pokemon-color-search/nearest.js

// Selectors + DOM Manipulation
const getColorInputNode = () => document.getElementById("color-input");
const getMetricDropdownNode = () => document.getElementById("metric");
const getClusterChoiceDropdownNode = () => document.getElementById("image-summary");
const getClusterScaleToggleNode = () => document.getElementById("scale-by-cluster-size");
const getClusterMeanWarning = () => document.getElementById("cluster-mean-warning");
const getIncludeXToggleNode = () => document.getElementById("include-x");
const getNormQYToggleNode = () => document.getElementById("norm-q-y");
const getCloseCoeffSliderNode = () => document.getElementById("close-coeff");
const getCloseCoeffDisplayNode = () => document.getElementById("close-coeff-display");
const getLimitSliderNode = () => document.getElementById("num-poke");
const getLimitDisplayNode = () => document.getElementById("num-poke-display");
const getNameInputNode = () => document.getElementById("pokemon-name");
const getScoreListJABNode = () => document.getElementById("best-list-jab");
const getScoreListRGBNode = () => document.getElementById("best-list-rgb");
const getSearchSpaceDisplayNode = () => document.getElementById("search-space-display");
const getSearchListNode = () => document.getElementById("search-list");
const getHideableControlNodes = () => document.querySelectorAll(".hideable_control");
const getQJABDisplay = () => document.getElementById("q-vec-jab");
const getQRGBDisplay = () => document.getElementById("q-vec-rgb");
const getObjFnDisplay = () => document.getElementById("obj-fn");

const clearNodeContents = node => { node.innerHTML = ""; };
const hideCustomControls = () => getHideableControlNodes()
  .forEach(n => n.setAttribute("class", "hideable_control hideable_control--hidden"));
const showCustomControls = () => getHideableControlNodes()
  .forEach(n => n.setAttribute("class", "hideable_control"));

// Vector Math
const vectorDot = (u, v) => u.map((x, i) => x * v[i]).reduce((x, y) => x + y);
const vectorSqMag = v => vectorDot(v, v);
const vectorMag = v => Math.sqrt(vectorSqMag(v));
const vectorSqDist = (u, v) => vectorSqMag(u.map((x, i) => x - v[i]));
const vectorDist = (u, v) => Math.sqrt(vectorSqDist(u, v));
const vectorNorm = v => { const n = vectorMag(v); return [ n, v.map(c => c / n) ]; };

// Angle Math
const angleDiff = (a, b) => { const raw = Math.abs(a - b); return raw < 180 ? raw : (360 - raw); };
const rad2deg = 180 / Math.PI;

// Conversions
const jab2hex = jab => d3.jab(...jab).formatHex();
const rgb2hex = rgb => d3.rgb(...rgb).formatHex();
const jab2hue = ([, a, b]) => rad2deg * Math.atan2(b, a);
const rgb2hue = rgb => d3.hsl(d3.rgb(...rgb)).h || 0;
const hex2rgb = hex => {
  const { r, g, b } = d3.color(hex);
  return [r, g, b];
};

// Arg Compare
const argComp = comp => ra => ra.map((x, i) => [x, i]).reduce((a, b) => comp(a[0], b[0]) > 0 ? b : a)[1];
const argMin = argComp((a, b) => a - b);
const argMax = argComp((a, b) => b - a);

// Pre-Compute Data
const computeVectorData = (vector, toHex, toHue) => {
  const [ magnitude, unit ] = vectorNorm(vector);
  return {
    vector,
    magnitude,
    magSq: magnitude * magnitude,
    unit,
    hex: toHex(vector),
    hue: toHue(vector),
  };
};

const computeStats = (inertia, trueMeanVec, kMeanStruct, toHex, toHue) => ({
  inertia,
  trueMean: computeVectorData(trueMeanVec, toHex, toHue),
  kMeans: kMeanStruct.slice(0, 3).map(z => computeVectorData(z, toHex, toHue)),
  kWeights: kMeanStruct[3],
  largestCluster: argMax(kMeanStruct[3]),
  smallestCluster: argMin(kMeanStruct[3]),
});

