nfv
Version:
[Updated to use ndv package instead dv] FormVision is a node.js library for extracting data from scanned forms
132 lines (113 loc) • 5.2 kB
text/coffeescript
dv = require 'ndv'
async = require 'async'
{findBarcodes} = require './find_barcodes'
{findCheckboxes} = require './find_checkboxes'
{findText} = require './find_text'
{estimateTransform} = require './estimate_transform'
{matchBarcodes} = require './match_barcodes'
{matchText} = require './match_text'
{matchCheckboxes} = require './match_checkboxes'
{unpack} = require './schema'
matchByPath = (field, formData) ->
return unless field?.box? and field?.path?
tail = formData
for item in field.path.split('.')
tail = tail[item]
return unless tail?
return tail if tail?.box?
distinctColor = (index) ->
colors = [
"#000000", "#FFFF00", "#1CE6FF", "#FF34FF", "#FF4A46", "#008941", "#006FA6", "#A30059",
"#FFDBE5", "#7A4900", "#0000A6", "#63FFAC", "#B79762", "#004D43", "#8FB0FF", "#997D87",
"#5A0007", "#809693", "#FEFFE6", "#1B4400", "#4FC601", "#3B5DFF", "#4A3B53", "#FF2F80",
"#61615A", "#BA0900", "#6B7900", "#00C2A0", "#FFAA92", "#FF90C9", "#B903AA", "#D16100",
"#DDEFFF", "#000035", "#7B4F4B", "#A1C299", "#300018", "#0AA6D8", "#013349", "#00846F",
"#372101", "#FFB500", "#C2FFED", "#A079BF", "#CC0744", "#C0B9B2", "#C2FF99", "#001E09",
"#00489C", "#6F0062", "#0CBD66", "#EEC3FF", "#456D75", "#B77B68", "#7A87A1", "#788D66",
"#885578", "#FAD09F", "#FF8A9A", "#D157A0", "#BEC459", "#456648", "#0086ED", "#886F4C",
"#34362D", "#B4A8BD", "#00A6AA", "#452C2C", "#636375", "#A3C8C9", "#FF913F", "#938A81",
"#575329", "#00FECF", "#B05B6F", "#8CD0FF", "#3B9700", "#04F757", "#C8A1A1", "#1E6E00",
"#7900D7", "#A77500", "#6367A9", "#A05837", "#6B002C", "#772600", "#D790FF", "#9B9700",
"#549E79", "#FFF69F", "#201625", "#72418F", "#BC23FF", "#99ADC0", "#3A2465", "#922329",
"#5B4534", "#FDE8DC", "#404E55", "#0089A3", "#CB7E98", "#A4E804", "#324E72", "#6A3A4C",
"#83AB58", "#001C1E", "#D1F7CE", "#004B28", "#C8D0F6", "#A3A489", "#806C66", "#222800",
"#BF5650", "#E83000", "#66796D", "#DA007C", "#FF1A59", "#8ADBB4", "#1E0200", "#5B4E51",
"#C895C5", "#320033", "#FF6832", "#66E1D3", "#CFCDAC", "#D0AC94", "#7ED379", "#012C58";
]
color = colors[index % colors.length]
return [parseInt(color[1..2], 16), parseInt(color[3..4], 16), parseInt(color[5..6], 16)]
class Form
constructor: (, ) ->
match: (formSchema, cb) =>
formData = {}
# Test if schema to page mapping was provided, estimate otherwise.
if typeof formSchema.schemaToPage is 'function'
schemaToPage = formSchema.schemaToPage
else if formSchema.page?
fallbackScale = [0].width / formSchema.page.width
schemaToPage = estimateTransform formSchema.words, [2], fallbackScale
# Match barcodes invariant to transformation changes.
matchBarcodes formData, formSchema, [1], schemaToPage
# Match text and verify cleanliness of empty text fields.
{anchors} = matchText formData, formSchema, [2], schemaToPage, [0]
= anchors
# Estimate schema to fields transform and match checkboxes.
schemaToFields = estimateTransform formSchema.fields, formData, 1, 1, matchByPath
matchCheckboxes formData, formSchema, [3], [2], schemaToPage, schemaToFields
# Call form validators.
async.forEach formSchema.fields, (field, nextField) ->
if field.formValidator?
field.formValidator formData, nextField
else
nextField()
, (err) ->
return cb err if err?
cb null, formData
return
toImage: =>
resultImage = new dv.Image [0].width * .length, [0].height, 32
imageOffset = (box, index) =>
return {
x: box.x + [0].width * index
y: box.y
width: box.width
height: box.height
}
imageBox = {x: 0, y: 0, width: [0].width, height: [0].height}
for image, index in when image?
resultImage.drawImage(image.toColor(), imageOffset(imageBox, index))
for data, index in [1..] when data?
for boxed, boxedIndex in data
try for candidateBox in boxed.candidate
color = distinctColor boxedIndex
resultImage.drawBox(imageOffset(candidateBox, index), 6, color[0], color[1], color[2], 0.5)
try resultImage.drawBox(imageOffset(boxed.box, index), 2, 0, 0, 255, 0.5)
for anchor, index in ? []
box =
x: anchor.word.box.x + anchor.offset.x
y: anchor.word.box.y + anchor.offset.y
width: anchor.word.box.width
height: anchor.word.box.height
resultImage.drawBox(imageOffset(box, 1), 4, 0, 255, 50, 0.5)
resultImage.drawLine(imageOffset(box, 1), imageOffset(anchor.word.box, 1), 4, 0, 255, 255, 0.5)
return resultImage
toObject: =>
return {
barcodes: [1]
text: [2]
checkboxes: [3]
}
module.exports = class FormReader
constructor: (language = 'eng', = null) ->
= new dv.Tesseract language
.pageSegMode = 'single_block'
.classify_enable_learning = 0
.classify_enable_adaptive_matcher = 0
= new dv.ZXing()
find: =>
data = [null, null, null, null]
images = [, null, null, null]
[data[1], images[1]] = findBarcodes images[0],
[data[2], images[2]] = findText images[1],
[data[3], images[3]] = findCheckboxes images[2]
return new Form data, images