2c:null 2d:["$","title",null,{"children":"Warping LCSS | Lawrence Thatcher | Observable"}] 32:T49f,md`## References - [_Supporting One-Time Point Annotations for Gesture Recognition_](https://ieeexplore.ieee.org/abstract/document/7778186/) by Long-Van Nguyen-Dinh, Alberto Calatroni, and Gerhard Tröster - [_Improving Online Gesture Recognition with Template Matching Methods inAccelerometer Data_](https://www.researchgate.net/publication/258031645_Improving_Online_Gesture_Recognition_with_Template_Matching_Methods_in_Accelerometer_Data) by Long-Van Nguyen-Dinh, Daniel Roggen, Alberto Calatroni, and Gerhard Tröster - [_An Inertial Measurement Framework for Gesture Recognition and Applications_](https://link.springer.com/chapter/10.1007/3-540-47873-6_2) by Benbasat and Paradiso - [_The use of fast Fourier transform for the estimation of power spectra: a method based on time averaging over short, modified periodograms_](https://scholar.google.com/scholar?hl=en&as_sdt=0%2C45&q=The+use+of+the+fast+Fourier+transform+for+the+estimation+of+power+spectra%3A+A+method+based+on+time+averaging+over+short%2C+modified+periodograms&btnG=) by Peter Welch - [_Fast time series classification using numerosity reduction_](https://dl.acm.org/citation.cfm?id=1143974) by Xi _et al._ `33:T6eb,class Environment extends Annotations { constructor(svg, points=undefined, name="unknown", include_margin=true) { super(points); this.svg = svg; this.name = name; this.margin = include_margin ? [margin.left, margin.right, margin.top, margin.bottom] : [0,0,0,0] let [X,Y] = this.xy(); this.x = X; this.y = Y; this.zoom = d3.zoom().scaleExtent([1,30]).extent(ext).translateExtent(ext); this.line = d3.line() .x((d,i) => X(i)) .y((d,i) => Y(d)) console.debug("setup environment:", name, this.width, this.height, this); } get width() { return this.svg.attr("width") } get height() { return this.svg.attr("height") } xy() { let [ml,mr,mt,mb] = this.margin; let x = d3.scaleLinear().domain([0, data.length-1]).range([ml, this.width - mr]) let y = d3.scaleLinear().domain(extents).range([this.height - mb, mt]) return [x,y] } drawAnnotations() { drawAnnotations(this.svg, this.points) } drawSignals() { drawSignals(this.svg, line) } drawAxes() { drawAxes(this.svg) } setupClipPath() { setupClipPath(this.svg) } zoomed() { let t = d3.event.transform; x.domain(t.rescaleX(x0).domain()); // rescale x this.svg.selectAll(".line").attr("d", line); // redraw lines let ax = this.svg.select("g.x-axis"); // redraw x-axis ax.selectAll("*").remove(); ax.call(xAxis); this.drawAnnotations(); } setupZoom() { this.svg.append("rect") .attr("class", "zoom") .attr("width", w) .attr("height", h + margin.bottom) .attr("transform", "translate(" + margin.left + "," + margin.top + ")") this.svg.call(this.zoom.on("zoom", () => this.zoomed() )); } }34:T561,class BoundaryInitialization extends Environment { constructor(svg, points) { super(svg,points,"boundary-init"); } get singlepoints() { return this.points.filter((d) => d.point) } get boundpoints() { return this.points.filter((d) => !d.point) } clicked(d,i) { if (d.point) this.expand(d,i); else this.contract(d,i); } expand(d,i, delay=0) { let bounds = this.bounds(d,i)[strategy] this.svg.selectAll("rect.boundary").filter((r) => r.t === d.t) .transition().duration(300).delay(delay) .attr("x", x(bounds[0])) .attr("width", W(bounds)) .on("end", () => { this.points[i].point = false; this.points[i].bounds = bounds; console.log('expand', this.points); viewof markedBounds.value = this.points; }); } contract(d,i,delay=0) { this.svg.selectAll("rect.boundary").filter((r) => r.t === d.t) .transition().duration(400).delay(delay) .attr("x", (d) => {return x(d.t) -.5}) .attr("width", 1) .on("end", () => { this.points[i].point = true; console.log('contract', this.points); viewof markedBounds.value = this.points; }); } drawAnnotations() { drawAnnotations(this.svg, this.points, "annotation annotation-ghost"); drawBoundaries(this.svg, this.points).on("click", (d,i) => this.clicked(d,i)); } }35:T6e5,class AutoSizeEnvironment extends Annotations { constructor(svg, regions, name="unknown", xscale=xe, include_margin=false) { super([]); this.svg = svg; this.name = name; this.xscale = xscale.copy(); this.x0 = xscale.copy(); this.margin = include_margin ? [margin.left, margin.right, margin.top, margin.bottom] : [0,0,0,0] let [X,Y] = this.xy(); this.x = X; this.y = Y; this.regions = regions.map(r => {return {start: X.invert(xscale(r.start)), end: X.invert(xscale(r.end))} }) this.zoom = d3.zoom().scaleExtent([1,30]).extent(ext).translateExtent(ext); this.line = d3.line() .x((d,i) => X(i)) .y((d,i) => Y(d)) console.debug("setup environment:", name, this.width, this.height, this); } get width() { return this.svg.attr("width") } get height() { return this.svg.attr("height") } xy() { let [ml,mr,mt,mb] = this.margin; let x = d3.scaleLinear().domain([0, data.length-1]).range([ml, this.width - mr]) let y = d3.scaleLinear().domain(extents).range([this.height - mb, mt]) return [x,y] } drawAnnotations() { drawRegions(this.svg, this.regions, this.x) } drawSignals() { drawSignals(this.svg, this.line, "line noclip") } drawAxes() { drawAxes(this.svg) } setupClipPath() { } zoomed() { let t = d3.event.transform; this.x.domain(t.rescaleX(x0).domain()); // rescale x this.svg.selectAll(".line").attr("d", this.line); // redraw lines this.drawAnnotations(); } setupZoom() { this.svg.append("rect") .attr("class", "zoom") .attr("width", this.width) .attr("height", this.height) this.svg.call(this.zoom.on("zoom", () => this.zoomed() )); } }36:T687,class NonMotionRemoval extends Environment { constructor(svg, points) { super(svg,points,"nonmotion-removal"); } get trimmed() { return this.points.filter(d => d.trimmed) } get untrimmed() { return this.points.filter(d => !d.trimmed) } toggle(d) { if (!d.trimmed) this.trim(d); else this.untrim(d); } trim(d) { let i = this.points.findIndex(p => p.t === d.t); this.svg.selectAll("rect.boundary.noclick").filter((r) => r.t === d.t) .transition().duration(500) .attr("x", (d) => x(d.L)) .attr("width", (d) => x(d.R) - x(d.L)) .on("end", () => { this.points[i].oldbounds = this.points[i].bounds; this.points[i].bounds = [d.L, d.R]; this.points[i].trimmed = true; this.drawAnnotations(); }) } untrim(d) { let i = this.points.findIndex(p => p.t === d.t); let origBounds = d.oldbounds || d.bounds; this.svg.selectAll("rect.boundary.noclick").filter((r) => r.t === d.t) .transition().duration(500) .attr("x", (d) => x(origBounds[0])) .attr("width", (d) => x(origBounds[1]) - x(origBounds[0])) .on("end", () => { this.points[i].bounds = origBounds; delete this.points[i].oldbounds; this.points[i].trimmed = false; this.drawAnnotations(); }) } toTemplates() { viewof markedBounds2.value = this.points; } drawAnnotations() { drawAnnotations(this.svg, this.points, "annotation