2f:T4f6,rules = [ fuzzy.and({lineWidth:"equal", lineExtent:"equal"}, {blockBody:"clear"}), fuzzy.or({lineExtent:'above'},{blockBody:'vague'}) // fuzzy.or({lineWidth:"small"}, {blockBody:"maybe"}), // fuzzy.or({lineWidth:"large"}, {blockBody:"vague"}), /* fuzzy.or({sameLine:"clear"}, {blockBody:"clear"}), fuzzy.or({lineEnd: "clear"}, {blockBody:'clear'}), fuzzy.or({numberEnd: "clear"}, {blockBody:'maybe'}), fuzzy.or({duplicate:"clear"}, {blockBody:"vague"}), fuzzy.and({lineWidth: "basic",hangIndent: "equal"}, {blockBody: "clear"}), fuzzy.and({lineWidth: "basic",leadIndent: "equal"}, {blockBody: "clear"}), fuzzy.and({lineWidth: "basic",leadIndent:"equal",hangIndent:"equal"},{blockBody:"clear"}), fuzzy.and({lineWidth: "small",leadExtent:"equal",hangExtent:"equal"},{blockBody:"maybe"}),*/ //fuzzy.and({leadIndent:"equal",hangIndent:"equal"},{blockBody:"clear"},2), //fuzzy.and({leadSpacing:"apart",hangSpacing:"apart"},{blockBody:"vague"},2), //fuzzy.and({lineWidth:"basic",leadSpacing: "equal",hangSpacing:"equal"}, {blockBody: "clear"}), // prev punc + current enum == sentence // multiple absence of left coords where other pages have them ~~ table, figure, list, etc. ];30:T44a,classed = temp.outer.map((d,i,f)=> { let out = fuzzy.defuzz( inputs2, outputs[0], rules, { lineWidth:d.dims.span[2], charWidth:d.dims.norm[5], lineExtent:d.dims.norm[4] /*lineWidth:d.dims.norm[2], leadWidth: d.lead.norm[2],hangWidth: d.hang.norm[2], lineExtent:d.dims.norm[4], leadHeight: d.lead.norm[3],hangHeight: d.hang.norm[3], charWidth:d.dims.norm[5], pageTurn:d.turn, leadIndent:d.lead.norm[0], hangIndent:d.hang.norm[0], leadExtent:d.lead.norm[4], hangExtent:d.hang.norm[4], sameLine: Math.abs(d.dims.span[1] - f[i-1]?.dims.span[1]) < 2, // tune titlecase: d.caps.trim().length / d.line.length > 0.5, numerated: d.enum, duplicate: d.dupl, numberEnd: d.enum && (f[i-1].punc == '.' || f[i-1].punc == '”'), lineEnd:d.punc === '.' || d.punc === ":" */}, // fuzzy.meanOfMaximaStrategy() ) for(let item in out) { out[item] = parseFloat(out[item].toFixed(2)) } return {...out,line:d.line,d} })//.sort((b,a)=>a.title - b.title) //.filter(d=>d.blockBody<=0.5 && d.dupl) 31:T1e6c,temp = { let t1 = performance.now(); let input = open(file); let empty = (n) => ({ min: [], med: [], mod: [], max: [], avg: [], var: [], dev: [], // geo: Array.from({ length:n },()=>1), sum: Array.from({ length:n },()=>0), nom: Array.from({ length:n },()=>0), all: Array.from({ length:n },()=> { let self = new Float32Array(input.length); let set; return Object.defineProperties(self, { percentile: { get: ()=>(p) => self[~~((input.length - 1) * p)]}, occurance: { get: ()=>(set = set ?? [...new Set(self)], (p)=> set.find((_,i,f)=> i/(f.length-1) >= p ))}, fullRank: { get: ()=>(v, band = 0)=> self.findIndex((a,i,f)=> a >= v-(v*band) || i == (f.length-1)) / (self.length-1)}, flatRank: { get: ()=>(set = set ?? [...new Set(self)], (v, band = 0)=> set.findIndex((a,i,f)=> a >= v-(v*band) || i == (f.length-1)) / (set.length-1))} }); }), }); let slots = 6, n = -1; let stats = empty(slots); let delts = empty(slots); let freqs = new Map(); let outer = []; let abs = Math.abs; let lines = input.map((entry,i,f)=> { let prev = Math.max(i-1,0)// ~i-1 >= 0? i-1 : 0 let mult = entry.line.length - entry.line.replaceAll('\n','').length; // let p =0.5, a=0.5; let diff = abs(entry.top - f[prev].height - f[prev].top); return { page:entry.page,line:entry.line, dims: Array.of(entry.left,entry.top,entry.width,entry.height, Math.log10(diff + 1), entry.width/entry.line.length/mult, ), mult }}) for(let entry of lines) { ++ n let above = lines[n-1] || entry; let below = lines[n+1] || entry; let lead = [],//new Float32Array(slots), hang = [];//new Float32Array(slots); let hash = entry.line.slice(0,26).replace(/(\p{L})|\n|.{4}|.*/gu,'$1') //[...new Set(entry.line.replace(/\P{L}/gu,''))].join('') freqs.set(hash, (freqs.get(hash) ?? 0)+1); let j = -1; for(let point of entry.dims) { ++ j stats.sum[j] += point ; // stats.geo[j] += Math.log(point) ; stats.max[j] = point > stats.max[j] ? point : stats.max[j] ?? point stats.min[j] = point < stats.min[j] ? point : stats.min[j] ?? point //let deltaA = abs(point - above.dims[j]) let deltaB = abs(point - below.dims[j]) ** 0.5 delts.sum[j] = delts.sum[j] + deltaB || deltaB; delts.max[j] = deltaB > delts.max[j] ? deltaB : delts.max[j] ?? deltaB delts.min[j] = deltaB < delts.min[j] ? deltaB : delts.min[j] ?? deltaB // lead[j] = deltaA; hang[j] = deltaB; stats.all[j][n] = point; delts.all[j][n] = deltaB; } outer.push({...entry,lead,hang,hash}) } stats.avg = stats.sum.map(s=> s/(n+1)) delts.avg = delts.sum.map(d=> d/(n+1)) // stats.geo = stats.geo.map(s=> Math.exp(s/(n+1))) // delts.geo = delts.geo.map(d=> Math.exp(d/(n+1))) stats.all = stats.all.map(s=> s.sort()) delts.all = delts.all.map(d=> d.sort()) stats.med = stats.all.map(s=> median(s)) delts.med = delts.all.map(d=> median(d)) stats.mod = stats.all.map(s=> mode(s)) delts.mod = delts.all.map(s=> mode(s)) stats.var = stats.all.map((s,j)=> variance(s,stats.avg[j])) delts.var = delts.all.map((d,j)=> variance(d,delts.avg[j])) stats.dev = stats.var.map(s=> Math.sqrt(s)) delts.dev = delts.var.map(s=> Math.sqrt(s)) stats.nom = stats.nom.map((_,j)=> [ // denominators stats.max[j]/stats.dev[j], stats.max[j]-stats.min[j], //stats.all[j].occurance(0.75) - stats.all[j].occurance(0.25) // IQR ]) delts.nom = delts.nom.map((_,j)=> [ delts.max[j]/delts.dev[j], delts.max[j]/*min is always 0*/ ]) // ------- cache deltas let prev = {dims:null,norm:null,auto:null} outer = outer.map((item,i,f)=> { let span = { dims : item.dims, lead : prev.dims ?? f.at(-1).dims, // loop around array for first entry hang : item.hang } let norm = { // z-score scaling & max dev normalisation dims : item.dims.map((d,j)=> (d-stats.med[j])/stats.dev[j]/stats.nom[j][0]), lead : prev.norm ?? // calculate norm lead when i == 0 span.lead.map((d,j)=>(d-delts.med[j])/delts.dev[j]/delts.nom[j][0]), hang : item.hang.map((d,j)=>(d-delts.med[j])/delts.dev[j]/delts.nom[j][0]), } let auto = { // range scaling dims : item.dims.map((d,j)=>(d - stats.min[j]) / stats.nom[j][1]), lead : prev.auto ?? // calculate auto lead when i == 0 span.lead.map((d,j)=>(d - delts.min[j]) / delts.nom[j][1]), hang : item.hang.map((d,j)=>(d - delts.min[j]) / delts.nom[j][1]), } prev.dims = span.hang; prev.norm = norm.hang; prev.auto = auto.hang; let meta = { caps : item.line.replace(/(\p{Lu})|\n|\S+/gu,'$1'), capt : // item.line.replace(/(\p{Lu}\p{L}+ [0-9\.]{0,3}[:\.]([0-9\.]{0,3})?)|\n|.*/gu,'$1'), item.line.substring(0,item.line.indexOf(" ", item.line.indexOf(" ")+1)).replace(/(.* [0-9\.]{0,3}[:\.])$|.*/,'$1'), gaps : item.line.length - item.line.replaceAll(' ','').length, //mult : item.line.length - item.line.replaceAll('\n','').length, numb : item.line.length < 4 && Number.isInteger(+item.line.replace(/\s+/g,'')), mark : among(item.line, { at: 0, either:`•●` }), seam : among(item.line, { at: 0, either:`-—,;)` }) || among(item.line, { at: 1, either:`-—;)` }), more : among(item.line, { at:-2, either:`-—,;()` }), punc : among(item.line, { at:-2, either:`.?:!,”`}) || among(item.line, { at:-3, either:`.?:!,”`}), case : /^\p{Lu}/u.test(item.line[0]) && /^\p{Ll}/u.test(item.line[1]) && /^\p{L}+/u.test(item.line.substring(item.line.indexOf(' ')+1)), enum : /^[0-9\.]+$/.test(item.line.substring(0, item.line.indexOf(' '))) && /^[A-Z]/.test(item.line.substring(item.line.indexOf(' ')+1)), turn : item.page > f[i-1]?.page, axis : item.dims[1] < f[i-1]?.dims.span[1], same : Math.round(item.dims[1]) == Math.round(f[i-1]?.dims.span[1]), // same line dupl : freqs.get(item.hash) > 4 } for(let feat in norm) { // pivot norms + auto scaled features (mutates array) item[feat] = {span:span[feat],norm:norm[feat],auto:auto[feat]} } return {...item,...meta} }) return Object.defineProperties({stats,delts,outer,freqs}, { ticks: { get:()=>performance.now()-t1 } }) function among(string,{ either = '', at = 0} = {}) { let char = string.at(at); return either.indexOf(char) > -1 ? char : false } function median(array) { const middle = ~~(array.length * 0.5); if ( array.length % 2 === 0) { return (array[middle - 1] + array[middle]) * 0.5 } return array[middle]; } function mode(arr,{ rank = 0, table = false} = {}) { let prev = 0; let outer = []; for (let i = 1; i < arr.length; ++i) { let item = [i-prev,arr[i-1]]; if(arr[i] !== arr[i-1]) { prev = i; outer.push(item) } } return table ? outer.sort((a,b)=>b[0]-a[0]) : outer.sort((a,b)=>b[0]-a[0])[rank]?.[1] } function variance(arr, avg) { let acc = 0; for (let i = 0; i < arr.length; ++i) { acc += ((arr[i] - avg) ** 2) } return acc / (arr.length - 1) // return arr.reduce((a, n)=> a + ((n - avg) ** 2),0) / ( arr.length - 1 ) } }32:[{"id":"9d2c83d1557225097ad838483e279c3b2ea48e67e24b45f536bbd2a94ed24c3321a8ba1906e8ce297a259c967d9d47a725f0abdaca3686ec3dc63969b3cf44ff","href":"https://static.observableusercontent.com/files/9d2c83d1557225097ad838483e279c3b2ea48e67e24b45f536bbd2a94ed24c3321a8ba1906e8ce297a259c967d9d47a725f0abdaca3686ec3dc63969b3cf44ff","name":"drive(23@2.01.30)lines(2).json","type":"application/json","lastModified":1675049219113.836,"size":103015},{"id":"7c7aca759eacf0ef1c555754bf3f225bd29915535375a9e5a423e8e1542371d5901b61c1e85b9a7dc0c3ab00b5755c38a262ed322ebae8c8347ae0a107a2a373","href":"https://static.observableusercontent.com/files/7c7aca759eacf0ef1c555754bf3f225bd29915535375a9e5a423e8e1542371d5901b61c1e85b9a7dc0c3ab00b5755c38a262ed322ebae8c8347ae0a107a2a373","name":"drive(23@3.01.30)lines(1).json","type":"application/json","lastModified":1675049219006.762,"size":72677},{"id":"25a65db919abeaaecbd8e5a041e3931ba26c03166c9d1c9109794481ea8e863e950eac05583563ca5fdba38760658411988f0d9c4161ded73187e0172dabfee0","href":"https://static.observableusercontent.com/files/25a65db919abeaaecbd8e5a041e3931ba26c03166c9d1c9109794481ea8e863e950eac05583563ca5fdba38760658411988f0d9c4161ded73187e0172dabfee0","name":"drive(23@2.01.30)lines(3).json","type":"application/json","lastModified":1675049219766.706,"size":1188581},{"id":"67ddfb95df5ea0ed6d93f4cea7eba7aaded017bc9b9960659ddfa07356202e291e765b07cabc3b7480d745a99205508255568d8e0453e4363256f3d14d9e18ed","href":"https://static.observableusercontent.com/files/67ddfb95df5ea0ed6d93f4cea7eba7aaded017bc9b9960659ddfa07356202e291e765b07cabc3b7480d745a99205508255568d8e0453e4363256f3d14d9e18ed","name":"drive(23@4.01.30)lines(1).json","type":"application/json","lastModified":1675574180729.442,"size":76199},{"id":"e41033a87e8f395d607f853a4f5b71d02a537a0727fa017b704c26bd2560d781ad87a8ff66d99d920ac974e4a043e074b8ad236d3968a0d0ebe5fe0f587a15d1","href":"https://static.observableusercontent.com/files/e41033a87e8f395d607f853a4f5b71d02a537a0727fa017b704c26bd2560d781ad87a8ff66d99d920ac974e4a043e074b8ad236d3968a0d0ebe5fe0f587a15d1","name":"drive(23@3.01.30)lines(2).json","type":"application/json","lastModified":1675574196753.081,"size":108703},{"id":"c09191f03349a67c2b8218684719b9ba3488f434216bf7448b68c8e95f151256ee2ecc54ed43ee0abca2e5b762ed86267691b712dbc06b81e21959e35ce27f1b","href":"https://static.observableusercontent.com/files/c09191f03349a67c2b8218684719b9ba3488f434216bf7448b68c8e95f151256ee2ecc54ed43ee0abca2e5b762ed86267691b712dbc06b81e21959e35ce27f1b","name":"drive(23@3.01.30)lines(3).json","type":"application/json","lastModified":1675574215132.379,"size":1571194}] 2d:{"title":"Line classifier","cells":[{"id":0,"value":"# Line classifier","pinned":false,"hidden":false,"mode":"md"},{"id":10,"value":"___","pinned":false,"hidden":false,"mode":"md"},{"id":5691,"value":"outputs = [\n\n /*\n blockStart: fuzzy.variable([0, 1], { \n vague:fuzzy.gaussian(0,0.2), \n maybe:fuzzy.gaussian(0.5,0.15), \n clear:fuzzy.gaussian(1,0.2) \n }),*/\n\n \n {blockBody: outVar('p')},\n {title: outVar('h')},\n\n /* blockEnd: fuzzy.variable([0, 1], { \n vague:fuzzy.gaussian(0,0.2), \n maybe:fuzzy.gaussian(0.5,0.2), \n clear:fuzzy.gaussian(1,0.2) \n }),*/\n\n]","pinned":false,"hidden":false,"mode":"ojs"},{"id":13658,"value":"Plot.plot({\n marks: [\n Plot.rectY(\n classed \n // temp.outer.map(d=>d.dims.norm[2])\n , Plot.binX({y: \"count\"}, {x: d=>d?.title ?? d?.blockBody ?? d})),\n Plot.ruleY([0])\n ]\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":8835,"value":"title = [\n\n fuzzy.and({charWidth:\"above\", lineExtent:\"close\"},{title:\"clear\"}),\n \n fuzzy.and({leadIndent:\"close\",hangIndent:\"close\", sameLine:\"vague\", lineEnd:\"vague\"},{title:\"clear\"}),\n\n \n // fuzzy.or({leadWidth:\"equal\",hangWidth:\"equal\"},{title:\"maybe\"}),\n\n fuzzy.and({leadIndent:\"equal\",leadExtent:\"equal\"},{title:\"vague\"}),\n\n\n \n \n /* fuzzy.or({lineExtent:\"equal\"},{title:\"maybe\"}),\n fuzzy.or({lineWidth:\"above\"},{title:\"clear\"}),\n fuzzy.or({charWidth:\"above\"},{title:\"clear\"}),\n fuzzy.and({leadIndent:\"equal\",hangIndent:\"equal\"},{title:\"vague\"}),\n \n fuzzy.and({leadExtent:\"apart\",hangExtent:\"apart\"},{title:\"maybe\"}),*/\n \n \n fuzzy.and({duplicate:\"clear\"},{title:\"vague\"})\n\n \n // fuzzy.or({leadSpacing:\"close\",hangSpacing:\"close\"},{title:\"clear\"})\n\n \n]","pinned":false,"hidden":false,"mode":"ojs"},{"id":5749,"value":"$2f","pinned":false,"hidden":false,"mode":"ojs"},{"id":5774,"value":"$30","pinned":false,"hidden":false,"mode":"ojs"},{"id":6134,"value":"viewof useLineWidth = Inputs.range([-1,1],{label:\"lineWidth\",step:0.01})","pinned":false,"hidden":false,"mode":"ojs"},{"id":6120,"value":"viewof useLeadIndent = Inputs.range([-1,1],{label:\"leadIndent\",step:0.01})","pinned":false,"hidden":false,"mode":"ojs"},{"id":6318,"value":"viewof useHangIndent = Inputs.range([-1,1],{label:\"hangIndent\",step:0.01})","pinned":false,"hidden":false,"mode":"ojs"},{"id":11334,"value":"balVar = (desc)=> \n({desc,...fuzzy.variable([-1,1], { \n below:fuzzy.invSigmoid(-0.05,10), \n equal:fuzzy.gaussian(0,0.125), \n above:fuzzy.sigmoid(0.05,10) \n })})","pinned":true,"hidden":false,"mode":"ojs"},{"id":6905,"value":"simVar = (prop)=> \n({prop,...fuzzy.variable([-1,1], { \n equal:fuzzy.invSigmoid(0.05,10), \n close:fuzzy.gaussian(0.25,0.20), \n apart:fuzzy.sigmoid(0.45,10) \n })})","pinned":true,"hidden":false,"mode":"ojs"},{"id":7522,"value":"outVar= (prop) =>\n({prop,...fuzzy.variable([0, 1], { \n vague:\n //fuzzy.invSigmoid(0.4,10),\n fuzzy.invRamp(0,1), \n maybe:\n // fuzzy.gaussian(0.5,0.175),\n fuzzy.triangle(0,0.5,1), \n clear:\n // fuzzy.sigmoid(0.75,5),\n // fuzzy.triangle(0.25,1) \n fuzzy.ramp(0,1)\n })})","pinned":true,"hidden":false,"mode":"ojs"},{"id":6908,"value":"dimVar = (prop)=>\n({prop,...fuzzy.variable([-1,1], {\n small:fuzzy.invSigmoid(-0.2,10), \n basic:fuzzy.gaussian(0,0.1),\n large:fuzzy.sigmoid(0.2,10),\n })})","pinned":true,"hidden":false,"mode":"ojs"},{"id":6934,"value":"binVar = (prop)=>\n({prop,...fuzzy.variable([0, 1], {\n vague: fuzzy.invRamp(0,1),\n clear: fuzzy.ramp(0,1)\n })})","pinned":true,"hidden":false,"mode":"ojs"},{"id":5542,"value":"inputs = ({\n \n lineWidth: dimVar('dims[2]'),\n charWidth: balVar('dims[5]'), // + divide by line count\n \n leadWidth: simVar('lead[2]'),\n hangWidth: simVar('hang[2]'),\n\n leadHeight: simVar('lead[3]'),\n hangHeight: simVar('hang[3]'),\n\n leadIndent: simVar('lead[0]'),\n hangIndent: simVar('hang[0]'),\n\n lineExtent: simVar('dims[4]'),\n leadExtent: simVar('lead[4]'),\n hangExtent: simVar('hang[4]'),\n\n duplicate: binVar('dupl'),\n\n titlecase: binVar('caps>gaps'),\n numerated: binVar('enum'),\n\n pageTurn: binVar('turn'),\n lineEnd: binVar('punc'),\n numberEnd: binVar('enum[i + 1] + punc'),\n\n sameLine: binVar('same'),\n\n})","pinned":true,"hidden":false,"mode":"ojs"},{"id":10881,"value":"[...temp.outer].map(d=>({feat:\n (d.dims.norm[3] + d.dims.norm[6] * 100 * (1-1/(d.caps.replaceAll(' ','').length+1)))/d.line.length/d.mult,\n \n line:d.line,hash:d.hash} )).sort((a,b)=>b.feat - a.feat)","pinned":false,"hidden":false,"mode":"ojs"},{"id":16119,"value":"temp.ticks","pinned":true,"hidden":false,"mode":"ojs"},{"id":6778,"value":"viewof file = Inputs.select([{fileA},{fileB},{fileC}],{label:'File',format: d=>Object.keys(d)[0]})","pinned":false,"hidden":false,"mode":"ojs"},{"id":15512,"value":"viewof feat = Inputs.select(Object.getOwnPropertyNames(buildRules()),{label:'Feature',format: d=>d,value:this?.value})","pinned":false,"hidden":false,"mode":"ojs"},{"id":80,"value":"Plot.plot({//y:{axis:null},\n x:{domain:inputs2[feat].domain,\n // zero:true\n },\n marks: [\n Plot.rectY( feature ,\n Plot.binX({y: \"count\",title: d=>d.map(l=>l.line.slice(0,48)).slice(0,8).join('\\n')\n },{x:d=>d.feat,inset:0})),\n Plot.ruleY([0])\n ]\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":5546,"value":"{ \n let node = svg`${vizzy.varToSvg(inputs2[feat])}`\n\n node.setAttribute('style',`position:relative;left:30px; transform:scale(0.97,1);`)\n\n return node\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":16676,"value":"feature.filter(d=>d.class=='below')","pinned":true,"hidden":false,"mode":"ojs"},{"id":14614,"value":"buildRules = (stats,delts) => {\n let wrap = {\n \n ... mixVar({lineWidth:col({stats,at:2,mix:0.25})}), \n ... mixVar({charWidth:col({stats,at:5})}), \n\n ... mixVar({lineHeight:col({stats,at:3})}),\n ... mixVar({lineExtent:col({stats,at:4,mix:0.5})}),\n\n ... mixVar({leadIndent:col({delts,at:0,mix:0.5})}),\n ... mixVar({hangIndent:col({delts,at:0,mix:0.5})})\n \n }\n return wrap\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":15073,"value":"function mixVar({ domain, ...feature } = {}) {\n \n let name = Object.keys(feature).at(0);\n feature = Object.values(feature).at(0);\n\n Object.assign(feature, {\n ref: [\"lead\", \"hang\"].includes(name.slice(0, 4))\n ? name.slice(0, 4)\n : \"dims\" });\n\n let { min, low, mid, top, max, dev } = feature;\n return {\n [name]: {\n ...fuzzy.variable(domain ?? [min, max], {\n lower: fuzzy.invSigmoid((min+low*2)/3, 1 / dev),\n below: (i)=>fuzzy.invSigmoid(low, 1 / dev)(i) - fuzzy.invSigmoid((min+low*2)/3, 1 / dev)(i),\n equal: fuzzy.gaussian(mid,dev),\n above: (i)=>fuzzy.sigmoid(top, 1 / dev)(i) - fuzzy.sigmoid((max+top*2)/3, 1 / dev)(i),\n upper: fuzzy.sigmoid((max+top*2)/3, 1 / dev)\n }),\n feature\n },\n };\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":11344,"value":"feature = [...temp.outer]\n .map(d=> ({\n feat: d[inputs2[feat].feature.ref].span[inputs2[feat].feature.col].toFixed(2),\n class:'',\n line: d.line\n }))\n .map(d=>({...d,class:fuzzy.classify(inputs2[feat], d.feat)} ))\n//.filter(d=>d.class == 'upper')\n\n\n /*\n[...temp.outer].map((d,i,f)=>({\n feat: \n d.dims.span[3],\n line: d.line})).map(d=>({...d,class:fuzzy.classify(inputs2.lineHeight, d.feat)} ))\n \n .sort((a,b)=>b.feat -a.feat)//.filter(d=>d.feat>0.1)//.slice(220)*/","pinned":false,"hidden":false,"mode":"ojs"},{"id":15798,"value":"inputs2 = buildRules(temp.stats,temp.delts)","pinned":false,"hidden":false,"mode":"ojs"},{"id":15205,"value":"function col({ stats, delts, at = 0, IQR = 0.5, mix = 0.5, aim = true, ...rest } = {}) {\n\n let src = stats ? 'stats' : delts ? 'delts' : null //new Error().stack.split(/at (.*) /)[1]\n let self = Object.entries(stats ?? delts ?? {}).reduce((pool,[k,v]) => (pool[k] = v[at],pool),{})\n Object.assign(self,{\n low : self.all?.percentile((50 - (IQR * 100) / 2) / 100),\n mid : self.med,\n dev : self.dev * mix,\n top : self.all?.percentile((50 + (IQR * 100) / 2) / 100),\n col : at, ref: '', src\n })\n \n if (aim) self.mid = (self.low+self.top) / 2 //=> IQM, midhinge\n\n rest = Object.entries(rest).reduce((pool,[k,v])=> (pool[k] = typeof v == 'function' ? v(self) : v, pool) ,{})\n \n return {...self,...rest};\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":3724,"value":"j = 3","pinned":false,"hidden":false,"mode":"ojs"},{"id":4625,"value":"Plot.plot({\n marks: [\n Plot.ruleY([0]),\n Plot.lineY([...new Set(temp.outer.map(d=>d.dims.auto[j]))].sort(), {x: d=>d, y: (d,i,f)=>i/(f.length-1), stroke:\"#bbb\"}),\n Plot.lineY(//[...new Set(temp.delts.all[2])], {y:d=>temp.delts.all[2].fullRank(d), x: (d,i)=>d}\n Array.from({length:4}, (_,i)=>(temp.stats.all[j].occurance(i/3)-temp.stats.all[j].occurance(0))/\n (temp.stats.all[j].occurance(1)-temp.stats.all[j].percentile(0))), \n { y: (d,i,f)=>i/(f.length-1),x:d=>d, curve:\"basis\"})\n ]\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":39,"value":"$31","pinned":false,"hidden":false,"mode":"ojs"},{"id":35,"value":"___\n\n### Data","pinned":false,"hidden":false,"mode":"md"},{"id":6786,"value":"open = ()=> { for(let prop in file) {\n return file[prop] }\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":24,"value":"fileA = FileAttachment(\"drive(23@4.01.30)lines(1).json\").json().then(a=>a.map(d=> ({...d,line:decodeURI(d.string)})))","pinned":false,"hidden":false,"mode":"ojs"},{"id":25,"value":"fileB = FileAttachment(\"drive(23@3.01.30)lines(2).json\").json().then(a=>a.map(d=> ({...d,line:decodeURI(d.string)})))","pinned":false,"hidden":false,"mode":"ojs"},{"id":26,"value":"fileC = FileAttachment(\"drive(23@3.01.30)lines(3).json\").json().then(a=>a.map(d=> ({...d,line:decodeURI(d.string)})))","pinned":false,"hidden":false,"mode":"ojs"},{"id":5538,"value":"___\n\n### Static","pinned":false,"hidden":false,"mode":"md"},{"id":499,"value":"function sortedIndex(array, value) {\n\tvar low = 0,\n\t\thigh = array.length;\n\n\twhile (low < high) {\n\t\tvar mid = low + high >>> 1;\n\t\tif (array[mid] >= value) low = mid + 1;\n\t\telse high = mid;\n\t}\n\treturn low;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":12737,"value":"function clamp(num, min, max) {\n return num <= min \n ? min\n : num >= max \n ? max\n : num\n}","pinned":true,"hidden":false,"mode":"ojs"},{"id":10602,"value":"function roll(arr, {mode = 'entries', sort = true } = {}) {\n\n const counts = new Map();\n var count = 0;\n \n for (const item of arr) {\n count = (counts.get(item) ?? 0) + 1\n counts.set(item,count)\n }\n return sort ? [...counts[mode]()].sort((a,b)=> b[1]-a[1])\n : [...counts[mode]()]\n}","pinned":true,"hidden":false,"mode":"ojs"},{"id":11980,"value":"columns = { let t1 = performance.now();\n let result = //roll(\n roll(roll(\n temp.outer.filter(d=>d.axis).map(d=>d.page),{mode:'values'},\n {mode:'values',sort:false}))[0]\n // chain(temp.outer.filter(d=>d.axis).map(d=>d.page),\n // Array(2).fill(count({mode:'values',sort:false})))\n return {t:performance.now()-t1,result}\n}","pinned":true,"hidden":false,"mode":"ojs"},{"id":12077,"value":"function count({mode = 'entries', sort = true } = {}) {\n\n return (arr) => {\n\n const counts = new Map();\n var count = 0;\n \n for (const item of arr) {\n count = (counts.get(item) ?? 0) + 1\n counts.set(item,count)\n }\n return sort ? [...counts[mode]()].sort((a,b)=> b[1]-a[1] || a-b )\n : [...counts[mode]()]\n}\n\n}","pinned":true,"hidden":false,"mode":"ojs"},{"id":12075,"value":"chain = (array,...steps) => steps.flat().reduce((pipe,fn) => fn(pipe),array)","pinned":true,"hidden":false,"mode":"ojs"},{"id":2284,"value":"function derive(f) {\n var h = 0.001;\n return function (x) { return (f(x+h) - f(x-h)) / (2 * h)}\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":4746,"value":"function smooth(array, bw = 4) {\n let max = 0\n return array.map((d,i,f)=> { \n let sum = [...f].slice(Math.max(i-bw,0),i+1).reduce((all,cur)=> all+cur)/bw\n if(sum > max) max = sum\n return sum\n }).map(d=>d/max)\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":3945,"value":"___\n\n### Imports","pinned":false,"hidden":false,"mode":"md"},{"id":7454,"value":"import {Plot} from \"@mkfreeman/plot-tooltip\"","pinned":true,"hidden":false,"mode":"ojs"},{"id":3937,"value":"fuzzy = import('https://cdn.skypack.dev/@thi.ng/fuzzy@2.1.23?min')","pinned":true,"hidden":false,"mode":"ojs"},{"id":3939,"value":"vizzy = import('https://cdn.skypack.dev/@thi.ng/fuzzy-viz@2.1.45?min')","pinned":true,"hidden":false,"mode":"ojs"}],"files":"$W32","stdlib":"1","titling":"auto"}