2f:[] 2d:{"title":"Work for a model evaluation website","cells":[{"id":0,"value":"# Work for a model evaluation website","pinned":true,"hidden":false,"mode":"md"},{"id":36,"value":"viewof username = Inputs.password({label: html`Set username`, value: \"\"})\n","pinned":false,"hidden":false,"mode":"ojs"},{"id":1027,"value":"password","pinned":false,"hidden":false,"mode":"ojs"},{"id":41,"value":"viewof password = Inputs.password({label: html`Set password`, value: \"\"})\n","pinned":false,"hidden":false,"mode":"ojs"},{"id":4,"value":"viewof csvfile = Inputs.file({\n label: md`__Observations to evaluate__ 📁 ([Example here](https://raw.githubusercontent.com/ThomasFaria/codif-ape-API/main/website/exemples/exemple.csv))`,\n accept: \".csv\",\n required: true\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":366,"value":"grid = await html`

${viewof quantile}

Accuracy: ${Accuracy}

Seuil: ${Seuil}

${distrib}

${distrib_adjusted}

`","pinned":false,"hidden":false,"mode":"ojs"},{"id":212,"value":"viewof quantile = Inputs.range([0, 100], {label: \"Taux de classification automatique\", step: 1, value: 80})","pinned":false,"hidden":false,"mode":"ojs"},{"id":118,"value":"distrib = Plot.plot({\n width: screen.width * 0.7,\n color: { scheme: \"BuRd\", legend: true },\n marks: [\n Plot.rectY(\n processed_response,\n Plot.binX(\n { y: \"count\" },\n { x: \"IC\", fill: \"Result\", mixBlendMode: \"multiply\", tip: true }\n )\n ),\n Plot.ruleY([0]),\n Plot.ruleX([Seuil], { strokeWidth: 2, tip: true })\n ]\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":997,"value":"distrib_adjusted = Plot.plot({\n width: screen.width * 0.7,\n y: { percent: true },\n color: { scheme: \"BuRd\", legend: true },\n marks: [\n Plot.rectY(\n processed_response,\n Plot.normalizeY(\n \"sum\", // normalize each series by the sum per series\n Plot.binX(\n { y2: \"count\" }, // disable implicit stack transform\n { x: \"IC\", fill: \"Result\", mixBlendMode: \"multiply\", tip: true }\n )\n )\n ),\n Plot.ruleY([0]),\n Plot.ruleX([Seuil], { strokeWidth: 2, tip: true })\n ]\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":1005,"value":"processed_response.filter((d) => d.IC > 1)","pinned":true,"hidden":false,"mode":"ojs"},{"id":794,"value":"grid2 = await html`