annotation-ghost"); drawBoundaries(this.svg, this.points, "boundary noclick"); drawTrims(this.svg, this.points).on("click", (d) => this.toggle(d)) } }37:T551,css = html``38:[] 30:{"id":"d2855d78e0939872","slug":"warping-lcss","title":"Warping LCSS","thumbnail":"e696bcfde435b0030c8d70ac1ca4544a80441f1c91c02c6c84c66a2ef886921b","access_level":"public","create_time":"$D2018-08-29T18:31:20.307Z","update_time":"$D2019-01-26T02:02:36.976Z","view_time":null,"first_public_time":"$D2018-08-29T20:07:01.034Z","has_unpublished_edits":null,"fork":false,"stars":6,"starred":false,"owner":{"type":"team","id":"fe780737d88b9b25","login":"lwthatcher","name":"Lawrence Thatcher","avatar_url":"https://avatars-next.observableusercontent.com/0a65531a6560640a422e9d867df27dfaf6b6082c1ff4f92a53372ecbeb785ba0","tier":"starter_2024"},"version":2454,"publish_version":2454,"role":null,"authored":false,"authors":[{"type":"individual","id":"8450d56c8907b6b4","login":"lwthatcher","name":"Lawrence Thatcher","avatar_url":"https://avatars-next.observableusercontent.com/0a65531a6560640a422e9d867df27dfaf6b6082c1ff4f92a53372ecbeb785ba0","tier":null}],"members":[{"type":"individual","id":"8450d56c8907b6b4","login":"lwthatcher","name":"Lawrence Thatcher","avatar_url":"https://avatars-next.observableusercontent.com/0a65531a6560640a422e9d867df27dfaf6b6082c1ff4f92a53372ecbeb785ba0","tier":null,"role":"editor"}],"comments":[],"custom_thumbnail":false,"trashed":false,"lapsed":null,"subscribed":null,"forks":1,"collection":null,"type":"notebook","body":{"titling":"auto","title":"Warping LCSS","cells":[{"id":0,"value":"md`# Warping LCSS\nThis is meant to be an interactive notebook illustrating Nguyen-Dinh _et al.'s_ [**BoundarySearch**](https://ieeexplore.ieee.org/document/7778186/#) algorithm for temporal gesture recognition using single time point annotations.\nThis method employs a variant of Dynamic Time Warping (DTW) called [**Warping LCSS**](https://www.researchgate.net/publication/258031645_Improving_Online_Gesture_Recognition_with_Template_Matching_Methods_in_Accelerometer_Data).\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":55,"value":"md`## Step 1: Quanitization\nThe first step is to reduce the dimensionality of the input data and quanitize into discrete groups.\nIn other words, we need to _cluster_ the data points.\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":10,"value":"md`### The Dataset\nIn this example we'll use the simple [pills-blue](https://gist.github.com/lwthatcher/da8f095431b5d00bf5257444677b5245) dataset.\nFor simplicity, we only use the accelerometer and gyroscope sensor readings,\nboth of which have three axes:\n\n| Accelerometer | | Gyroscope | |\n|----|--------------------|----|--------------------|\n| Ax | ${sparkline(Ax,0)} | Gx | ${sparkline(Gx,3)} |\n| Ay | ${sparkline(Ay,1)} | Gy | ${sparkline(Gy,4)} |\n| Az | ${sparkline(Az,2)} | Gz | ${sparkline(Gz,5)} |\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":6,"value":"data = {\n let url = \"https://gist.githubusercontent.com/lwthatcher/da8f095431b5d00bf5257444677b5245/raw/b4ccc9b60f715623a7862482217950b5ea5aa555/data.short.csv\"\n let parseRow = (r) => {\n return { data: [+r.Ax, +r.Ay, +r.Az, +r.Gx, +r.Gy, +r.Gz],\n label: +r.label}\n }\n return d3.csv(url, parseRow)\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":58,"value":"md`### Clustering\nThere are a variety of ways to cluster the data.\nThe _BoundarySearch_ implementation uses \\`k-means\\` clustering, so we'll use that for this example as well.\n\n#### the \\`k-means\\` algorithm\nThere are multiple other references that introduce and explain the \\`k-means\\` algorithm (included below),\nso for now we will content ourselves to precompute these clusters _offline_.\n\n> **Other Notebooks on \\`k-means\\`:**\n> - [k-means clustering](https://beta.observablehq.com/@akngs/k-means-clustering) by Alan Kang\n> - [kmeans workshop](https://beta.observablehq.com/@vndrewlee/kmeans-workshop) by Andrew Lee\n\n\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":90,"value":"clusters = d3.csv(\"https://gist.githubusercontent.com/lwthatcher/0facab62c8d14bacb8eea9c8b97d4cbc/raw/d39a2a9d848353ac6ecb3ef5ac32de81852af9f1/kmeans.short.csv\")","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":84,"value":"md`#### choice of \\`k\\`\nOne of the challenges to clustering algorithms such as \\`k-means\\` \nis selecting an appropriate value for the user defined parameter \\`k\\` (the number of clusters). \nThe _BoundarySearch_ paper says the following on selecting a value for \\`k\\`:\n\n> _The number of clusters k (i.e., number of symbols) should be chosen so that the vocabulary can capture variation intrinsic to gestures. The choice of k is done through cross-validation or empirically._\n\nIn other words: try a bunch of different values for \\`k\\` until you get one that works.\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":76,"value":"viewof k = html``","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":79,"value":"md`**k = ${k}**`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":206,"value":"md`\n${genTable()}\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":235,"value":"md`\n> **Note:** a _streak_ is a continuous sequence of time steps with the same cluster assignment.\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":92,"value":"kmeans = clusters.map(row => +row[k])","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":310,"value":"md`\n---\n## Step 2: Annotation\nNext we need to provide the actual user annotations.\nHere we're going to part from the _BoundarySearch_ paper which samples points from each of the known labels.\nInstead for the sake of interactivity, we will provide the labels ourselves by clicking on the graph below:\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1024,"value":"md`### Interactive Example\n- Create **_single-point annotations_** for the data by \\`clicking\\` on the chart below.\n\n> - Use the \\`scroll-wheel\\` to _zoom in and out_.\n\n> - You can _remove_ annotations by \\`clicking\\` the annotation.\n\n- For simplicity, the correct label for your annotation is automatically provided.\n\n> - The highlight bar's **color** indicates the label.\n\n> - The number above the highlight bar displays the cluster assignment from the quantization step for that point.