| | |
|---|
| | | if (trim) { |
|---|
| | | const includesCheck = (tok: string) => |
|---|
| | | tokenizedTerms.some((term) => tok.toLowerCase().startsWith(term.toLowerCase())) |
|---|
| | | const occurencesIndices = tokenizedText.map(includesCheck) |
|---|
| | | const occurrencesIndices = tokenizedText.map(includesCheck) |
|---|
| | | |
|---|
| | | let bestSum = 0 |
|---|
| | | let bestIndex = 0 |
|---|
| | | for (let i = 0; i < Math.max(tokenizedText.length - contextWindowWords, 0); i++) { |
|---|
| | | const window = occurencesIndices.slice(i, i + contextWindowWords) |
|---|
| | | const window = occurrencesIndices.slice(i, i + contextWindowWords) |
|---|
| | | const windowSum = window.reduce((total, cur) => total + (cur ? 1 : 0), 0) |
|---|
| | | if (windowSum >= bestSum) { |
|---|
| | | bestSum = windowSum |
|---|
| | |
|---|
| | | const termLower = term.toLowerCase() |
|---|
| | | let matching = tags.filter((str) => str.includes(termLower)) |
|---|
| | | |
|---|
| | | // Substract matching from original tags, then push difference |
|---|
| | | // Subtract matching from original tags, then push difference |
|---|
| | | if (matching.length > 0) { |
|---|
| | | let difference = tags.filter((x) => !matching.includes(x)) |
|---|
| | | |
|---|
| | |
|---|
| | | // setup index if it hasn't been already |
|---|
| | | if (!index) { |
|---|
| | | index = new Document({ |
|---|
| | | cache: true, |
|---|
| | | charset: "latin:extra", |
|---|
| | | optimize: true, |
|---|
| | | encode: encoder, |
|---|