const pokemonColorData = database.map(({
  name, xJAB, xRGB, yJAB, yRGB, zJAB, zRGB,
}) => ({
  name,
  jabStats: computeStats(xJAB, yJAB, zJAB, jab2hex, jab2hue),
  rgbStats: computeStats(xRGB, yRGB, zRGB, rgb2hex, rgb2hue),
}));

const pokemonLookup = new Fuse(pokemonColorData, { keys: [ "name" ] });

// Color Calculations
const getContrastingTextColor = rgb => vectorDot(rgb, [0.3, 0.6, 0.1]) >= 128 ? "#222" : "#ddd";

const readColorInput = () => {
  const colorInput = "#" + (getColorInputNode()?.value?.replace("#", "") ?? "FFFFFF");

  if (colorInput.length !== 7) {
    return;
  }

  const rgb = d3.color(colorInput);
  const { J, a, b } = d3.jab(rgb);

  return {
    jabData: computeVectorData([ J, a, b ], jab2hex, jab2hue),
    rgbData: computeVectorData([ rgb.r, rgb.g, rgb.b ], rgb2hex, rgb2hue),
  };
};

// State
const state = {
  metric: null,
  clusterChoice: null,
  includeScale: null,
  includeX: null,
  normQY: null,
  closeCoeff: null,
  numPoke: null,
  searchTerm: null,
  searchSpace: null,
  targetColor: null,
  searchResults: null,
  clusterToggles: {},
  currentScores: {},
};

// Metrics
const getBestKMean = (stats, q) => argMin(stats.kMeans.map((z, i) => vectorDist(z.vector, q.vector) / stats.kWeights[i]));
const getWorstKMean = (stats, q) => argMax(stats.kMeans.map((z, i) => vectorDist(z.vector, q.vector) / stats.kWeights[i]));

const getScale = weight => state.includeScale ? (1 / weight) : 1;

const summarySelectors = [
  // true mean
  stats => [stats.trueMean, 1],
  // largest cluster
  stats => [stats.kMeans[stats.largestCluster], getScale(stats.kWeights[stats.largestCluster])],
  // smallest cluster
  stats => [stats.kMeans[stats.smallestCluster], getScale(stats.kWeights[stats.smallestCluster])],
  // best fit cluster
  (stats, q) => {
    const best = getBestKMean(stats, q);
    return [stats.kMeans[best], getScale(stats.kWeights[best])];
  },
  // worst fit cluster
  (stats, q) => {
    const worst = getWorstKMean(stats, q);
    return [stats.kMeans[worst], getScale(stats.kWeights[worst])];
  },
];

const selectedSummary = (stats, q) => summarySelectors[state.clusterChoice](stats, q);

// TODO unfortunately the addition of the scaling factor means we have to compute *real*
// metrics instead of cheaper approximations. would be nice to fix this
const metrics = [
  // RMS
  (stats, q) => { 
    const [ mean, scale ] = selectedSummary(stats, q);
    return (stats.inertia - 2 * vectorDot(mean.vector, q.vector)) * scale; 
  },
  // mean angle
  (stats, q) => {
    const [ mean, scale ] = selectedSummary(stats, q);
    return rad2deg * Math.acos(vectorDot(mean.unit, q.unit)) * scale;
  },
  // mean dist
  (stats, q) => {
    const [ mean, scale ] = selectedSummary(stats, q);
    // TODO I know there's some way to avoid recalculation here but I'm just too lazy right now
    return vectorDist(mean.vector, q.vector) * scale;
  },
  // hue angle
  (stats, q) => {
    const [ mean, scale ] = selectedSummary(stats, q);
    return angleDiff(mean.hue, q.hue) * scale;
  },
  // max inertia
  (stats, q) => {
    const [ , scale ] = selectedSummary(stats, q);
    return -stats.inertia * scale;
  },
  // custom
  (stats, q) => {
    const [ mean, scale ] = selectedSummary(stats, q);
    return (
      (state.includeX ? stats.inertia : 0) 
      - 
      state.closeCoeff * vectorDot(
        mean[state.normQY ? "unit" : "vector"],
        state.normQY ? q.unit : q.vector,
      )
    ) * scale;
  },
];