\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":312,"value":"viewof annotation = {\n const svg = d3.select(DOM.svg(width, height));\n let env = new Environment(svg, await marked, \"annotation\");\n env.setupClipPath();\n env.drawSignals();\n env.drawAxes();\n env.setupZoom();\n \n let [highlight, clear] = setupHighlight(svg);\n \n svg.on(\"mousemove touchmove\", moved);\n svg.on(\"click\", clicked);\n \n function moved() {\n const t = getT(this);\n highlight(t);\n }\n function clicked() {\n let [t, target] = [getT(this), d3.event.target];\n if (target.classList.contains(\"annotation\")) {\n removeAnnotation(target);\n }\n else { addAnnotation(t); }\n }\n\n env.drawAnnotations();\n \n svg.property(\"value\", env).dispatch(\"input\")\n return svg.node();\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":613,"value":"md`### Labels\n| | \"null\" | \"pills\" | \"move\" |\n|:--------------|--------|------------|-----------|\n| **yours:** | ${n.Ø} | ${n.pills} | ${n.move} |\n| **expected:** | 9 | 5 | 3 |\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":517,"value":"{\n const button = html``;\n button.onclick = () => {\n viewof marked.value = [];\n viewof markedBounds.value = [];\n }\n return button;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":61,"value":"md`\n---\n## Step 3: Boundary Initialization\nOnce the single time-point annotations are available, \ninitial start/end bounds can be provided to convert them into temporal labels.\n_**Warping LCSS**_ uses these initial bounds to find the optimal matching sub-sequence inside them to infer the label boundaries.\nConsequently this initial estimate must both be _*_[admissable](https://en.wikipedia.org/wiki/Admissible_heuristic),\nbut also as tight as possible to reduce computational complexity.\n\n> ***Note:** The initial boundaries _must_ encompass the actual boundaries to provide a search space for the _Warping LCSS_ algorithm. Its dynamic programming aspect cannot arbitrarily \"grow\" its search space while still remaining computationally efficient.\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":706,"value":"md`### Initialization Strategies\nThe _BoundarySearch_ paper presents two methods for initializing these boundaries, but others could also be used.\nFor simplicity we'll look at the two presented in the paper: **MaxLen** and **NoLen**.\n\n#### MaxLen\nThe _**maximum length**_ strategy expands the initial boundaries in either direction from the annotation point by the maximum length of any label within that class, or the location of any other single-point annotations in that direction, whichever is shorter.\n\nThis method can drastically reduce the search space by providing tighter bounds, but also _can only be used with prior knowledge of label lengths_ already known.\n\n#### NoLen\nWhen prior knowledge about label lengths is not known, the _**no length**_ strategy expands the initial boundaries to the nearest single-point annotations in either direction.`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":715,"value":"md`### Select Stratgey\nFortunately for us, we _do_ already known the maxium lengths of each label type.\nThese will be used when selecting the **_MaxLen_** option.\n\n| | \"null\" | \"pills\" | \"move\" |\n|:--------------------|------------------|------------------|------------------|\n| **maximum length:** | ${max_bounds[0]} | ${max_bounds[1]} | ${max_bounds[2]} |\n\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":720,"value":"md`### Interactive Example:\n- To _**initialize the boundaries**_, simply \\`click\\` on the single-point annotation.\n\n- To avoid overlap, the boundaries are now stacked vertically by label.\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1609,"value":"viewof stacking = html``","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":702,"value":"viewof strategy = html``","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":704,"value":"viewof boundary_init = {\n const svg = d3.select(DOM.svg(width, height));\n let env = new BoundaryInitialization(svg, await markedBounds)\n env.setupClipPath();\n env.drawSignals();\n env.drawAxes();\n env.setupZoom();\n \n let [highlight, clear] = setupHighlight(svg);\n\n env.drawAnnotations();\n \n svg.on(\"mousemove touchmove\", moved);\n \n function moved() {\n if (d3.event.shiftKey) {\n const t = getT(this);\n highlight(t); \n }\n else { clear(); }\n }\n \n svg.property(\"value\", env).dispatch(\"input\")\n return svg.node();\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":916,"value":"{\n const btn1 = html``;\n const btn2 = html``;\n btn1.onclick = () => {\n let env = boundary_init;\n for (let i = 0; i < env.points.length; i++) { env.contract(env.points[i],i) }\n }\n btn2.onclick = () => {\n let env = boundary_init;\n for (let i = 0; i < env.points.length; i++) { env.expand(env.points[i],i) }\n }\n return html`${btn1} ${btn2}`;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":64,"value":"md`\n---\n## Step 4: Non-motion Removal\nThis step is **optional**, but can help in reducing the size and search spaces of your templates.\nThe basic idea is to try to trim areas of non-motion.\n\nThe are various methods for automatically detecting periods of motion/non-motion,\nhere we will discuss two such approaches: _variance thresholding_ and _energy thresholding_.\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1169,"value":"md`### Variance Thresholding\nThis is the technique used in the _BoundarySearch_ paper, which takes the variance over a sliding window for each axis in the data. The variance over a 16-sample window has been precomputed and is imported below.\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1160,"value":"variances = d3.csv(\"https://gist.githubusercontent.com/lwthatcher/1be1e2aca15ff1336819de2f67c8ecf5/raw/ca9e303e449e73cc89bb4909e9123944aad99b71/var.short.csv\", (row, i) => {\n return [+row.Ax, +row.Ay, +row.Az, +row.Gx, +row.Gy, +row.Gz]\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1479,"value":"md`\nThe periods of motion are inferred through a _**start motion threshold**_ and a _**stop motion threshold**_.\nWhen the variance of _any_ axis exceed the start threshold the motion part begins, and ends when _all_ axes are below the stop threshold.\n\nThese values can be adjusted below using the respective sliders:\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1280,"value":"viewof THRESHOLDS = {\nlet s1 = html``\nlet s2 = html``\n\ns1.oninput = (e) => { M.property(\"value\", [+s1.value, +s2.value]).dispatch(\"input\") }\ns2.oninput = (e) => { M.property(\"value\", [+s1.value, +s2.value]).dispatch(\"input\") }\n\nlet m = md`\n**Thresholds:**\n\n| Start Motion Threshold | Stop Motion Threshold |\n|:----------------------:|:---------------------:|\n| ${s1} | ${s2} |\n`\nlet M = d3.select(m)\nM.property(\"value\", [100,50]).dispatch(\"input\");\nreturn m\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1361,"value":"md`\n| | |\n|:--------------------------------------------:|:-------------------------------------------:|\n| start motion threshold: **${THRESHOLDS[0]}** | stop motion threshold: **${THRESHOLDS[1]}** |`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1447,"value":"nonmotion_var = {\n const svg = d3.select(DOM.svg(width, height/2));\n let regs = v_thresholds.filter(r => r.label === \"non-motion\");\n let env = new AutoSizeEnvironment(svg, regs, \"variance-thresholding\", xv);\n env.drawSignals();\n env.setupZoom();\n env.drawAnnotations();\n return svg.node();\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1177,"value":"md`### Energy Thresholding\nAn alternate technique is to use a threshold on the spectral energy of the signal.