

# Change this in sourcing script to get error messages.
ERR_MSG="UNKNOWN"

set -eE -o functrace

failure() {
    local lineno=$2
    local fn=$3
    local exitstatus=$4
    local cmd=$5
    local msg=$6
    local lineno_fns=${1% 0}
    if [[ "$lineno_fns" != "0" ]]; then
        lineno="${lineno} ${lineno_fns}"
    fi
    printf "\n####### ERROR #######\n\n"
    echo "  Failure in file:                    ${BASH_SOURCE[1]}"
    echo "  Function trace [line numbers]:      ${fn} [${lineno}]"
    echo "  Exit status:                        ${exitstatus}"
    echo "  Command that failed:                $cmd"
    echo "  Error message:                      $msg"
}
trap 'failure "${BASH_LINENO[*]}" "$LINENO" "${FUNCNAME[*]:-script}" "$?" "$BASH_COMMAND" "$ERR_MSG"' ERR

############################################
################ DEFAULTS ##################
############################################

DEFAULT_PTAU_SIZE=18
DEFAULT_SWAP_SIZE=400G

############################################
################ CLI FLAGS #################
############################################

# https://stackoverflow.com/questions/7069682/how-to-get-arguments-with-flags-in-bash#21128172

CLOUDWATCH=false
SWAP=false
PTAU=false
REPO=false
VERBOSE=false

print_usage() {
    printf "
Setup various software required to used the zk-proof-of-asset repo.

USAGE:
    ./machine_initialization.sh [FLAGS] [OPTIONS]

DESCRIPTION:
    This script does the following:
    1. Installs a bunch of software using the APT package manager
    2. Installs Rust via their install script
    3. Clones the Circom repo, and builds from source
    4. Clones the pyenv repo, and installs Python 3.10 & 3.6
    5. Clones the node repo, selects a particular commit, and builds from source
    6. Installs nvm via their install script, and installs latest npm
    7. Clones rapidsnark repo, and builds from source
    8. Installs pnpm via their install script
    9. Print out the 'export' command that needs to be run manually before trying to run full_workflow.sh

    The script can be run multiple times without repeating work e.g. if a command
    fails then you can run the script again and it will pick up where it left off.

TROUBLESHOOTING:

    If the npm install fails with this error:
      > Installing latest LTS version.
      > Downloading and installing node v20.11.0...
      > Binary download failed, trying source.
    Then run this command manually:
      export NVM_DIR=\"$HOME/.nvm\" && [ -s \"$NVM_DIR/nvm.sh\" ] && \. \"$NVM_DIR/nvm.sh\" && nvm install --lts.
    And restart the process.

    If the pnpm install fails with this error:
      > ==> Downloading pnpm binaries 8.15.1
      > WARN using --force I sure hope you know what you are doing
      > Copying pnpm CLI from /tmp/tmp.a13YBtCUZy/pnpm to /root/.local/share/pnpm/pnpm
      > ERR_PNPM_UNKNOWN_SHELL Could not infer shell type.
    Then run this command manually (with different tmp file):
      SHELL=\"$SHELL\"  /tmp/tmp.PZoYjFP8NI/pnpm setup --force
    And restart the process.

FLAGS:

     -c            AWS CloudWatch memory metrics
                   See this for more info https://stackoverflow.com/questions/42317062/how-to-monitor-ec2-instances-by-memory

     -P            Download ptau file number $DEFAULT_PTAU_SIZE & put in $HOME/zk-proof-of-assets

     -S            Create a swap file with default size: $DEFAULT_SWAP_SIZE

     -v            Print commands that are run (set -x)

     -h            Help

OPTIONS:

     -b <BRANCH>   Checkout <BRANCH> in $HOME/zk-proof-of-assets

     -p <NUM>      Download ptau file <NUM> & put in $HOME/zk-proof-of-assets
                   See all ptau files here https://github.com/iden3/snarkjs?tab=readme-ov-file#7-prepare-phase-2

     -r <DIR>      Clone zk-proof-of-assets repo into <DIR>
                   Also install dependencies and apply patches to dependencies

     -s <SIZE>     Create swap file of size <SIZE> (recommended for large circuits)
"
}

while getopts 'cvhr:Ss:b:Pp:' flag; do
    case "${flag}" in
    c) CLOUDWATCH=true ;;
    s)
        SWAP=true
        SWAP_SIZE_INPUT="${OPTARG}"
        ;;
    S) SWAP=true ;;
    r)
        REPO=true
        REPO_PARENT_DIR="${OPTARG}"
        ;;
    b) BRANCH="${OPTARG}" ;;
    p)
        PTAU=true
        PTAU_SIZE_INPUT="${OPTARG}"
        ;;
    P) PTAU=true ;;
    v) VERBOSE=true ;;
    h)
        print_usage
        exit 1
        ;;
    *)
        print_usage
        exit 1
        ;;
    esac
done

