pandora_typst_report/report.typ

#let cr_colors = (
  dark_grey: rgb("#333333"), beige: rgb("#fdf0d5"), light_grey: rgb("#eeeeee"), dark_red: rgb("#780000"), red: rgb("#c1121f"), blue: rgb("#669bbc"), dark_blue: rgb("#003049"), green: rgb("#29bf12"),
)

#import "@preview/fletcher:0.5.1" as fletcher: diagram, node, edge
#import "@local/svg-emoji:0.1.0": setup-emoji, noto, github
#import "@preview/metro:0.3.0": *
#import "@preview/cetz:0.2.2"

#show: setup-emoji.with(font: noto)

#set page(paper: "a4", fill: cr_colors.light_grey, footer: locate(loc => [
  #set text(10pt)
  #if loc.page() != 1 {
    align(right, counter(page).display("1 / 1", both: true))
  }
]))

#show heading.where(level: 1): it => [
  #set align(center)
  #set text(fill: cr_colors.dark_blue)
  #it.body
  #v(18pt)
]
#show image: set text(font: "FreeSans")

#set text(size: 16pt, font: "Futura", fill: cr_colors.dark_blue)
#let contigs = (
  "chr1", "chr2", "chr3", "chr4", "chr5", "chr6", "chr7", "chr8", "chr9", "chr10", "chr11", "chr12", "chr13", "chr14", "chr15", "chr16", "chr17", "chr18", "chr19", "chr20", "chr21", "chr22", "chrX", "chrY",
)
#let parseCustomDate(dateString) = {
  let parts = dateString.split("T")
  let datePart = parts.at(0).replace("-", "/")
  let timePart = parts.at(1).split(":")
  let hour = timePart.at(0)
  let minute = timePart.at(1)

  return datePart + " " + hour + "h" + minute
}

#let formatString(input) = {
  let words = input.split("_")
  let capitalizedWords = words.map(word => {
    if word.len() > 0 {
      upper(word.first()) + word.slice(1)
    } else {
      word
    }
  })
  capitalizedWords.join(" ")
}
#let si-fmt(val, precision: 1, sep: "\u{202F}", binary: false) = {
  let factor = if binary { 1024 } else { 1000 }
  let gt1_suffixes = ("k", "M", "G", "T", "P", "E", "Z", "Y")
  let lt1_suffixes = ("m", "μ", "n", "p", "f", "a", "z", "y")
  let scale = ""
  let unit = ""

  if type(val) == content {
    if val.has("text") {
      val = val.text
    } else if val.has("children") {
      val = val.children.map(content => content.text).join()
    } else {
      panic(val.children.map(content => content.text).join())
    }
  }
  // if val contains a unit, split it off
  if type(val) == str {
    unit = val.find(regex("(\D+)$"))
    val = float(val.split(unit).at(0))
  }

  if calc.abs(val) > 1 {
    for suffix in gt1_suffixes {
      if calc.abs(val) < factor {
        break
      }
      val /= factor
      scale += " " + suffix
    }
  } else if val != 0 and calc.abs(val) < 1 {
    for suffix in lt1_suffixes {
      if calc.abs(val) > 1 {
        break
      }
      val *= factor
      scale += " " + suffix
    }
  }

  let formatted = str(calc.round(val, digits: precision))

  formatted + sep + scale.split().at(-1, default: "") + unit
}

#let reportCoverage(prefix) = {
  image(prefix + "_global.svg", width: 100%)
  for contig in contigs {
    heading(level: 4, contig)
    let path = prefix + "_" + contig
    image(path + "_chromosome.svg")
    let data = json(path + "_stats.json")
    grid(
      columns: (1fr, 2fr), gutter: 3pt, align(
        left + horizon,
      )[
        #set text(size: 12pt)
        #table(
          stroke: none, columns: (auto, 1fr), gutter: 3pt, [Mean], [#calc.round(data.mean, digits: 2)], [Standard dev.], [#calc.round(data.std_dev, digits: 2)], ..data.breaks_values.map(r => ([#r.at(0)], [#calc.round(r.at(1) * 100, digits: 1)%])).flatten(),
        )
      ], align(right, image(path + "_distrib.svg", width: 100%)),
    )

    parbreak()
  }
}

#let reportBam(path) = {
  let data = json(path)
  table(
    gutter: 3pt, stroke: none, columns: (auto, 1fr), ..for (key, value) in data {
      if key != "cramino" and key != "composition" and key != "path" and key != "modified" {
        ([ #formatString(key) ], [ #value ])
      } else if key == "modified" {
        ([ Modified Date (UTC) ], [ #parseCustomDate(value) ])
      } else if key == "composition" {
        ([ Run(s) ], [
          #for (i, v) in value.enumerate() {
            if i > 0 [ \ ]
            [#v.at(0).slice(0, 5): #calc.round(v.at(1), digits: 0)%]
          }
        ])
      } else if key == "cramino" {
        for (k, v) in value {
          if k == "normalized_read_count_per_chromosome" {} else if k != "path" and k != "checksum" and k != "creation_time" and k != "file_name" {
            let k = formatString(k)
            let v = if type(v) == "integer" { si-fmt(v) } else { v }
            ([ #k ], [ #v ])
          } else {
            ()
          }
        }.flatten()
      } else {
        ()
      }
    }.flatten(),
  )
}