\nThe spectral signal energy has already been approximated using _*Welch's Method_ in python using the method [\\`scipy.signal.welch\\`](https://docs.scipy.org/doc/scipy-0.14.0/reference/generated/scipy.signal.welch.html) from the \\`scipy\\` library,\nand is imported below.\n\n> ***Note:** refer to the \\`scipy\\` documentation for more information on the implementation.\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":875,"value":"energy = d3.csv(\"https://gist.githubusercontent.com/lwthatcher/36933f118fe5ecba2b44b8c5f8ac26e2/raw/de8b5d93b7206e2a92f4bca39b16eb766c71d112/energy.csv\",(row) => {\n return [+row.Ax, +row.Ay, +row.Az, +row.Gx, +row.Gy, +row.Gz]\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1495,"value":"md`\nThis time we will use a single threshold value over the magnitiude of the energy over time,\nhowever we also have a _smoothing_ parameter, which helps filter out \"changes\" in motion beneath a certain length.\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2262,"value":"viewof ENERGY_THRESHOLDS = {\nlet s1 = html``\nlet s2 = html``\n\ns1.oninput = (e) => { M.property(\"value\", [+s1.value, +s2.value]).dispatch(\"input\") }\ns2.oninput = (e) => { M.property(\"value\", [+s1.value, +s2.value]).dispatch(\"input\") }\n\nlet m = md`\n**Thresholds:**\n\n| Threshold | Smoothing |\n|:----------------------:|:---------------------:|\n| ${s1} | ${s2} |\n`\nlet M = d3.select(m)\nM.property(\"value\", [1,0]).dispatch(\"input\");\nreturn m\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2404,"value":"md`\n| | |\n|:--------------------------------------------------------:|:--------------------------------------:|\n| energy threshold: **${ENERGY_THRESHOLDS[0].toFixed(3)}** | smoothing: **${ENERGY_THRESHOLDS[1]}** |`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1522,"value":"nonmotion_energy = {\n const svg = d3.select(DOM.svg(width, height/2));\n let regs = smoothed_regions.filter(r => r.label === \"non-motion\");\n let env = new AutoSizeEnvironment(svg, regs, \"energy-thresholding\", xe);\n env.drawSignals();\n env.setupZoom();\n env.drawAnnotations();\n return svg.node();\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1607,"value":"md`### Trimming Templates\nWe can now use the inferred areas of motion/non-motion to narrow our template boundaries.\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1632,"value":"viewof trim_method = html``","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1894,"value":"md`##### Interactive Example\n- Click on the highlighted portion of a label to **trim** it to the specified length.\n\n> - Undo the trim by clicking on it again, or pressing the button.\n\n- Select _energy thresholding_ or _variance thresholding_ using the \\`trim_method\\` dropdown above.\n\n- When finished, press the button below to move on to the next step.\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1138,"value":"viewof nonmotion_removal = {\n const svg = d3.select(DOM.svg(width, height));\n let points = clip_bounds;\n \n let env = new NonMotionRemoval(svg, points);\n env.setupClipPath();\n env.drawSignals();\n env.drawAxes();\n env.setupZoom();\n env.drawAnnotations();\n \n svg.on(\"mousemove touchmove\", moved);\n function moved() {\n let t = getT(this);\n \n }\n \n svg.property(\"value\", env).dispatch(\"input\")\n return svg.node();\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1819,"value":"{\n const btn1 = html``;\n const btn2 = html``;\n btn1.onclick = () => {\n let env = nonmotion_removal;\n for (let p of env.points) { env.untrim(p) }\n viewof markedBounds2.value = [];\n }\n btn2.onclick = () => {\n let env = nonmotion_removal;\n viewof markedBounds2.value = nonmotion_removal.points;\n }\n return html`${btn1} ${btn2}`;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":66,"value":"md`\n---\n## Step 5: Template Matching\nThis is the step that finally uses _WarpingLCSS_ to refine the template boundaries.\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1841,"value":"{\nif (markedBounds2.length > 0) {\n return md`#### Templates\n | | \"null\" | \"pills\" | \"move\" |\n |:----------------------|------------------|------------------|------------------|\n | **count:** | ${cts[0].length} | ${cts[1].length} | ${cts[2].length} |\n | **template lengths:** | ${tlens(cts[0])} | ${tlens(cts[1])} | ${tlens(cts[2])} |\n `\n}\nelse { return md`- Use the above to initialize templates`}\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2072,"value":"md`### Comparison Matrices\nDuring this step we use dynamic programming to compare each template to all other templates in the same class.\n\nWe'll start by initializing each of the appropriate \\`M x N\\` matricies to all zeros.\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1900,"value":"MATRICIES = {\n let Ls = templates.map(group => group.map((t) => t.length)).map(pairs)\n return Ls.map(group => group.map((p) => math.zeros(...p)))\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2088,"value":"md`### Warping Parameters\nThere are a few custom parameters that are used here:\n\n| parameter | definition | current value |\n|-----------|------------------------|---------------|\n| **\\`p\\`** | penalty term | ${p} |\n| **\\`r\\`** | maximum warping window | ${r} |\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2101,"value":"viewof p = html``","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2108,"value":"viewof r = html``","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2077,"value":"md`### Computing Distances\nThe equations specifying the warping distance relies on the original distance between the discretized clusters. For convenience these have also been precomputed and imported to correspond to our choice of k.\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2305,"value":"dists = cluster_distances.get(k)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2307,"value":"md`\n> **Note:** This array does _not_ contain the entire distance matrix, but instead is only the collapsed distance matrix. For more details please see the appendix below.