QUESTDB_website/content/multipledataset.html

---
title: "Multiple dataset"
date: 2019-09-27 16:41
draft: false
---
<link rel="stylesheet" type="text/css" href="/css/modal.css" />
<link rel="stylesheet" type="text/css" href="/css/form.css" />
<link rel="stylesheet" type="text/css" href="/css/table.css" />
<script src="/js/data.js" type="text/javascript"></script>
<script src="/js/loadAllData.js" type="text/javascript"></script>
<script src="/js/getFullDataPath.js" type="text/javascript"></script>
<script src="/js/getTextFromFile.js" type="text/javascript"></script>
<script src="/js/trueTypeOf.js" type="text/javascript"></script>
<script src="/js/uniq.js"></script>
<script src="/js/processingIndicator.js"></script>
<script src="/js/noNan.js"></script>
<script src="/js/websiteFile.js"></script>
<script src="/js/getSets.js"></script>
<script src="https://cdn.rawgit.com/larsgw/citation.js/archive/citation.js/citation-0.4.0-9.js"
  type="text/javascript"></script>
<script src="/js/DOICache.js"></script>
<script src="/js/nestedCheckbox.js"></script>
<script src="https://unpkg.com/simple-statistics@7.0.2/dist/simple-statistics.min.js"></script>
<script>
  window.onload = async () => {
    const LatexInline = ['\\(', '\\)']
    var cbextl = $("#cb_exTypeList")
    var allc = $("#AllowList:last-child")
    for (const [name, value] of excitationTypes.All) {
      var description = null
      switch (Number(value)) {
        case Number(excitationTypes.VALENCE):
          description = name.toLowerCase()
          break;
        case Number(excitationTypes.RYDBERG):
          const word = name
          description = word.charAt(0).toUpperCase() + word.substring(1).toLowerCase()
          break;
        case Number(excitationTypes.Singulet):
          description = name.toLowerCase()
          break;
        case Number(excitationTypes.Doublet):
          description = name.toLowerCase()
          break;
        default:
          description = null;
          break;
      }
      $("<li/>").text(LatexInline[0] + value.LaTeX + LatexInline[1] + (description == null ? '' : ": " + description)).insertBefore(allc)
      var cbli = $("<li/>")
      $("<input/>", { type: "checkbox", id: "cb_" + name, name: name, value: Number(value) }).change(nestedCheckbox_change).appendTo(cbli);
      $('<label />', { 'for': 'cb_' + name, text: LatexInline[0] + value.LaTeX + LatexInline[1] }).appendTo(cbli);
      cbextl.append(cbli);
    }
    await MathJax.typesetPromise();
    delete (cbextl)
    delete (allc)
    window.doiCache = new DOICache()
    $('#form_dat > fieldset > input[type=radio]').on('change', async function (event) {
      processingIndicator.isActive = true
      $("#Geom_th,#GSGeom_th,#ESGeom_th").each(function () {
        $(this).attr("hidden", true)
      })
      var result = $(this).val();
      var proms = [];
      switch (result) {
        case "abs":
          proms = getAbsFilesName().map((f) => AbsDataFile.loadAsync(f));
          $("#Geom_th").attr("hidden", false)
          break;
        case "fluo":
          proms = getFluoFilesName().map((f) => FluoDataFile.loadAsync(f));
          $("#Geom_th").attr("hidden", false)
          break;
        case "zpe":
          proms = getZPEFilesName().map((f) => ZPEDataFile.loadAsync(f));
          $("#GSGeom_th,#ESGeom_th").each(function () {
            $(this).attr("hidden", false)
          })
          break;
      }
      window.dats = await Promise.all(proms)
      processingIndicator.isActive = false
      reloadFileSelector()
      await reloadSelect()
    })
  }
  function reloadFileSelector() {
    $("#customFiles_input").attr("value", "").attr("disabled", false)
  }

  async function reloadSelect() {
    processingIndicator.isActive = true
    radioval = $('#form_dat > fieldset > input[name=datatype]:checked').val();
    const files = $("#customFiles_input")[0].files
    for (file of [...files]) {
      switch (radioval) {
        case "abs":
          var dat = await AbsDataFile.loadAsync(file)
          window.dats.push(dat)
          break;
        case "fluo":
          dat = await FluoDataFile.loadAsync(file)
          window.dats.push(dat)
          break;
        case "ZPE":
          dat = await ZPEDataFile.loadAsync(file)
          window.dats.push(dat)
          break;
      }
    }
    $('#form_dat > fieldset > div > select,#sel_ref').each(function () {
      $(this).find('option[value!=""]').remove()
    })
