Multi-column CSV parser for Javascript (depends on https://gist.github.com/u8sand/6b1ebfd40331af69764b9754ae6d1be4 and https://gist.github.com/u8sand/39892f1ac91920b55f9246076c64debd)

parse.js

import { matrix_flatten, matrix_slice, slice } from './matrix'
import { range } from './range'

export function count_first_na(L) {
  for (let i = 0; i < L.length; i++) {
    if (L[i] != null)
      return i
  }
  throw new Error('NaNs not identified')
}

export function dictzip(header, data) {
  const D = {}
  for (let i = 0; i < Math.min(header.length, data.length); i++)
    D[header[i]] = data[i]
  return D
}

export function *parse(matrix) {
  const border_x = count_first_na(matrix[0])
  const border_y = count_first_na(matrix_flatten(matrix_slice(matrix, 0, null)))

  if (border_y <= 0 || border_x <= 0)
    throw new Error('Invalid formatting')

  const header_x = matrix_flatten(matrix_slice(matrix, border_x, slice(null, border_y + 1)))
  const header_y = matrix_flatten(matrix_slice(matrix, slice(null, border_x + 1), border_y))

  for (const y of range(border_y + 1, matrix.length)) {
    for (const x of range(border_x + 1, matrix[0].length)) {
      yield {
        'meta': {
          ...dictzip(
            header_x,
            matrix_flatten(matrix_slice(matrix, x, slice(null, border_y + 1)))
          ),
          ...dictzip(
            header_y,
            matrix_flatten(matrix_slice(matrix, slice(null, border_x + 1), y))
          ),
        },
        'data': matrix[y][x],
      }
    }
  }
}

export function parse_csv(data) {
  const lines_re = /[\n\r]/
  const line_re = /(^(("([^"]*)")|([^,]*)),)|((("([^"]*)")|([^,]*)),)|((("([^"]+)")|([^,]+))$)/g

  function *parse_line(line) {
    let m
    while (m = line_re.exec(line)) {
      let r =
           m[2] || m[7] || m[12]
      try {
        r = JSON.parse(r)
      } catch(e) {}

      if (r === undefined || r.length === 0) {
        yield null
      } else {
        yield r
      }
    }
    if (line[line.length - 1] === ',') {
      yield null
    }
  }

  function *parse_lines(lines) {
    for (const line of lines) {
      yield [...parse_line(line)]
    }
  }

  return [...parse_lines(data.split(lines_re))]
}

export function parse_file(data) {
  return parse(parse_csv(data))
}

parse.test.js

import assert from 'assert'
import { count_first_na, dictzip, parse, parse_csv } from './parse'

describe('util.parse', () => {
  it('count_first_na', () => {
    assert.equal(
      count_first_na([null, null, 1]),
      2
    )
  })

  it('dictzip', () => {
    assert.deepEqual(
      dictzip(['a', 'b', 'c'], [1, 2, 3]),
      {
        'a': 1,
        'b': 2,
        'c': 3,
      }
    )
  })

  it('parse csv', () => {
    assert.deepEqual(
      parse_csv(',Cell Type,MCF10\n,Drug,mydrug\nGene,Cell Line,\nSTAT3,mycellline,0.4'),
      [
        [ null, "Cell Type", "MCF10" ],
        [ null, "Drug", "mydrug" ],
        [ "Gene", "Cell Line", null ],
        [ "STAT3", "mycellline", 0.4 ],
      ]
    )
  })

  it('parse', () => {
    assert.deepEqual(
      [...parse([
        [         null,  'Cell Type',      'MCF10', 'mycelltype',],
        [         null,       'Drug',     'mydrug','myotherdrug',],
        [      'Gene',  'Cell Line',          null,          null,],
        [     'STAT3',       'Blah',          0.1,          0.2,],
      ])],
      [
        {
          'data': 0.1,
          'meta': {
            'Cell Line': 'Blah',
            'Cell Type': 'MCF10',
            'Drug': 'mydrug',
            'Gene': 'STAT3'
          }
        },
        {
          'data': 0.2,
          'meta': {
            'Cell Line': 'Blah',
            'Cell Type': 'mycelltype',
            'Drug': 'myotherdrug',
            'Gene': 'STAT3'
          }
        }
      ]
    )
  })
})