\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2117,"value":"ΨA = (αi, αj) => {\n if (αi === αj) return 1;\n else return -p * dist(αi, αj);\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2419,"value":"ΨW = (αi, αj) => {return -p * dist(αi, αj)}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2422,"value":"md`### Computing Warping Paths\n> -DO STUFF-\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2424,"value":"B = (matrix) => {\n let [M,N] = matrix.size()\n return [M,N]\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2427,"value":"B(MATRICIES[1][0])","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":69,"value":"md`\n---\n## Step 6: Boundary Refinement\n> -INSERT DOCUMENTATION-\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":656,"value":"md`\n---\n## Conclusion\n> -INSERT CONCLUSION-\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":659,"value":"md`\n---\n\n# Appendix\nThe appendix includes the references for the various citations used throughout, \nas well as additional code cells used to help facilitate the graphical representations.\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2356,"value":"md`## Collapsed Distance Matrices\nThe \\`k-means\\` clusters for this example were precomputed along with their corresponding distance matrices.\nSince the distances between two clusters are symmetric, to save space when importing only the **upper triangle** of each distance matrix was stored.\n\n\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2381,"value":"md`### Indexing into the Collapsed Array\nNormally indexing into the full 2-dimensional distance matrix is of the form \\`dist[i,j]\\`, however the collapsed array is only 1-dimensional. The \\`i, j\\` indices can be converted into a single index into the collapsed matrix with the following equation:\n\n${index_equation} where ${tex`i \\leq j`}\n\n\n> **Note:** If ${tex`i \\geq j`}, simply swap \\`i\\` and \\`j\\`.\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2389,"value":"dist = (i,j) => {\n if (i > j) [i,j] = [j,i];\n let idx = (k*i) - ((i-1)*i)/2 + (j-i)\n return dists[idx]\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":665,"value":"$32","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":663,"value":"md`## Code\nThe remaining cells simply contain some of the extracted code fragments and intermediate values used throughout the notebook:\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":389,"value":"md`#### display`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":98,"value":"margin = ({top: 20, right: 20, bottom: 30, left: 40, middle: 10})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":100,"value":"height = 200","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":348,"value":"h = height - margin.top - margin.bottom","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":350,"value":"w = width - margin.left - margin.right","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":352,"value":"ext = [[margin.left, margin.top], [margin.left + w, margin.top + h]]","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":391,"value":"md`### plotting`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":103,"value":"x = d3.scaleLinear()\n .domain([0, data.length-1])\n .range([margin.left, width - margin.right])","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":365,"value":"x0 = d3.scaleLinear()\n .domain([0, data.length])\n .range([margin.left, width - margin.right])","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2247,"value":"x00 = d3.scaleLinear()\n .domain([0, data.length])\n .range([0, width])","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1264,"value":"xv = d3.scaleLinear()\n .domain([0, variances.length-1])\n .range([0, width])\n .clamp(true)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":883,"value":"xe = d3.scaleLinear()\n .domain([0, energy.length-1])\n .range([0, width])","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":107,"value":"y = d3.scaleLinear()\n .domain(extents)\n .range([height - margin.bottom, margin.top])","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":109,"value":"xAxis = g => g\n .attr(\"transform\", `translate(0,${height - margin.bottom})`)\n .call(d3.axisBottom(x).tickSizeOuter(0))\n .call(g => g.append(\"text\")\n .attr(\"x\", width - margin.right)\n .attr(\"y\", -4)\n .attr(\"fill\", \"#000\")\n .attr(\"font-weight\", \"bold\")\n .attr(\"text-anchor\", \"end\")\n .text('t'))","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":111,"value":"yAxis = g => { \n let axis = d3.axisLeft(y)\n return g.attr(\"transform\", `translate(${margin.left},0)`)\n .call(axis)\n .call(g => g.select(\".domain\").remove())\n .call(g => g.select(\".tick:last-of-type text\").clone()\n .attr(\"x\", 4)\n .attr(\"text-anchor\", \"start\")\n .attr(\"font-weight\", \"bold\")\n .text('f(t)'))\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":538,"value":"md`##### behaviors`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":316,"value":"line = d3.line()\n .x((d,i) => x(i))\n .y((d,i) => y(d))","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":354,"value":"zoom = d3.zoom().scaleExtent([1,10]).extent(ext).translateExtent(ext)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":393,"value":"md`##### line-charts`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":523,"value":"getT = (owner) => {\n const xy = d3.mouse(owner);\n return Math.round(x.invert(xy[0]));\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":399,"value":"drawSignals = (svg, lineFunc=line, cls=\"line\") => {\n for (let i=0; i < sensors.length; i++) {\n let [sensor, name] = [sensors[i], sensorNames[i]];\n svg.append(\"g\").append(\"path\")\n .datum(sensor)\n .attr(\"class\", cls + \" \" + name)\n .attr(\"stroke\", linecolor(i))\n .attr(\"d\", lineFunc)\n }\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":417,"value":"drawAxes = (svg) => {\n svg.append(\"g\").attr(\"class\", \"x-axis\").call(xAxis);\n svg.append(\"g\").attr(\"class\", \"y-axis\").call(yAxis);\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":455,"value":"setupClipPath = (svg) => {\n svg.append(\"defs\").append(\"clipPath\")\n .attr(\"id\", \"clippy\")\n .append(\"rect\")\n .attr(\"width\", w)\n .attr(\"height\", h)\n .attr(\"transform\", \"translate(\" + margin.left + \",\" + margin.top + \")\")\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":837,"value":"class Annotations {\n constructor(points) {\n this.points = points || marked; \n }\n \n getPoint(i) {\n if (i < 0) return 0;\n if (i >= this.points.length) return data.length;\n return this.points[i].t;\n }\n \n bounds(d, i) {\n let {t, lbl} = d;\n let nolen = [this.getPoint(i-1), this.getPoint(i+1)]\n let dt = max_bounds[lbl];\n let maxlen = [clamp(t-dt), clamp(t+dt)]\n maxlen = [Math.max(nolen[0],maxlen[0]), Math.min(nolen[1],maxlen[1])]\n return {nolen, maxlen}\n }\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":770,"value":"$33","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":981,"value":"hovered = (t) => {\n t = Math.min(t, data.length-1);\n