    $('#form_dat > fieldset > div > select').each(async function () {
      const vals = uniq(window.dats.map((d) => { return d[$(this).attr("name")] }))
      var sets = await getSets()
      for (const val of vals) {
        var op = $("<option/>", {
          value: JSON.stringify(val)
        })
        if ($(this).attr("name") == "DOI" && sets.has(val.toString())) {
          $(op).text(sets.get(val.toString()))
        }
        else {
          $(op).text(val.toString())
        }
        op.appendTo(this)
      }
    })
    $('#form_dat > input[type="submit"').prop('disabled', false);
    processingIndicator.isActive = false
    await reloadContent()
  }
  async function reloadStat() {
    processingIndicator.isActive = true
    var stb = $("#stat_table > tbody")
    $("#graph_div").empty()
    $(stb).empty()
    var refstr = $("#sel_ref option:selected").val()
    var sdatdic = new Map()
    for (const d of window.filtData) {
      const key = JSON.stringify([d.code, d.method, d.DOI])
      if (!(sdatdic.has(key))) {
        sdatdic.set(key, new Map())
      }
      for (const exc of d.excitations) {
        const key2 = JSON.stringify([d.molecule, exc.initial, exc.final])
        const keydic = sdatdic.get(key)
        if ((!exc.isUnsafe || window.filterParams[0]) && (exc.type & window.filterParams[1])) {
          if (!(keydic.has(key2))) {
            keydic.set(key2, [])
          }
          keydic.get(key2).push(exc.value) 
        }
      }
    }
    var sdic = new Map()
    for (const [key, sdat] of sdatdic) {
      for (const [key2, exval] of sdat) {
        if (!(sdic.has(key))) {
          sdic.set(key, [])
        }
        sdic.get(key).push(exval - ((sdatdic.has(refstr)) ? sdatdic.get(refstr).get(key2) : NaN))
      }
    }
    sdic.delete(refstr)
    var graphdat = []
    for (const [keystr, vals] of sdic) {
      row = $("<tr/>")
      key = JSON.parse(keystr)
      key.shift()
      Reflect.setPrototypeOf(key[0], method.prototype)
      Reflect.setPrototypeOf(key[1], DOI.prototype)
      for (const el of key) {
        var td = $("<td/>")
        if (trueTypeOf(el) == DOI.name) {
          var publi = doiCache.get(el.string, true)
          const sets = await getSets()
          if (sets.has(el.toString())) {
            $(td).append(sets.get(el.toString()))
          }
          $("<a/>", {
            href: el.url,
            target: "_blank"
          }).html(publi.format('citation', {
            format: 'html',
            lang: 'en-US'
          })).appendTo(td)
        }
        else {
          $(td).text(el)
        }
        $(row).append(td)
      }
      const noNanVals = (vals.every((v) => Number.isNaN(v))) ? vals : (vals.filter((v) => !Number.isNaN(v)))
      const avals = noNanVals.map(v => Math.abs(v))
      for (const val of [ss.min(noNanVals), ss.max(noNanVals), ss.mean(noNanVals), ss.mean(avals), ss.median(noNanVals), ss.median(avals), ss.rootMeanSquare(noNanVals), ss.variance(noNanVals), ss.standardDeviation(noNanVals)]) {
        $("<td/>").text(noNanPrecision(val, 3)).appendTo(row)
      }
      $(stb).append(row)
      var box = {
        x: noNanVals,
        amean: ss.mean(avals).toFixed(3),
        name: key[0].toString() + " MAD : " + ss.mean(avals).toPrecision(4),
        type: 'box',
        boxmean: 'sd'
      };
      graphdat.push(box)
    }
    var layout = {
      paper_bgcolor: 'rgba(0,0,0,0)',
      plot_bgcolor: 'rgba(0,0,0,0)',
      gap: 0,
      legend: {
        traceorder: 'reversed',
      },
      xaxis: {
        title: {
          text: 'Energy (eV)',
        }
      },
      bgcolor: '#E2E2E2',
      bordercolor: '#FFFFFF',
      borderwidth: 2,
      // autosize: false,
      width: 850,
      height: 500,
      margin: {
        l: 0,
        r: 10,
        b: 15,
        t: 20,
        pad: 0,
      },
    }
    Plotly.newPlot('graph_div', graphdat, layout);
    processingIndicator.isActive = false
  }
  async function reloadContent() {
    window.filterParams = [$("#cb_unsafe").prop("checked"), Array.from(Array.from($("#cb_exTypeList>li>input[type=checkbox]:checked")).map(el => parseInt(el.value))).reduce((pv, cv) => pv + cv, 0)]
    doiCache.clear()
    processingIndicator.isActive = true
    const LatexInline = ['\\(', '\\)']
    window.filtData = window.dats
    $('#form_dat > fieldset > div > select').each(function () {
      const prop = $(this).attr("name")
      const values = $(this).val()