const scorePokemon = pkmn => ({
  jab: metrics[state.metric](pkmn.jabStats, state.targetColor.jabData),
  rgb: metrics[state.metric](pkmn.rgbStats, state.targetColor.rgbData),
});

const calcDisplayMetrics = (meanData, q) => ({
  theta: rad2deg * Math.acos(vectorDot(q.unit, meanData.unit)),
  delta: vectorDist(q.vector, meanData.vector),
  phi: angleDiff(q.hue, meanData.hue),
});

// Math Rendering
const renderQVec = (q, node, sub) => {
  node.innerHTML = TeXZilla.toMathMLString(String.raw`\vec{q}_{\text{${sub}}} = \left(\text{${q.join(", ")}}\right)`);
};

const mathArgBest = (mxn, arg) => `\\underset{${arg}}{\\arg\\${mxn}}`;

const mathDefinitions = {
  "main-definition": String.raw`
    \begin{aligned}
      \vec{\mu}\left(P\right) &= \frac{1}{\left|P\right|}\sum_{p\in P}{\vec{p}} \\
      I\left(P\right) &= \frac{1}{\left|P\right|}\sum_{p\in P}{\left|\left|\vec{p}\right|\right|^2} \\
      \delta\left(P\right) &= \left|\left| \vec{q} - \vec{\mu}\left(P\right) \right|\right| \\
      \theta\left(P\right) &= \angle \left(\vec{q}, \vec{\mu}\left(P\right)\right) \\
      \vec{x}_{\perp} &= \text{oproj}_{\left\{\vec{J}, \vec{L}\right\}}{\vec{x}} \\
      \phi\left(P\right) &= \angle \left(\vec{q}_{\perp}, \vec{\mu}\left(P\right)_{\perp} \right) \\
      \sigma\left(P\right) &= \sqrt{E\left[\left(\vec{q} - P\right)^2\right]} \\
        &= \sqrt{\frac{1}{|P|}\sum_{p \in P}{\left|\left|\vec{p} - \vec{q}\right|\right|^2}}
    \end{aligned}
  `,
  "cluster-definition": String.raw`
    \begin{aligned}
      \left\{P_1, P_2, P_3\right\} &= ${mathArgBest("max", String.raw`\left\{P_1, P_2, P_3\right\}`)} \sum_{i=1}^3 \sum_{p\inP_i} \left|\left| \vec{p} - \vec{\mu}\left(P_i\right) \right|\right|^2 \\
      \pi_i &= \frac{\left|P_i\right|}{\left|P\right|} \\
      M\left(P\right) &= ${mathArgBest("max", "P_i")} \left( \left|P_i\right| \right) \\
      m\left(P\right) &= ${mathArgBest("min", "P_i")} \left( \left|P_i\right| \right) \\
      \alpha\left(P\right) &= ${mathArgBest("min", "P_i")} \left[ \frac{1}{\pi_i} \left|\left| \vec{q} - \vec{\mu}\left(P_i\right) \right|\right| \right] \\
      \omega\left(P\right) &= ${mathArgBest("max", "P_i")} \left[ \frac{1}{\pi_i} \left|\left| \vec{q} - \vec{\mu}\left(P_i\right) \right|\right| \right]
    \end{aligned}
  `,
};

const includeScaleFactor = muArg => state.clusterChoice > 0 && state.includeScale ? String.raw`\frac{\left|P\right|}{\left|${muArg}\right|}` : ""