t = Math.max(t, 0);\n return marked.find((d) => d.t === t);\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":530,"value":"md`###### annotation`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":544,"value":"setupHighlight = (svg) => {\n let highlightbar = svg.append(\"line\")\n .attr(\"class\", \"highlight\")\n .attr(\"y1\", margin.top)\n .attr(\"y2\", margin.top + h)\n \n let cLbl = svg.append(\"text\")\n .attr(\"class\", \"highlight\")\n .attr(\"y\", 10)\n \n function drawHighlight(t) {\n if (hovered(t)) {\n clearHighlight();\n return;\n }\n t = Math.min(t, data.length-1);\n t = Math.max(t, 0);\n highlightbar.attr(\"x1\", x(t))\n .attr(\"x2\", x(t))\n .attr(\"stroke\", labelcolor(data[t].label))\n .classed(\"hidden\", false)\n cLbl.attr(\"class\", \"cluster-label\")\n .attr(\"x\", x(t))\n .attr(\"dx\", -3)\n .classed(\"hidden\", false)\n .text(kmeans[t])\n }\n \n function clearHighlight() {\n highlightbar.classed(\"hidden\", true);\n cLbl.classed(\"hidden\", true);\n }\n return [drawHighlight, clearHighlight]\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":560,"value":"drawAnnotations = (svg, points, cls=\"annotation\") => {\n let bars = svg.selectAll(\"line.annotation\").data(points);\n bars.enter()\n .append(\"line\")\n .attr(\"class\", cls)\n .attr(\"y1\", margin.top)\n .attr(\"y2\", margin.top + h)\n .attr(\"stroke\", (d) => labelcolor(d.lbl))\n .merge(bars)\n .attr(\"x1\", (d) => x(d.t))\n .attr(\"x2\", (d) => x(d.t))\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":949,"value":"addAnnotation = (t) => {\n let [m, m2] = [marked, markedBounds];\n if (hovered(t)) {console.warn('duplicate annotation:',t); return;}\n m.push({t, lbl: data[t].label})\n m.sort((a,b) => a.t-b.t);\n m2.push({t, lbl: data[t].label, point: true})\n m2.sort((a,b) => a.t-b.t);\n viewof marked.value = m;\n viewof markedBounds.value = m2;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1004,"value":"removeAnnotation = (target) => {\n let el = d3.select(target);\n let t = el.data()[0].t;\n el.transition().duration(300)\n .attr(\"y1\", margin.top + (h/2))\n .attr(\"y2\", margin.top + (h/2))\n .remove()\n .on(\"end\", () => {\n viewof marked.value = marked.filter((d) => d.t !== t)\n viewof markedBounds.value = markedBounds.filter((d) => d.t !== t)\n console.log('REMOVED', t);\n })\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":748,"value":"md`###### boundary initialization`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":802,"value":"$34","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":864,"value":"W = (bounds) => {return x(bounds[1]) - x(bounds[0])}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":894,"value":"H = (d,i) => { return stack(d,i)[1] - stack(d,i)[0] }","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1084,"value":"_x = (d) => {\n if (d.point) return x(d.t) -.5;\n else return x(d.bounds[0]);\n }","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1087,"value":"_w = (d) => {\n if (d.point) return 1;\n else return W(d.bounds);\n }","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1611,"value":"stack = (d,i) => {\n if (stacking === \"label\") return y_regions[d.lbl];\n if (stacking === \"fancy\") return y_regions[(d.lbl + i) % 3 ];\n if (stacking === \"alternate\") return y_regions[i % 3];\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":873,"value":"drawBoundaries = (svg, points, cls=\"boundary\") => {\n let lbls = svg.selectAll(\"rect.boundary\").data(points);\n lbls = lbls.enter()\n .append(\"rect\")\n .attr(\"class\", cls)\n .attr(\"t\", (d) => d.t)\n .attr(\"y\", (d,i) => { return stack(d,i)[0] + margin.top })\n .attr(\"height\", (d,i) => H(d,i))\n .attr(\"stroke\", (d) => labelcolor(d.lbl))\n .attr(\"fill\", (d) => labelcolor(d.lbl))\n .merge(lbls)\n .attr(\"x\", _x)\n .attr(\"width\", _w)\n return lbls;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1144,"value":"md`###### non-motion removal`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2202,"value":"function drawRegions(svg, regs, xscale, cls='non-motion') {\n let rects = svg.selectAll(\"rect.\" + cls).data(regs);\n rects = rects.enter()\n .append('rect')\n .attr('class', cls)\n .attr(\"y\", 0)\n .attr(\"height\", svg.attr(\"height\"))\n .merge(rects)\n .attr(\"x\", (d) => xscale(d.start))\n .attr(\"width\", (d) => xscale(d.end) - xscale(d.start))\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1584,"value":"$35","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1148,"value":"$36","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1741,"value":"xb = {\n if (trim_method === \"energy\") return xe;\n if (trim_method === \"var\") return xv;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1744,"value":"convert = (rt) => x00.invert(xb(rt))","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1790,"value":"drawTrims = (svg, points, cls=\"boundary trim\") => {\n let lbls = svg.selectAll(\"rect.boundary.trim\").data(points);\n lbls = lbls.enter()\n .append(\"rect\")\n .attr(\"class\", cls)\n .attr(\"t\", (d) => d.t)\n .attr(\"y\", (d,i) => { return stack(d,i)[0] + margin.top })\n .attr(\"height\", (d,i) => H(d,i))\n .attr(\"stroke\", (d) => labelcolor(d.lbl))\n .attr(\"fill\", (d) => labelcolor(d.lbl))\n .merge(lbls)\n .attr(\"x\", (d) => x(d.L))\n .attr(\"width\", (d) => x(d.R) - x(d.L))\n return lbls;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":369,"value":"md`#### sensors`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":13,"value":"Ax = data.map(d => d.data[0])","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":20,"value":"Ay = data.map(d => d.data[1])","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":22,"value":"Az = data.map(d => d.data[2])","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":24,"value":"Gx = data.map(d => d.data[3])","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":26,"value":"Gy = data.map(d => d.data[4])","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":28,"value":"Gz = data.map(d => d.data[5])","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":320,"value":"sensors = [Ax, Ay, Az, Gx, Gy, Gz]","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":327,"value":"sensorNames = [\"Ax\", \"Ay\", \"Az\", \"Gx\", \"Gy\", \"Gz\"]","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":299,"value":"extents = {\n let exts = sensors.map(s => d3.extent(s));\n let result = [d3.min(exts, e => e[0]), d3.max(exts, e => e[1])]\n return [-5000, 5000]\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1203,"value":"v = transpose(variances)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1219,"value":"e = transpose(energy)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":374,"value":"md`### helper methods`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":122,"value":"boundaries = (labels) => {\n const lbls = Array.from(labels.entries()); // lists both value and index\n let change = (l) => {\n let i = l[0];\n let lbl = l[1];\n let prev = labels[i-1];\n return prev !