Installing CMEnt
================

Install the package and its R dependencies using BiocManager:

    if (!requireNamespace("BiocManager", quietly = TRUE))
        install.packages("BiocManager")
    BiocManager::install("CMEnt")

Optional annotation, reference genome, and analysis packages depend on the
requested functionality. CMEnt reports missing packages and gives installation
commands; package functions do not install software. Install the indicated
packages before repeating the analysis.

External command-line tools
===========================

Tabix-backed file workflows use bgzip and tabix from HTSlib. On Linux and macOS,
conversion also uses GNU sort, with head, tail, and a POSIX shell. These tools
must be visible on the PATH inherited by R. Matrix-based workflows do not need
the external tools. R does not install them automatically.

Linux (Debian/Ubuntu):

    sudo apt-get -o Acquire::Retries=3 update
    sudo apt-get -o Acquire::Retries=3 install tabix coreutils

macOS (with Homebrew already installed):

    brew install htslib coreutils
    export PATH="$(brew --prefix coreutils)/libexec/gnubin:$PATH"

The PATH setting selects GNU sort rather than the macOS sort command. Start R
from that shell, or arrange for the same PATH in the R session.

Windows:

The Windows conversion path sorts in R and does not use GNU sort. It still
requires compatible bgzip.exe and tabix.exe on PATH for compression/indexing.
Native Windows execution of those external tools has not been verified here.
For these file workflows, use the Linux Docker image or run R inside WSL2,
installing the Linux tools there as described above. Installing tools only
inside WSL2 does not make them available to a native Windows R session.

Check tool discovery in R with:

    Sys.which(c("bgzip", "tabix", "sort"))

Container setup
===============

The source distribution includes the Dockerfile and related scripts under
inst/docker. From the repository root, build with:

    docker build -f inst/docker/Dockerfile -t cment .

The container entrypoint installs missing optional R packages for the selected
command, inputs, and genome before starting analysis. Viewer setup uses the
saved metadata to prepare packages for that analysis's plots and interactions.
This requires network access and a writable R library when packages are missing.
It does not enable installation from CMEnt package functions or worker processes.

Running inst/docker/run_cment.R directly provides the CLI without container
installation. For arbitrary interactive R scripts, install their optional
dependencies explicitly before calling CMEnt.