      window.filtData = window.filtData.filter((d) => {
        if (typeof values == "undefined" || values == null) {
          return false
        }
        return values.includes(JSON.stringify(d[prop]))
      })
    })
    await window.doiCache.addRange(window.filtData.map((d) => d.DOI.string))
    for (const el of uniq(window.filtData.map(d => [d.code, d.method, d.DOI]))) {
      op = $("<option/>", {
        value: JSON.stringify(el)
      }).text(el[1])
      if (el[1].name.includes("TBE")) {
        $(sel_ref).prepend(op)
      }
      else {
        $(sel_ref).append(op)
      }
    }
    $(sel_ref).prop("selectedIndex", 0);
    var data = $("#data")
    $(data).empty();
    if (window.filtData.length > 0) {
      var sets = await getSets()
      for (const doi of doiCache.keys()) {
        paperdata = window.filtData.filter((d) => {
          return d.DOI.string == doi
        })
        var methods = uniq(paperdata.map(d => d.method))
        const sortedMethods = methods.sort((a, b) => {
          if (a.isTBE && !b.isTBE) {
            return -1
          }
          else if (a.isTBE && b.isTBE) {
            if (!a.isCorrected && b.isCorrected) {
              return -1
            }
            else {
              return 1
            }
          }
          else if (!a.isTBE && b.isTBE) {
            return 1
          }
          else {
            return 0
          }
        })
        var div = $('<div/>').appendTo(data)
        $("<h1/>").text(sets.get(doi)).appendTo(div)
        var table = $("<table/>").addClass("datatable").appendTo(div)
        var head = $("<tr/>")
        $("<thead/>").append(head).appendTo(table)
        var tbody = $("<tbody/>").appendTo(table)
        head.append(["Molecule", "Transition"].map(x => $("<th/>").text(x)))
        head.append(sortedMethods.map(x => $("<th/>").text(x.toString())))
        datadic = new Map()
        for (const dat of paperdata) {
          const key1 = dat.molecule;
          if (!datadic.has(key1)) {
            datadic.set(key1, new Map())
          }
          const key3 = JSON.stringify(dat.method)
          for (const ex of dat.excitations) {
            Reflect.setPrototypeOf(ex.type,excitationType.prototype)
            if (window.filterParams[1] & ex.type != 0) {
              const key2 = JSON.stringify([ex.initial, ex.final, ex.type])
              if (!datadic.get(key1).has(key2)) {
                datadic.get(key1).set(key2, new Map())
              }
              datadic.get(key1).get(key2).set(key3, [ex.value, ex.isUnsafe])
            }
          }
        }
        for (const [molecule, moldat] of datadic.entries()) {
          var printmol = true;
          for (const [jsonex, exdat] of moldat.entries()) {
            const ex = JSON.parse(jsonex)
            Reflect.setPrototypeOf(ex[0], state.prototype)
            Reflect.setPrototypeOf(ex[1], state.prototype)
            Reflect.setPrototypeOf(ex[2], excitationType.prototype)
            var tr = $("<tr/>")
            var tdmol = $("<td/>")
            if (printmol) {
              $(tdmol).text(molecule)
              printmol = false
            }
            tr.append(tdmol)
            $("<td/>").text(String.raw`${LatexInline[0]}${ex[0].toLaTeX()} \rightarrow ${ex[1].toLaTeX()} (${ex[2].LaTeX}) ${LatexInline[1]}`).appendTo(tr)
            var entries = Array.from(exdat.entries())
            for (const method of sortedMethods) {
              td = $("<td/>")
              kv = entries.find(x => {
                return JSON.stringify(method) == x[0]
              })
              if (kv !== undefined) {
                const [val, unsafe] = kv[1]
                if (unsafe) {
                  td.append($("<div/>", { title: "unsafe value", style: "float: left" }).text('⚠'))
                }
                if (unsafe && !$("#cb_unsafe").is(':checked')) {
                  td.append($("<s/>").append(val))
                }
                else {
                  td.append(val)
                }
              }
              tr.append(td)
            }
            tbody.append(tr)
          }
        }
      }
      await MathJax.typesetPromise()
    }
    await reloadStat()
  }
  async function submitdat(){
    window.onbeforeunload = () => {return '';}
    await reloadContent();
  }
</script>
{{< getDataFilesName >}}
<p>
  In this page you can compare multiple dataset and have somme statistics to evaluate method or basis accuracy.