#let formatedReadCount(path) = {
  let data = json(path)
  let data = data.cramino.normalized_read_count_per_chromosome
  let res = ()
  for contig in contigs {
    res.push(data.at(contig))
  }
  res.push(data.at("chrM"))
  return res
}

#let printReadCount(diag_path, mrd_path) = {
  let index = 14;
  let c = contigs
  c.push("chrM")
  let diag = formatedReadCount(diag_path)
  let mrd = formatedReadCount(mrd_path)
  c.insert(0, "")
  diag.insert(0, "diag")
  mrd.insert(0, "mrd")
  let arrays1 = (c.slice(0, index), diag.slice(0, index), mrd.slice(0, index))
  table(
    columns: arrays1.at(0).len(), ..arrays1.map(arr => arr.map(item => [#item])).flatten(),
  )

  let arrays2 = (c.slice(index), diag.slice(index), mrd.slice(index))
  arrays2.at(0).insert(0, "")
  arrays2.at(1).insert(0, "diag")
  arrays2.at(2).insert(0, "mrd")

  table(
    columns: arrays2.at(0).len(), ..arrays2.map(arr => arr.map(item => [#item])).flatten(),
  )
}

#let variantsFlow(path) = {
  import fletcher.shapes: diamond, parallelogram, chevron
  let data = json(path)
  diagram(
    spacing: (1pt, 60pt), node-fill: gradient.radial(cr_colors.light_grey, cr_colors.blue, radius: 300%), node-stroke: cr_colors.dark_blue + 1pt, edge-stroke: 1pt, node-inset: 14pt, node(
      (0.2, 0), [Variants MRD: #num(data.vcf_stats.n_tumoral_init)], corner-radius: 2pt, extrude: (0, 3), name: <input_mrd>,
    ), node(
      (1.8, 0), [Variants Diag: #num(data.vcf_stats.n_constit_init)], corner-radius: 2pt, extrude: (0, 3), name: <input_diag>,
    ), node(
      (1, 1), align(center)[Variant in MRD ?], shape: diamond, name: <is_in_mrd>,
    ), edge(<input_mrd>, "s", <is_in_mrd>, "-|>"), edge(<input_diag>, "s", <is_in_mrd>, "-|>"), edge(<is_in_mrd>, <is_low_mrd>, "-|>", [Yes], label-pos: 0.8), node(
      (0.25, 2), [MRD variant depth \ < 4 ?], shape: diamond, name: <is_low_mrd>,
    ), edge(<is_low_mrd>, <low_mrd>, "-|>"), node(
      (0, 3), [Low MRD depth: #num(data.vcf_stats.n_low_mrd_depth)], shape: parallelogram, name: <low_mrd>,
    ), edge(<is_in_mrd>, <next>, "-|>", [No], label-pos: 0.8), node(
      (1.85, 2), [To BAM filter: #num(data.bam_stats.n_lasting)], shape: chevron, extrude: (-3, 0), name: <next>, stroke: cr_colors.green,
    ), edge(<is_low_mrd>, <homo>, "-|>"), node((1.5, 3), [VAF = 100% ?], shape: diamond, name: <homo>), edge(<homo>, <constit>, "-|>", [Yes], label-pos: 0.5, bend: -80deg), edge(<homo>, <chi>, "-|>", [No], label-pos: 0.6), node(
      (1.5, 4), [$#sym.chi^2$ VAF MRD vs Diag ?], shape: diamond, name: <chi>,
    ), edge(<chi>, <constit>, "-|>", label-pos: 0.8), node(
      (1, 5), [Constit: #num(data.vcf_stats.n_constit)], shape: parallelogram, name: <constit>,
    ), edge(<chi>, <loh>, "-|>", [p < 0.01], label-pos: 0.8), node(
      (2, 5), [LOH: #num(data.vcf_stats.n_loh)], shape: parallelogram, name: <loh>,
    ),
  )
}