const metricText = [
  muArg => String.raw`
    ${mathArgBest("min", "P")}\left[
      ${includeScaleFactor(muArg)}I\left(P\right) 
      - 2\vec{q}\cdot ${includeScaleFactor(muArg)}\vec{\mu}\left(${muArg}\right)
    \right]`,
  muArg => String.raw`${mathArgBest("min", "P")}\left[${includeScaleFactor(muArg)}\angle \left(\vec{q}, \vec{\mu}\left(${muArg}\right)\right)\right]`,
  muArg => String.raw`${mathArgBest("min", "P")}\left[${includeScaleFactor(muArg)}\left|\left|\vec{q} - \vec{\mu}\left(${muArg}\right) \right|\right|\right]`,
  muArg => String.raw`${mathArgBest("min", "P")}\left[${includeScaleFactor(muArg)}\angle \left(\vec{q}_{\perp}, \vec{\mu}\left(${muArg}\right)_{\perp}\right)\right]`,
  muArg => String.raw`${mathArgBest("min", "P")}\left[-${includeScaleFactor(muArg)}I\left(P\right)\right]`,
].map(s => muArg => TeXZilla.toMathML(s(muArg)));

const muArgs = [
  "P",
  String.raw`M\left(P\right)`,
  String.raw`m\left(P\right)`,
  String.raw`\alpha\left(P\right)`,
  String.raw`\omega\left(P\right)`,
];

const renderVec = math => String.raw`\vec{${math.charAt(0)}}${math.substr(1)}`;
const renderNorm = vec => String.raw`\frac{${vec}}{\left|\left|${vec}\right|\right|}`;
const updateObjective = () => {
  const muArg = muArgs[state.clusterChoice];
  let tex = metricText?.[state.metric]?.(muArg);
  if (!tex) {
    const { includeX, normQY, closeCoeff } = state;
    if (!includeX && closeCoeff === 0) {
      tex = TeXZilla.toMathML(String.raw`\text{Malamar-ness}`);
    } else {
      const qyMod = normQY ? renderNorm : c => c;
      tex = TeXZilla.toMathML(String.raw`
        ${mathArgBest("min", "P")}
        \left[
          ${includeX ? String.raw`${includeScaleFactor(muArg)}I\left(P\right)` : ""}
          ${closeCoeff === 0 ? "" : String.raw`
              -
              ${closeCoeff}
              ${qyMod("\\vec{q}")}
              \cdot
              ${includeScaleFactor(muArg)}${qyMod(String.raw`\vec{\mu}\left(${muArg}\right)`)}
          `}
        \right]
      `);
    }
  }
  const objFnNode = getObjFnDisplay();
  clearNodeContents(objFnNode);
  objFnNode.appendChild(tex);
};

// Pokemon Rendering
const stripForm = ["flabebe", "floette", "florges", "vivillon", "basculin", "furfrou", "magearna"];

const getSprite = pokemon => {
  pokemon = pokemon
    .replace("-alola", "-alolan")
    .replace("-galar", "-galarian")
    .replace("darmanitan-galarian", "darmanitan-galarian-standard");
  if (stripForm.find(s => pokemon.includes(s))) {
    pokemon = pokemon.replace(/-.*$/, "");
  }
  return `https://img.pokemondb.net/sprites/sword-shield/icon/${pokemon}.png`;
};