  For this follow this steps
  <ol>
    <li>Select a value type (Absorption, fluorescence or \(\Delta \text{ZPE}\))</li>
    <li>You can add custom files if you want</li>
    <li>Choose each parameters (of course you can select multiples values for each)</li>
    <ul>
      <li>Molecule</li>
      <li>Method</li>
      <li>Publication</li>
      <li>Allow:</li>
      <ul id="AllowList">
        <li>Unsafe</li>
      </ul>
    </ul>
    <li>Select a reference from <strong>already selected data</strong> (by default first is selected -it's the <abbr
        title="Theoretical best estimate">TBE</abbr> if present- is already selected)</li>
    <li>Now you can see the list of selected data and some statistics about these data</li>
  </ol>
</p>
<form id="form_dat" action="javascript:submitdat()" method="post">
  <fieldset>
    <legend>Data type</legend>
    <input type="radio" id="abs" value="abs" name="datatype">
    <label for="abs">Absorption</label>
    <input type="radio" id="fluo" value="fluo" name="datatype">
    <label for="fluo">Fluorescence</label>
    <input type="radio" id="zpe" value="zpe" name="datatype">
    <label for="zpe">\(\Delta \text{ZPE}\)</label>
  </fieldset>
  <fieldset>
    <div>
      <label for="customFiles_input">Add custom data file</label>
      <input type="file" multiple onchange="reloadSelect()" id="customFiles_input" disabled=true></input>
    </div>
  </fieldset>
  <fieldset class="table">
    <div>
      <label for="mol_select">Molecule</label>
      <select multiple id="mol_select" name="molecule"></select>
    </div>
    <div>
      <label for="method_select">Method</label>
      <select multiple id="method_select" name="method"></select>
    </div>
    <div>
      <label for="DOI_select">Publication</label>
      <select multiple id="DOI_select" name="DOI"></select>
    </div>
  </fieldset>
  <fieldset id="excitationFilter">
    <ul>
      <li>
        <h1>excitations types</h1>
        <input type="checkbox" onchange="nestedCheckbox_change(event)" id="cb_exType_All"></input>
        <label for="cb_exType_All"> All</label>
      </li>
      <ul id="cb_exTypeList">
      </ul>
      <li id="li_cb_unsafe">
        <input type="checkbox" id="cb_unsafe" name="unsafe">
        <label for="cb_unsafe">Unsafe</label>
      </li>
  </fieldset>
  <input type="submit" disabled=true value="Load"></input>
</form>
<form id="form_ref">
  <label for="sel_ref">Reference</label>
  <select id="sel_ref" onchange="reloadStat()"></select>
</form>
<section id="data">
</section>
<section>
  <table id="stat_table" class="datatable">
    <thead>
      <th>Method</th>
      <th>Publication</th>
      <th>Min</th>
      <th>Max</th>
      <th>Mean</th>
      <th>Absolute Mean</th>
      <th>Median</th>
      <th>Absolute Median</th>
      <th>RMS</th>
      <th>Variance</th>
      <th>Standard deviation</th>
    </thead>
    <tbody>
    </tbody>
    <div id="graph_div"></div>
  </table>
</section>
{{< waitModal >}}