SWAP_SIZE=${SWAP_SIZE_INPUT:-$DEFAULT_SWAP_SIZE}
PTAU_SIZE=${PTAU_SIZE_INPUT:-$DEFAULT_PTAU_SIZE}

if $VERBOSE; then
    # print commands before executing
    set -x
fi

############################################
####### APT SOFTWARE INSTALLATION ##########
############################################

ERR_MSG="Initial setup failed"

sudo apt update && sudo apt upgrade -y

sudo apt install -y build-essential gcc pkg-config libssl-dev libgmp-dev libsodium-dev nasm nlohmann-json3-dev cmake m4 curl wget git time patch parallel jq bc

# for pyenv
sudo apt install -y --no-install-recommends make zlib1g-dev libbz2-dev libreadline-dev libsqlite3-dev llvm libncurses5-dev xz-utils tk-dev libxml2-dev libxmlsec1-dev libffi-dev liblzma-dev

############################################
############### CLOUDWATCH #################
############################################

# Setup memory metrics for CloudWatch
# https://stackoverflow.com/questions/42317062/how-to-monitor-ec2-instances-by-memory

if $CLOUDWATCH; then
    ERR_MSG="CloudWatch setup failed"

    sudo wget https://amazoncloudwatch-agent.s3.amazonaws.com/ubuntu/amd64/latest/amazon-cloudwatch-agent.deb
    sudo dpkg -i -E ./amazon-cloudwatch-agent.deb
    echo '{"metrics":{"metrics_collected":{"mem":{"measurement":["mem_used_percent"],"metrics_collection_interval":30}}}}' | sudo tee -a /opt/aws/amazon-cloudwatch-agent/bin/config.json >/dev/null
    sudo /opt/aws/amazon-cloudwatch-agent/bin/amazon-cloudwatch-agent-ctl -a fetch-config -m ec2 -c file:/opt/aws/amazon-cloudwatch-agent/bin/config.json -s
fi

############################################
############## SYSTEM CONFIG ###############
############################################

ERR_MSG="System config setup failed"

# Increase kernel memory map per process, and persist after reboot
# https://www.kernel.org/doc/Documentation/sysctl/vm.txt
# https://github.com/iden3/snarkjs/issues/397#issuecomment-1876700914
#
# NOTE this may not work if you are using a Docker container, and will give message (exit code 0)
# `sysctl: setting key "vm.max_map_count", ignoring: Read-only file system`
# https://stackoverflow.com/questions/23537560/docker-build-read-only-file-system
# 6553000 is enough for at least 256 signatures in layer 1 circuit
sudo sysctl -w vm.max_map_count=6553000
if ! grep "vm.max_map_count=6553000" /etc/sysctl.conf; then
    sudo sh -c 'echo "vm.max_map_count=6553000" >>/etc/sysctl.conf'
fi

if $SWAP; then
    # Increase swap memory, and persist after reboot
    sudo fallocate -l $SWAP_SIZE /swapfile
    sudo chmod 600 /swapfile
    sudo mkswap /swapfile
    sudo swapon /swapfile
    # https://serverfault.com/questions/967852/swapfile-mount-etc-fstab-swap-swap-or-none-swap
    echo '/swapfile whatever swap sw 0 0' | sudo tee -a /etc/fstab
fi

############################################
##### CUSTOM SOFTWARE INSTALLATION #########
############################################

# Rust
ERR_MSG="Rust setup failed"
cd "$HOME"
if [[ ! -f "$HOME/.cargo/env" ]]; then
    curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs >>./rustup-init.sh
    chmod +x ./rustup-init.sh
    ./rustup-init.sh -y --no-modify-path
    rm -f ./rustup-init.sh
fi
source "$HOME/.cargo/env"

# Circom
ERR_MSG="Circom setup failed"
cd "$HOME"
if [[ ! -d "$HOME/circom" ]]; then
    git clone https://github.com/iden3/circom.git
    cd circom

    # this is circom 2.1.8 (there seems to be problems compiling circuits for later versions)
    git checkout 207a712e7830774a66189ef453b77a04cec97995

    cargo build --release
    cargo install --path circom
fi

# Pyenv
ERR_MSG="Pyenv setup failed"
cd "$HOME"
if [[ ! -d "$HOME/pyenv" ]]; then git clone https://github.com/pyenv/pyenv.git; fi
export PATH="$HOME/pyenv/bin:$PATH"
eval "$(pyenv init -)"
if ! pyenv versions | grep "3\.10\."; then pyenv install 3.10; fi
pyenv global 3.10
if ! pyenv versions | grep "3\.6\."; then pyenv install 3.6; fi

# Patched node
ERR_MSG="Node setup failed"
cd "$HOME"
if [[ ! -f "$HOME/node/out/Release/node" ]]; then
    git clone https://github.com/nodejs/node.git
    cd node
    pyenv local 3.6
    git checkout 8beef5eeb82425b13d447b50beafb04ece7f91b1
    python configure.py
    make -j16
fi
export PATCHED_NODE_PATH=$HOME/node/out/Release/node

# NPM
ERR_MSG="NPM setup failed"
cd "$HOME"
if ! which npm; then
    if [[ ! -d "$HOME/.nvm" ]]; then
        wget -qO- https://raw.githubusercontent.com/nvm-sh/nvm/v0.39.1/install.sh | bash
    fi
    export NVM_DIR="$HOME/.nvm"
    [ -s "$NVM_DIR/nvm.sh" ] && \. "$NVM_DIR/nvm.sh"

    printf "
    NOTE This sometimes fails with:
    > Installing latest LTS version.
    > Downloading and installing node v20.11.0...
    > Binary download failed, trying source.