${codif_seuil}

${acc_seuil}

${acc_codif}

`","pinned":false,"hidden":false,"mode":"ojs"},{"id":141,"value":"Seuil = round(\n d3.quantile(\n processed_response.map((d) => d.IC),\n 1 - quantile / 100\n ),\n 2\n)","pinned":false,"hidden":false,"mode":"ojs"},{"id":187,"value":"Accuracy = round(d3.mean(subset.map((d) => d.Result)) * 100, 2)","pinned":false,"hidden":false,"mode":"ojs"},{"id":310,"value":"## Analyse des erreurs","pinned":true,"hidden":false,"mode":"md"},{"id":827,"value":"Plot.barY(count_errors.slice(0, 15), {\n x: \"value\",\n y: \"frequency\",\n sort: { x: \"-y\" },\n tip: true\n}).plot()","pinned":true,"hidden":false,"mode":"ojs"},{"id":964,"value":"viewof search = Inputs.search(subset.filter((d) => 1 - d.Result))","pinned":true,"hidden":false,"mode":"ojs"},{"id":198,"value":"WrongPredictions = Inputs.table(search)","pinned":true,"hidden":false,"mode":"ojs"},{"id":933,"value":"## Impact du seuil de reprise sur la performance et le taux ","pinned":false,"hidden":false,"mode":"md"},{"id":935,"value":"codif_seuil = Plot.plot({\n grid: true,\n y: {\n label: \"↑ Taux de classification automatique (%)\",\n percent: true\n },\n x: {\n label: \"→ Seuil choisi\"\n },\n marks: [\n Plot.line(ic_tx_classif, {\n x: \"seuils\",\n y: \"codif_auto\",\n stroke: \"black\",\n strokeWidth: 2\n }),\n Plot.tip(\n ic_tx_classif,\n Plot.pointer({\n x: \"seuils\",\n y: \"codif_auto\",\n title: (d) => `Accuracy : ${round(d.accuracies, 2)}\\nSeuil : ${round(\n d.seuils,\n 2\n )}\\nTaux classification automatique: \n${round(d.codif_auto * 100, 2)} %`\n })\n ),\n Plot.tip([`Seuil à 0.6`], {\n x: ic_tx_classif[60].seuils,\n y: ic_tx_classif[60].codif_auto,\n anchor: \"bottom\"\n })\n ],\n color: { legend: true }\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":937,"value":"acc_seuil = Plot.plot({\n grid: true,\n y: {\n label: \"↑ Accuracy (%)\",\n percent: true\n },\n x: {\n label: \"→ Seuil choisi\"\n },\n marks: [\n Plot.line(ic_tx_classif, {\n x: \"seuils\",\n y: \"accuracies\",\n stroke: \"black\",\n strokeWidth: 2\n }),\n Plot.tip(\n ic_tx_classif,\n Plot.pointer({\n x: \"seuils\",\n y: \"accuracies\",\n title: (d) => `Accuracy : ${round(d.accuracies, 2)}\\nSeuil : ${round(\n d.seuils,\n 2\n )}\\nTaux classification automatique: \n${round(d.codif_auto * 100, 2)} %`\n })\n ),\n Plot.tip([`Seuil à 0.6`], {\n x: ic_tx_classif[60].seuils,\n y: ic_tx_classif[60].accuracies,\n anchor: \"bottom\"\n })\n ],\n color: { legend: true }\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":939,"value":"acc_codif = Plot.plot({\n grid: true,\n y: {\n label: \"↑ Accuracy (%)\",\n percent: true\n },\n x: {\n label: \"→ Taux de classification automatique (%)\",\n percent: true\n },\n marks: [\n Plot.line(ic_tx_classif, {\n x: \"codif_auto\",\n y: \"accuracies\",\n stroke: \"black\",\n strokeWidth: 2\n }),\n Plot.tip(\n ic_tx_classif,\n Plot.pointer({\n x: \"codif_auto\",\n y: \"accuracies\",\n title: (d) => `Accuracy : ${round(d.accuracies, 2)}\\nSeuil : ${round(\n d.seuils,\n 2\n )}\\nTaux classification automatique: \n${round(d.codif_auto * 100, 2)} %`\n })\n ),\n Plot.tip([`Seuil à 0.6`], {\n x: ic_tx_classif[60].codif_auto,\n y: ic_tx_classif[60].accuracies,\n anchor: \"bottom\"\n })\n ],\n color: { legend: true }\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":267,"value":"## Data stuffs","pinned":true,"hidden":false,"mode":"md"},{"id":941,"value":"ic_tx_classif = {\n const newArray = [];\n for (let i = 0; i <= 1.01; i += 0.01) {\n let seuil = round(i, 2);\n newArray.push({\n seuils: seuil,\n codif_auto:\n processed_response.filter((d) => d.IC > seuil).length /\n processed_response.length,\n accuracies: get_accuracies_from_threshold(processed_response, i, {\n round_int: 15\n })\n });\n }\n return newArray;\n}","pinned":true,"hidden":false,"mode":"ojs"},{"id":12,"value":"data = db.query(\n `\nSELECT text_description, type_, nature, surface, event, code\nFROM liasses\n`\n)","pinned":false,"hidden":false,"mode":"ojs"},{"id":77,"value":"transformedData = data.reduce((acc, obj) => {\n for (const [key, value] of Object.entries(obj)) {\n if (!acc.hasOwnProperty(key)) {\n acc[key] = [];\n }\n acc[key].push(value);\n }\n return