const renderCluster = ({
  index, big, small, best, worst, pi, theta, delta, phi, hex, vector,
}) => `
  <div
    class="pkmn_tile-cluster"
    style="grid-area: k${index + 1}; color: ${getContrastingTextColor(hex2rgb(hex))}; background-color: ${hex};"
  >
    <div class="pkmn_tile-cluster-top_label" style="grid-area: lbl;">
      ${index === big ? "<span>M</span>" : ""}
      ${index === small ? "<span>m</span>" : ""}
      ${index === best ? "<span>α</span>" : ""}
      ${index === worst ? "<span>ω</span>" : ""}
    </div>
    <div class="pkmn_tile-cluster-stat_label" style="grid-area: mu;">μ =</div>
    <div class="pkmn_tile-cluster-stat_label" style="grid-area: pi;">π =</div>
    <div class="pkmn_tile-cluster-stat_label" style="grid-area: th;">θ =</div>
    <div class="pkmn_tile-cluster-stat_label" style="grid-area: dl;">δ =</div>
    <div class="pkmn_tile-cluster-stat_label" style="grid-area: ph;">ϕ =</div>
    <div style="grid-area: mux">${hex}</div>
    <div style="grid-area: muv; justify-self: center;">(${vector.map(c => c.toFixed(2)).join(", ")})</div>
    <div style="grid-area: piv">${(pi * 100).toFixed(1)}%</div>
    <div style="grid-area: thv">${theta.toFixed(2)}°</div>
    <div style="grid-area: dlv">${delta.toFixed(2)}</div>
    <div style="grid-area: phv">${phi.toFixed(2)}°</div>
  </div>
`;

const getPokemonRenderer = targetList =>  (name, stats, q, score, idPostfix) => {
  let sigma, metrics, kMeanInfo, kMeanResults;
  if (q) {
    sigma = Math.sqrt(stats.inertia - 2 * vectorDot(stats.trueMean.vector, q.vector) + q.magSq)
    metrics = calcDisplayMetrics(stats.trueMean, q)
    kMeanInfo = {
      big: stats.largestCluster,
      small: stats.smallestCluster,
      best: getBestKMean(stats, q),
      worst: getWorstKMean(stats, q), // TODO yeah yeah this is a recalc whatever
    };
    kMeanResults = stats.kMeans.map(k => calcDisplayMetrics(k, q));
  } else {
    // no target color, just do all zeros
    sigma = 0;
    metrics = { theta: 0, delta: 0, phi: 0 };
    kMeanInfo = { big: 0, small: 0, best: 0, worst: 0 };
    kMeanResults = [ metrics, metrics, metrics ];
  }
  const clusterToggleId = `reveal_clusters-${name}-${idPostfix}`;

  const li = document.createElement("li");
  li.innerHTML = `
    <div class="pkmn_tile">
      <img class="pkmn_tile-img" src="${getSprite(name)}" />
      <span class="pkmn_tile-name">
        ${name.split("-").map(part => part.charAt(0).toUpperCase() + part.substr(1)).join(" ")}
      </span>
      <div class="pkmn_tile-fn">
        ${score?.toFixed(3) ?? ""}
      </div>
      <input 
        type="checkbox" 
        ${state.clusterToggles?.[clusterToggleId] ? "checked" : ""} 
        id="${clusterToggleId}"
        onchange="state.clusterToggles['${clusterToggleId}'] = event.target.checked"
        class="pkmn_tile-reveal_clusters" 
        role="button"
      >
      <label class="pkmn_tile-reveal_clusters_label" for="${clusterToggleId}">
        <div class="pkmn_tile-reveal_clusters_label--closed">►</div>
        <div class="pkmn_tile-reveal_clusters_label--open">▼</div>
      </label>
      <div
        class="pkmn_tile-true_mean"
        style="color: ${getContrastingTextColor(hex2rgb(stats.trueMean.hex))}; background-color: ${stats.trueMean.hex};"
      >
          <div class="pkmn_tile-true_mean-value">
              <div class="pkmn_tile-true_mean-mu_label">μ =</div>
              <div class="pkmn_tile-true_mean-mu_hex">${stats.trueMean.hex}</div>
              <div class="pkmn_tile-true_mean-mu_vec">
                (${stats.trueMean.vector.map(c => c.toFixed(2)).join(", ")})
              </div>
          </div>
          <div class="pkmn_tile-true_mean-stat pkmn_tile-true_mean-inertia">
            𝖨 = ${stats.inertia.toFixed(2)}
          </div>
          <div class="pkmn_tile-true_mean-stat pkmn_tile-true_mean-stat-sigma">
            σ = ${sigma.toFixed(2)}
          </div>
          <div class="pkmn_tile-true_mean-stat pkmn_tile-true_mean-stat-theta">
            θ = ${metrics.theta.toFixed(2)}°
          </div>
          <div class="pkmn_tile-true_mean-stat pkmn_tile-true_mean-stat-delta">
            δ = ${metrics.delta.toFixed(2)}
          </div>
          <div class="pkmn_tile-true_mean-stat pkmn_tile-true_mean-stat-phi">
            ϕ = ${metrics.phi.toFixed(2)}°
          </div>
      </div>
      ${stats.kMeans.map((data, index) => renderCluster({
          index,
          ...kMeanInfo,
          pi: stats.kWeights[index],
          ...kMeanResults[index],
          hex: data.hex,
          vector: data.vector,
      })).join("\n")}
    </div>
  `;
  targetList.appendChild(li);
};

