uniprotParser/index.ts

// https://ftp.uniprot.org/pub/databases/uniprot/current_release/knowledgebase/taxonomic_divisions/uniprot_sprot_human.xml.gz

import fs from 'fs'
import readline from 'readline'
import { XMLParser } from 'fast-xml-parser'

const line$ = (path: string) => readline.createInterface({
    input: fs.createReadStream(path),
    crlfDelay: Infinity
})

const makeIndex = async (filePath: string, indexPath?: string) => {
    indexPath = indexPath || filePath + '.jsi'

    let byteAcc = 0
    const fromSel = new RegExp("^<entry")
    const toSel = new RegExp("^</entry>")
    const valSel = new RegExp('<accession>')
    let tmp =  {values:[]} as any
    for await (const line of line$(filePath)) {
        if(fromSel.test(line)) tmp['from'] = byteAcc

        byteAcc += (line.length + 1)

        if (valSel.test(line)) tmp['values'].push(line.match("<accession>(.*?)</accession>")![1]) // 'uck 

        if(toSel.test(line)) {
            await fs.promises.appendFile(indexPath, tmp.values.join(';') + '\t' + tmp.from + '\t' + byteAcc + '\n')
            tmp = {values:[]}
        }
    }
}

const readOffset = (path: string, from:number, to:number) => {
    return new Promise<string>(async (resolve, reject) => {
        const size = to - from
        const buffer = Buffer.alloc(size);
        let filehandle = null;
        try {
            filehandle = await fs.promises.open(path, 'r+');
            await filehandle.read(buffer, 0, buffer.length, from);
        } finally {
            if (filehandle) {
                await filehandle.close()
                resolve(buffer.toString())
            }
        }
    })
}

const getEntryOffset = async (dbPath:string, accession:string): Promise<number[]> => {
    const indexPath = dbPath + '.jsi'
    if (!fs.existsSync(indexPath)) await makeIndex(dbPath)
    const lineSel = new RegExp(accession)
    for await (const line of line$(indexPath)) {
        if (lineSel.test(line)) return [Number(line.split('\t')[1]),Number(line.split('\t')[2])]
    }
    return [0, 0]
}

const getEnrty = async (dbPath:string, accession:string) => {
    const parser = new XMLParser({
        ignoreAttributes: false, 
        alwaysCreateTextNode: false, 
        attributeNamePrefix: "",
        textNodeName: "value",
        allowBooleanAttributes: true,
    })

    const offsets = await getEntryOffset(dbPath, accession)
    return parser.parse(await readOffset(dbPath, offsets[0], offsets[1]))
}

export { makeIndex, readOffset, getEnrty }