acc;\n}, {})","pinned":false,"hidden":false,"mode":"ojs"},{"id":10,"value":"db = DuckDBClient.of({\n liasses: await csvfile.csv()\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":120,"value":"processed_response = Object.keys(response.IC).map((key) => ({\n IC: response.IC[key] == 1 ? response.IC[key] - 1e-16 : response.IC[key],\n Probability: response.Probability[key],\n Prediction: response.Prediction[key],\n Code: response.Code[key],\n Result: response.Result[key],\n Lib: response.Lib[key]\n}))","pinned":false,"hidden":false,"mode":"ojs"},{"id":98,"value":"response = queryBatchApi(username, password, data)","pinned":true,"hidden":false,"mode":"ojs"},{"id":899,"value":"Inputs.table(processed_response)","pinned":false,"hidden":false,"mode":"ojs"},{"id":183,"value":"subset = processed_response.filter((d) => d.IC > Seuil)","pinned":false,"hidden":false,"mode":"ojs"},{"id":943,"value":"quantiles = Array.from({ length: 101 }, (_, i) => i / 100) // Array with quantile values from 0 to 1","pinned":false,"hidden":false,"mode":"ojs"},{"id":945,"value":"seuils = quantiles.map((quantile) =>\n d3.quantile(\n processed_response.map((d) => d.IC),\n 1 - quantile\n )\n)","pinned":false,"hidden":false,"mode":"ojs"},{"id":600,"value":"count_errors = valueCounts(\n subset.filter((d) => 1 - d.Result).map((d) => d.Code)\n)","pinned":false,"hidden":false,"mode":"ojs"},{"id":269,"value":"## Functions","pinned":false,"hidden":false,"mode":"md"},{"id":919,"value":"async function queryinfo(username, password) {\n const url = \"https://codification-ape.lab.sspcloud.fr/\";\n\n const response = await fetch(url, {\n method: \"GET\",\n headers: {\n \"Content-Type\": \"application/json\",\n Authorization: `Basic ${btoa(`${username}:${password}`)}`\n }\n });\n\n if (response.ok) {\n return response.json();\n } else if (response.status === 400) {\n console.log((await response.json()).detail);\n } else {\n console.log(\"Error occurred while querying the API.\");\n return response;\n }\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":27,"value":"async function queryBatchApi(username, password, data) {\n const url = \"https://codification-ape.lab.sspcloud.fr/evaluation\";\n\n // Create the request body\n const request_body = data.reduce((acc, obj) => {\n for (const [key, value] of Object.entries(obj)) {\n if (!acc.hasOwnProperty(key)) {\n acc[key] = [];\n }\n acc[key].push(value);\n }\n return acc;\n }, {});\n\n const response = await fetch(url, {\n method: \"POST\",\n headers: {\n \"Content-Type\": \"application/json\",\n Authorization: `Basic ${btoa(`${username}:${password}`)}`\n },\n body: JSON.stringify(request_body)\n });\n\n if (response.ok) {\n return response.json();\n } else if (response.status === 400) {\n console.log((await response.json()).detail);\n } else {\n console.log(\"Error occurred while querying the API.\");\n return response;\n }\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":719,"value":"function round(d, int) {\n return Math.round(d * 10 ** int) / 10 ** int;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":733,"value":"function get_accuracies_from_threshold(\n data,\n threshold,\n { round_int = 2 } = {}\n) {\n let acc = d3.mean(data.filter((d) => d.IC > threshold).map((d) => d.Result));\n acc = acc === undefined ? 1 : round(acc, round_int);\n return acc;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":602,"value":"function valueCounts(array, { freq = false } = {}) {\n const counts = {};\n\n for (let i = 0; i < array.length; i++) {\n const value = array[i];\n\n if (counts[value]) {\n counts[value]++;\n } else {\n counts[value] = 1;\n }\n }\n\n const result = [];\n\n const arrayLength = array.length;\n for (const value in counts) {\n const frequency = freq ? counts[value] / arrayLength : counts[value];\n result.push({ value: value, frequency: frequency });\n }\n result.sort((a, b) => b.frequency - a.frequency); // Sort by frequency in descending order\n\n return result;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":274,"value":"## Imports","pinned":false,"hidden":false,"mode":"md"},{"id":388,"value":"","pinned":false,"hidden":false,"mode":"html"}],"files":"$W2f","stdlib":"1","titling":"auto"}