// Update Search Results
const renderSearch = () => {
  const resultsNode = getSearchListNode();
  const append = getPokemonRenderer(resultsNode);
  clearNodeContents(resultsNode);
  const argMapper = state.searchSpace === "RGB"
    ? pkmn => [pkmn.rgbStats, state.targetColor?.rgbData, state.currentScores?.[pkmn.name]?.rgb ?? null]
    : pkmn => [pkmn.jabStats, state.targetColor?.jabData, state.currentScores?.[pkmn.name]?.jab ?? null]
  state.searchResults?.forEach(pkmn => append(
    pkmn.name, ...argMapper(pkmn), "search"
  ));
};

// Scoring
const renderScored = () => {
  const jabList = getScoreListJABNode();
  const appendJAB = getPokemonRenderer(jabList);
  const rgbList = getScoreListRGBNode();
  const appendRGB = getPokemonRenderer(rgbList);

  // extract best CIECAM02 results
  const bestJAB = pokemonColorData
    .sort((a, b) => state.currentScores[a.name].jab - state.currentScores[b.name].jab)
    .slice(0, state.numPoke);
  clearNodeContents(jabList);
  bestJAB.forEach(data => appendJAB(
    data.name, data.jabStats, state.targetColor.jabData, state.currentScores[data.name].jab, "jab"
  ));

  // extract best RGB results
  const bestRGB = pokemonColorData
    .sort((a, b) => state.currentScores[a.name].rgb - state.currentScores[b.name].rgb)
    .slice(0, state.numPoke);
  clearNodeContents(rgbList);
  bestRGB.forEach(data => appendRGB(
    data.name, data.rgbStats, state.targetColor.rgbData, state.currentScores[data.name].rgb, "rgb"
  ));
};

const rescore = () => {
  if (!state.targetColor) {
    return;
  }

  state.currentScores = {};
  pokemonColorData.forEach(data => {
    state.currentScores[data.name] = scorePokemon(data);
  });

  // update displays
  renderScored();

  // update the rendered search results as well
  renderSearch();
};

// Listeners
const onColorChanged = skipScore => {
  const readColor = readColorInput();
  if (readColor) {
    state.targetColor = readColor;

    renderQVec(state.targetColor.jabData.vector.map(c => c.toFixed(3)), getQJABDisplay(), "Jab");
    renderQVec(state.targetColor.rgbData.vector.map(c => c.toFixed()), getQRGBDisplay(), "RGB");

    const rootElem = document.querySelector(":root");
    rootElem.style.setProperty("--background", state.targetColor.rgbData.hex);
    rootElem.style.setProperty("--highlight", getContrastingTextColor(state.targetColor.rgbData.vector));

    if (!skipScore) {
      rescore();
    }
  }
};

const onRandomColor = () => {
  const color = [Math.random(), Math.random(), Math.random()].map(c => c * 255);
  getColorInputNode().value = d3.rgb(...color).formatHex();
  onColorChanged(); // triggers rescore
};