== lbl;\n }\n let bounds = lbls.filter(change); // filter list to boundary changes\n let result = bounds.map((b,i,array) => {\n let end = array[i+1] ? array[i+1][0] : labels.length;\n let start = b[0];\n let label = b[1];\n return {start, end, label}\n })\n return result;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":730,"value":"boundLength = (d) => d.end - d.start","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":737,"value":"linspace = (start, stop, nsteps) => {\n let delta = (stop-start)/(nsteps-1)\n return d3.range(start, stop+delta, delta).slice(0, nsteps)\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":756,"value":"clamp = (t) => {\n if (t < 0) return 0;\n if (t > data.length-1) return data.length-1;\n return t;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1215,"value":"transpose = (array) => { return array[0].map((col, i) => array.map(row => row[i])); }","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1398,"value":"thresh = (thresholds) => {\n return (d) => { \n if (d > thresholds[0]) return \"start\";\n if (d < thresholds[1]) return \"stop\";\n return \"-\";\n }\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1438,"value":"function* motionState(transitions) {\n let startMotion = (t) => t === \"start\";\n let stopMotion = (t) => t === \"stop\";\n let state = \"non-motion\";\n for (let T of transitions) {\n if (T.some(startMotion)) state = \"motion\";\n if (T.every(stopMotion)) state = \"non-motion\";\n yield state;\n }\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1639,"value":"magsquared = (acc,cur) => acc + cur*cur","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1725,"value":"overlaps = (A,B) => {\n let [a1,a2] = A;\n let [b1,b2] = B;\n return (a1 >= b1 && a1 <= b2) || (a2 >= b1 && a2 <= b2)\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2238,"value":"contains = (A,B) => {\n let [a1,a2] = A;\n let [b1,b2] = B;\n return (a1 <= b1 && a2 >= b2)\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1885,"value":"tlens = (d) => {\n let lengths = d.map(t => t.bounds[1] - t.bounds[0]);\n return lengths.join(', ')\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2027,"value":"pairs = (array) => {\n let result = [];\n for (let i=0; i < array.length; i++)\n for (let j=i+1; j < array.length; j++)\n result.push([array[i], array[j]]);\n return result;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":620,"value":"md`##### colors`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":120,"value":"clustercolor = d3.scaleLinear()\n .domain([0,k])\n .range([\"cyan\", \"indianred\"])","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":36,"value":"linecolor = (dim) => {\n return linecolormap[dim]\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":473,"value":"labelcolor = (label) => {\n return labelcolormap[label]\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":636,"value":"md`##### inline content`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":16,"value":"function sparkline(values, dim=0) {\n const width = 128;\n const height = 20;\n const x = d3.scaleLinear().domain([0, values.length - 1]).range([0.5, width - 0.5]);\n const y = d3.scaleLinear().domain(d3.extent(values)).range([height - 0.5, 0.5]);\n const context = DOM.context2d(width, height);\n const line = d3.line().x((d, i) => x(i)).y(y).context(context);\n context.strokeStyle=linecolor(dim);\n context.beginPath(), line(values), context.stroke();\n return context.canvas;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":183,"value":"genTable = () => {\n let result = \"| cluster | % data | # streaks | max streak | avg streak |\\n\"\n + \"|---------|--------| --------- | ---------- | ---------- |\\n\"\n for (let i of cIdxs) {\n result += \"| **\" + i + \"** | \" \n + getPercent(i) + \" | \" \n + numStreak(i) + \" | \" \n + maxStreak(i) + \" | \" \n + avgStreak(i) + \" |\\n\"\n }\n return result;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":638,"value":"md`###### lookups`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":210,"value":"getPercent = (i) => ((counts[i] / data.length)*100).toFixed(2) + \"%\"","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":213,"value":"numStreak = (i) => streaks[i].length","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":219,"value":"maxStreak = (i) => d3.max(streaks[i], (d) => d.end-d.start)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":258,"value":"avgStreak = (i) => d3.mean(streaks[i], (d) => d.end-d.start).toFixed(2)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":618,"value":"md`##### maps`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":37,"value":"linecolormap = {\n return {\n 0: \"cyan\",\n 1: \"aqua\",\n 2: \"skyblue\",\n 3: \"mediumspringgreen\",\n 4: \"palegreen\",\n 5: \"springgreen\"\n }\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":484,"value":"labelcolormap = {\n return {\n 0: \"lightgray\",\n 1: \"magenta\",\n 2: \"salmon\"\n }\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":609,"value":"labelmap = {\n return {\n 0: \"null\",\n 1: \"pills\",\n 2: \"move\"\n }\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":385,"value":"md`#### intermediate data`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":711,"value":"true_labels = data.map(d => d.label)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":709,"value":"true_bounds = boundaries(true_labels)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":142,"value":"bounds = boundaries(kmeans)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":158,"value":"counts = {\n let result = {}\n for (let i = 0; i < k; i++) {\n result[i] = kmeans.filter(d => d === i).length;\n }\n return result;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":225,"value":"cIdxs = {\n let result = []\n for (let i = 0; i < k; i++) result.push(i);\n return result;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":220,"value":"streaks = cIdxs.map(i => bounds.filter((b) => b.label === i))","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":727,"value":"max_bounds = {\n let result = [];\n result.push(d3.max(true_bounds.filter(d => d.label === 0).map(boundLength)))\n result.push(d3.max(true_bounds.filter(d => d.label === 1).map(boundLength)))\n result.push(d3.max(true_bounds.filter(d => d.label === 2).map(boundLength)))\n return result\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":746,"value":"y_regions = {\n let spaces = linspace(0,h,4);\n let result = [];\n for (let i = 0; i < 3; i++)\n result.push([spaces[i], spaces[i+1]]);\n return result;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1393,"value":"vtt = v.map((vi) => vi.map(thresh(THRESHOLDS)))","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1455,"value":"v_thresholds = boundaries(Array.from(motionState(transpose(vtt.slice(0,3)))))","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1655,"value":"e_thresholds = {\n let E = energy_magnitude.map((e) => e <= threshold ? \"non-motion\" : \"motion\")\n return boundaries(E)\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2157,"value":"smoothed_regions = {\n let regs = e_thresholds.map(r => r)\n let result = []\n let cur = regs.shift(); // take first element as initial\n for (let r of regs) {\n if (r.end-r.start < smoothing || r.label === cur.label) { cur.end = r.end; }\n else { result.push(cur); cur = r; }\n }\n result.push(cur);\n return result;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1636,"value":"energy_magnitude = energy.map(ei => ei.reduce(magsquared,0))","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1716,"value":"regions = {\n if (trim_method === \"energy\") return smoothed_regions;\n if (trim_method === \"var\") return v_thresholds;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1719,"value":"nonmotion_regions = regions.filter(r => r.label === \"non-motion\")","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1722,"value":"motion_regions = regions.filter(r => r.label === \"motion\")","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1797,"value":"clip_bounds = boundary_init.boundpoints.map((d) => {\n const regs = (d.lbl > 0) ? nonmotion_regions : motion_regions;\n let result = []\n for (let r of regs) {\n let rb = [convert(r.start), convert(r.end)]\n if (overlaps(rb, d.bounds)) {\n let side;\n if (rb[0] < d.t && rb[1] < d.t) side = \"left\";\n else if (rb[0] > d.t && rb[1] > d.t) side = \"right\";\n else side = \"?\";\n let R = {start: Math.round(rb[0]), end: Math.round(rb[1]), t:d.t, side, label:r.label}\n result.push(R);\n }\n }\n let L = d3.max(result.filter(d => d.side === \"left\"), (d) => d.end) || d.bounds[0]\n let R = d3.min(result.filter(d => d.side === \"right\"), (d) => d.start) || d.bounds[1]\n return Object.assign({},d,{L,R}, {trimmed: false});\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2123,"value":"clip_bounds2 = boundary_init.boundpoints.map((d) => {\n const regs = (d.lbl > 0) ? nonmotion_regions : motion_regions;\n let result = []\n for (let r of smoothed_regions) {\n let [s,e] = [convert(r.start), convert(r.end)]\n let se = [s,e]\n if (d.t > s && d.t <= e)\n result.push({r,s,e})\n }\n return {result,d}\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2183,"value":"nm_smooth = smoothed_regions.filter(r => r.label === \"non-motion\")","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2142,"value":"boundary_init.boundpoints.map((d) => {\n return {t: x0(d.t), l: x0(d.bounds[0]), r: x0(d.bounds[1])}\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1882,"value":"cts = [t.filter(d => d.lbl === 0), t.filter(d => d.lbl === 1), t.filter(d => d.lbl === 2)]","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2079,"value":"TEMPLATE_PAIRS = cts.map(group => group.map((d) => d.t)).map(pairs)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2085,"value":"templates = cts.map(ci => { \n return ci.map((d) => {\n let result = kmeans.slice(d.bounds[0], d.bounds[1]);\n console.count('RES');\n console.log('RES', result, d);\n return result;\n })\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":950,"value":"md`###### shared views`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":503,"value":"viewof marked = new View([])","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":963,"value":"viewof markedBounds = new View([])","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1843,"value":"viewof markedBounds2 = new View([])","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1586,"value":"ZOOMS = new View({\n \"annotation\": d3.zoomIdentity,\n \"boundary-init\": d3.zoomIdentity,\n \"nonmotion-removal\": d3.zoomIdentity\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1588,"value":"ZOOMS.value","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1866,"value":"md`###### aliases`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1868,"value":"n = {\n let result = {};\n let points = await marked;\n result.Ø = points.filter(d => d.lbl === 0).length;\n result.pills = points.filter(d => d.lbl === 1).length;\n result.move = points.filter(d => d.lbl === 2).length;\n return result\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1871,"value":"t = markedBounds2","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2396,"value":"threshold = ENERGY_THRESHOLDS[0]","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2401,"value":"smoothing = ENERGY_THRESHOLDS[0]","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2282,"value":"md`###### cluster distances`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2276,"value":"cluster_distances = {\n let text = await d3.text(\"https://gist.githubusercontent.com/lwthatcher/0facab62c8d14bacb8eea9c8b97d4cbc/raw/d39a2a9d848353ac6ecb3ef5ac32de81852af9f1/short.kmeans.dists.csv\")\n let kvs = d3.csvParseRows(text, (row) => {\n let values = row.map(v => +v);\n let N = inverse_length(values.length)\n return [N,values]\n })\n return new Map(kvs)\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2293,"value":"inverse_length = (x) => {\n return parseInt(0.5 * (Math.sqrt(8*x + 1) - 1))\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2360,"value":"md`###### Latex Equations`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2352,"value":"index_equation = tex`dist_{index}(i,j) = ki + \\cfrac{(i-1)i}{2} + (j-i)`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":449,"value":"md`#### imports/style sheets`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":4,"value":"d3 = require(\"d3\")","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1904,"value":"math = require(\"https://unpkg.com/mathjs@5.1.1/dist/math.min.js\")","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":507,"value":"class View {\n constructor(value) {\n this._list = [];\n this._value = value;\n }\n get value() {\n return this._value;\n }\n set value(value) {\n this._value = value;\n this.dispatchEvent({type: \"input\", value});\n }\n addEventListener(type, listener) {\n if (type != \"input\" || this._list.includes(listener)) return;\n this._list = [listener].concat(this._list);\n }\n removeEventListener(type, listener) {\n if (type != \"input\") return;\n this._list = this._list.filter(l => l !== listener);\n }\n dispatchEvent(event) {\n const p = Promise.resolve(event);\n this._list.forEach(l => p.then(l));\n }\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":336,"value":"$37","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"}],"files":"$W38","stdlib":"1"}} 31:"$30:body"