[Feature]: Add a one-click Visualizer setup script

[MillironX]

Summary

Create a script that goes from a clean install to fully functional Visualizer with just a double-click from the user that is output with the results.

More Info

Context

For many of our target users, installing npm and maintaining it is difficult. This would make it graphical and "easy."

Possible implementation

The script would do, in pseudocode:

nodeversion = Check for recent (>= 14) Node.js install
if ( nodeversion < Node.js LTS || nodeversion == null) {
    nbin = $(whereis n)
    if ( nbin == null ) {
        exec n-install-script.sh
        PATH += /path/to/n
    }
   $(n install lts)
}
npm install
npm start

A Windows version (batchfile) should also be included.

[MillironX]

Maybe it would be better to use a static site generator, instead. The workflow would look like:

1. Pipeline compiles output filenames/metadata, creates datasheet (CSV/JSON/YAML/TOML)
2. Pipeline passes datasheet to new process, SITE_GENERATOR
3. SITE_GENERATOR uses a static site generator to create the visualizer webpages, with links to the proper output files
4. SITE_GENERATOR has its publishDir directive and output: set to publish all the correct files to the results directory

flowchart TD
    subgraph Processes
    A[KRAKEN_CLASSIFICATION]
    B[FASTQC]
    C[MINIMAP_ALIGMENT]
    D[HAPLINK_HAPLOTYPES]
    E[RAXML_TREE]
    end
    subgraph Analysis
    X[MULTIQC]
    Y["Metadata compiler (Nextflow)"] -- CSV or YAML --> Z[STATIC_GENERATOR]
    end
    subgraph Output
    O[params.outdir]
    end
    A -- kreport --> X
    B -- zip --> X
    C -- "path(bam)" --> Y
    C -- bam --> O
    D -- YAML --> Y
    D -- YAML --> O
    E -- "path(newick)" --> Y
    E -- newick --> O
    X -- HTML --> O
    Z -- HTML --> O

Possible site generators:

Name	Repo	Pros	Cons
Jekyll	https://github.com/jekyll/jekyll	Ubiquitous
Hugo	https://github.com/gohugoio/hugo	Powerful, includes optimization, integrates well with NPM	Extending is hard (no sequence support)
Franklin.jl	https://github.com/tlienart/Franklin.jl	Easy extension with Julia (BioJulia support)	Doesn't seem to support data files?

[MillironX]

A static site generator for the final product is out due to Cross origin resource sharing issues. However, a static site generator could still create individual pieces, namely

• The genome chart (#14)
• The IGV page
• Phylogenetic tree page

Then the server would simply serve up the remote resources, including

• MultiQC Report
• Krona chart (?)
• BAM/Fasta files

As much as I hate CDNs, I think they will be required for this page in the end. Then the only dependencies will be a web server (Node.js, Python, Julia even). CDN resources include

• Bootstrap+Popper
• JQuery
• IGV
• Krona (img+js)
• phylotree.js
• Font Awesome
• freeze-table

[MillironX]

Starting with #40, the output is mostly a static page, but it still needs a server. Linux and Mac users can use Python3's http.server. We should try to compile a binary from https://github.com/tlienart/LiveServer.jl for Windows users.