const onCustomControlsChanged = skipScore => {
  state.includeX = getIncludeXToggleNode()?.checked ?? false;
  state.normQY = getNormQYToggleNode()?.checked ?? false;
  state.closeCoeff = parseFloat(getCloseCoeffSliderNode()?.value ?? 2);

  getCloseCoeffDisplayNode().innerHTML = state.closeCoeff;
  updateObjective();

  if (!skipScore) {
    rescore();
  }
}

const checkClusterMeanWarning = () => {
  const warning = getClusterMeanWarning();
  const unhidden = warning.getAttribute("class").replaceAll("hide", "");
  if (state.clusterChoice !== 0 && (state.metric === 0 || state.metric === 4)) {
    warning.setAttribute("class", unhidden);
  } else {
    warning.setAttribute("class", unhidden + " hide");
  }
}

const checkScaleByClusterToggle = () => {
  const toggle = getClusterScaleToggleNode()?.parentNode;
  const unhidden = toggle.getAttribute("class").replaceAll("hide", "");
  if (state.clusterChoice !== 0) {
    toggle.setAttribute("class", unhidden);
  } else {
    toggle.setAttribute("class", unhidden + " hide");
  }
}

const onScaleByClusterChanged = skipScore => {
  state.includeScale = getClusterScaleToggleNode()?.checked ?? true;

  updateObjective();

  if (!skipScore) {
    rescore();
  }
}

const onClusterChoiceChanged = skipScore => {
  const clusterChoice = getClusterChoiceDropdownNode()?.selectedIndex ?? 0;
  if (clusterChoice === state.clusterChoice) {
    return;
  }
  state.clusterChoice = clusterChoice;
  checkClusterMeanWarning();
  checkScaleByClusterToggle();
  updateObjective();
  if (!skipScore) {
    rescore();
  }
}

const onMetricChanged = skipScore => {
  const metric = getMetricDropdownNode()?.selectedIndex ?? 0;
  if (metric === state.metric) {
    return;
  }
  state.metric = metric;
  checkClusterMeanWarning();
  checkScaleByClusterToggle();
  if (state.metric === 5) { // Custom
    showCustomControls();
    onCustomControlsChanged(skipScore); // triggers rescore
  } else {
    hideCustomControls();
    updateObjective();
    if (!skipScore) {
      rescore();
    }
  }
};

const onLimitChanged = skipRenderScore => {
  state.numPoke = parseInt(getLimitSliderNode()?.value ?? 10);

  getLimitDisplayNode().textContent = state.numPoke;

  if (!skipRenderScore) {
    renderScored();
  }
};

const onSearchChanged = () => {
  state.searchTerm = getNameInputNode()?.value?.toLowerCase() ?? "";
  if (state.searchTerm.length === 0) {
    state.searchResults = [];
  } else {
    state.searchResults = pokemonLookup
      .search(state.searchTerm, { limit: 10 })
      .map(({ item }) => item);
  }
  renderSearch();
};

const onSearchSpaceChanged = () => {
  const old = state.searchSpace ?? "Jab";
  state.searchSpace = old === "RGB" ? "Jab" : "RGB";
  getSearchSpaceDisplayNode().textContent = old;
  renderSearch();
};

const onRandomPokemon = () => {
  getNameInputNode().value = "";
  state.searchResults = Array.from({ length: 10 }, () => pokemonColorData[Math.floor(Math.random() * pokemonColorData.length)]);
  renderSearch();
};

const onPageLoad = () => {
  // render static explanations
  Object.entries(mathDefinitions).forEach(([id, tex]) => {
    document.getElementById(id)?.appendChild(TeXZilla.toMathML(tex));
  });

  // fake some events but don't do any scoring
  onColorChanged(true);
  onMetricChanged(true);
  onClusterChoiceChanged(true);
  onScaleByClusterChanged(true);
  onLimitChanged(true);
  // then do a rescore directly, which will do nothing unless old data was loaded
  rescore();
  // finally render search in case rescore didn't
  onSearchChanged();
};