#let bamFilter(path) = {
  import fletcher.shapes: diamond, parallelogram, hexagon
  let data = json(path)
  diagram(
    spacing: (1pt, 60pt), node-fill: gradient.radial(cr_colors.light_grey, cr_colors.blue, radius: 300%), node-inset: 14pt, node-stroke: cr_colors.dark_blue + 1pt, edge-stroke: 1pt, node(
      (0.75, 0), [Variants not in MRD VCF: #num(data.bam_stats.n_lasting)], corner-radius: 2pt, extrude: (0, 3), name: <input_mrd>,
    ), edge(<input_mrd>, <depth>, "-|>"), node((0.75, 1), [MRD alignement depth ?], shape: diamond, name: <depth>), edge(<depth>, <low_depth>, "-|>", [< 4]), node(
      (0, 2), [Low MRD depth: #num(data.bam_stats.n_low_mrd_depth)], shape: parallelogram, name: <low_depth>,
    ), edge(<depth>, <seen>, "-|>"), node(
      (0.75, 3), [Alt. base seen in MRD pileup ?], shape: diamond, name: <seen>,
    ), edge(<seen>, <constit>, "-|>", [Yes]), node(
      (0, 4), [Constit: #num(data.bam_stats.n_constit)], shape: parallelogram, name: <constit>,
    ), edge(<seen>, <is_div>, "-|>", [No]), node(
      (1.1, 4), [Sequence #sym.plus.minus 20nt \ diversity ?], shape: diamond, name: <is_div>,
    ), edge(<is_div>, <low_div>, "-|>", [entropy < 1.8]), node(
      (0.25, 5), [Low diversity, artefact: #num(data.bam_stats.n_low_diversity)], shape: parallelogram, name: <low_div>,
    ), edge(<is_div>, <somatic>, "-|>"), node(
      (1.75, 5), [Somatic: #num(data.bam_stats.n_somatic)], shape: hexagon, extrude: (-3, 0), name: <somatic>, stroke: cr_colors.green,
    ),
  )
}

#let barCallers(path) = {
  import cetz.draw: *
  import cetz.chart
  
  let json_data = json(path).variants_stats
  let data = json_data.find(item => item.name == "callers_cat")
  let chart_data = data.counts.pairs().sorted(key: x => -x.at(1))
  let max_value = chart_data.first().at(1)

  cetz.canvas(
    length: 80%,
    {
      set-style(axes: (bottom: (tick: (label: (angle: 45deg, anchor: "north-east")))))
      chart.columnchart(
        chart_data,
        size: (1, 1),
      )
  })
}

#set heading(numbering: (..numbers) => {
  if numbers.pos().len() >= 2 and numbers.pos().len() <= 3 {
    numbering("1.1", ..numbers.pos().slice(1))
  }
})

#heading(level: 1, outlined: false)[Whole Genome Sequencing Report]

#outline(title: "Table of Contents", depth: 3)
#pagebreak()
== Identity
Camara

== Alignement
#grid(
  columns: (1fr, 1fr), gutter: 3pt, [
    ==== Diagnostic sample
    #set text(size: 11pt)
    #reportBam(
      "/Turbine-pool/LongReads/report/BECERRA/report/data/BECERRA_diag_hs1_info.json",
    )
  ], [
    ==== MRD sample
    #set text(size: 11pt)
    #reportBam(
      "/Turbine-pool/LongReads/report/BECERRA/report/data/BECERRA_mrd_hs1_info.json",
    )
    #set footnote(numbering: n => { " " })
    #footnote[Values computed by #link("https://github.com/wdecoster/cramino")[cramino] v0.14.5
    ]
  ],
)

#pagebreak()
=== Normalized read count by chromosome
#[
  #set text(size: 10pt)
  #printReadCount(
    "/Turbine-pool/LongReads/report/BECERRA/report/data/BECERRA_diag_hs1_info.json", "/Turbine-pool/LongReads/report/BECERRA/report/data/BECERRA_mrd_hs1_info.json",
  )
]

=== Coverage by chromosome
==== Proportion at given depth by chromosome
#reportCoverage("/Turbine-pool/LongReads/report/BECERRA/report/data/scan/BECERRA")
#set footnote(numbering: n => { " " })
#footnote[Values computed by Pandora development version]
#pagebreak()

== Variants
=== Variants calling
==== VCF filters
#pad(
  left: -0.8cm, top: 0.8cm, variantsFlow(
    "/Turbine-pool/LongReads/report/BECERRA/report/data/CAMARA_variants_stats.json",
  ),
)

==== BAM filters
#pad(
  top: 0.8cm, bamFilter(
    "/Turbine-pool/LongReads/report/BECERRA/report/data/CAMARA_variants_stats.json",
  ),
)
#pagebreak()

=== Somatic variants
#barCallers("/Turbine-pool/LongReads/report/BECERRA/report/data/CAMARA_variants_stats.json")

=== Selected Variants


#pagebreak()
== Conclusion
hello ???
#lorem(150)
❌
#emoji.rocket