    In this case run this command manually:
    export NVM_DIR=\"\$HOME/.nvm\" && [ -s \"\$NVM_DIR/nvm.sh\" ] && \. \"\$NVM_DIR/nvm.sh\" && nvm install --lts.
    \n"

    nvm install --lts
fi

# Rapidsnark (x86_64 architecture only)
ERR_MSG="Rapidsnark setup failed"
cd "$HOME"
if [[ ! -f "$HOME/rapidsnark/package/bin/prover" ]]; then
    git clone https://github.com/iden3/rapidsnark.git
    cd rapidsnark
    git submodule init
    git submodule update
    ./build_gmp.sh host
    mkdir build_prover && cd build_prover
    cmake .. -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=../package
    make -j4 && make install
fi
export RAPIDSNARK_PATH=$HOME/rapidsnark/package/bin/prover

# PNPM
ERR_MSG="PNPM setup failed"
cd "$HOME"
if [[ ! -f "$HOME/.local/share/pnpm/pnpm" ]]; then
    printf "
    NOTE This sometimes fails with:
    > ==> Downloading pnpm binaries 8.15.1
    > WARN using --force I sure hope you know what you are doing
    > Copying pnpm CLI from /tmp/tmp.a13YBtCUZy/pnpm to /root/.local/share/pnpm/pnpm
    > ERR_PNPM_UNKNOWN_SHELL Could not infer shell type.

    In this case just run this manually (with different tmp file):
    SHELL=\"\$SHELL\"  /tmp/tmp.PZoYjFP8NI/pnpm setup --force
    \n"

    curl -fsSL https://get.pnpm.io/install.sh | sh -
fi
export PNPM_HOME="$HOME/.local/share/pnpm"
export PATH="$PNPM_HOME:$PATH"
pnpm add npx -g
pnpm add snarkjs -g

# Setup repo.
if $REPO; then
    ERR_MSG="Repo setup failed"
    REPO_DIR="$REPO_PARENT_DIR/zk-proof-of-assets"

    if [[ ! -d "$REPO_DIR" ]]; then
        cd "$REPO_PARENT_DIR"
        git clone https://github.com/silversixpence-crypto/zk-proof-of-assets
    fi

    cd "$REPO_DIR"

    if [[ ! -z "$BRANCH" ]]; then
        git switch -c "$BRANCH" origin/"$BRANCH"
    fi

    pnpm i

    git submodule init
    git submodule update

    # Workaround for a problem in the circom compiler
    # https://github.com/iden3/circom/issues/230
    BATCH_ECDSA_DIR="$REPO_DIR/git_modules/batch-ecdsa"
    patch -u "$BATCH_ECDSA_DIR/circuits/batch_ecdsa.circom" -i ./batch-ecdsa.patch

    # Naming conflict resolution between ed25519-circom & circom-pairing
    ED25519_CIRCOM_DIR="$REPO_DIR/git_modules/ed25519-circom"
    cd "$ED25519_CIRCOM_DIR"
    git apply "$REPO_DIR"/ed25519-circom.patch
    cd -

    pip install -r requirements.txt

    cargo build --release
fi

# TODO instead of checking if the file exists rather check its checksum,
# because the download might have only gotten partway
if $PTAU; then
    if [[ $PTAU_SIZE -gt 27 ]]; then
        ptau_url=https://storage.googleapis.com/zkevm/ptau/powersOfTau28_hez_final.ptau
    else
        ptau_url=https://storage.googleapis.com/zkevm/ptau/powersOfTau28_hez_final_"$PTAU_SIZE".ptau
    fi

    wget "$ptau_url"
fi

# This is so that the user can set these env vars in their shell.
# This is the 2nd method mentioned here:
# https://stackoverflow.com/questions/16618071/can-i-export-a-variable-to-the-environment-from-a-bash-script-without-sourcing-i/16619261#16619261
# The 1st method in the above tends to break things, so try to avoid that.
echo "export SNARKJS_PATH=\$HOME/snarkjs/cli.js && export RAPIDSNARK_PATH=\$HOME/rapidsnark/package/bin/prover && export PATCHED_NODE_PATH=\$HOME/node/out/Release/node && export PATH=\$PATH:\$HOME/pyenv/bin && eval \"\$(pyenv init -)\" && export PNPM_HOME=\$HOME/.local/share/pnpm && export PATH=\$PNPM_HOME:\$PATH && export NVM_DIR=\$HOME/.nvm && [ -s \$NVM_DIR/nvm.sh ] && \. \$NVM_DIR/nvm.sh && source $HOME/.cargo/env"