Comparing two RegEx objects in Node.js

Question

I'm using NodeRED to perform some logic on a string which has been created from image analysis (OCR) on Microsoft Azure Cognitive Services. The image analysis doesn't allow for any pattern matching / input pattern.

The resulting string (let's call it 'A') sometimes interprets characters slightly incorrectly, typical things like 'l' = '1' or 's' = '5'.

The resulting string can be one of only a few different formats, for argument sake lets say:

1. [az]{4,5}
2. [ag]{3}[0-9]{1,2}
3. [0-9][az]{4}

What I need to do is determine which format the intepretted string ('A') most closely aligns to ('1', '2' or '3'). Once I establish this, I was planning to adjust the misinterpretted characters and hopefully be left with a string that is (near) perfect.

My initial plan was to convert 'A' into RegEx - so if 'A' came back as "12345", I would change this to a RegEx object [1|l][2|z]34[5|s], compare this object to the RegEx objects and hopefully one would come back as a match.

In reality, the interpretted string is more like 8 alphanumeric and five different (fairly complex) RegEx possibilities, but I've tried to simplify the problem for the purposes of this question.

So the question: is it possible to compare RegEx in this way? Does anyone have any other suggestions on how this image analysis could be improved?

Thanks

Answer 1

Here is a solution using a Cartesian product to compare a string for possible matches. Test string is 'abclz' , which could match pattern1 or pattern2 :

 const cartesian = (...a) => a.reduce((a, b) => a.flatMap(d => b.map(e => [d, e].flat()))); const charMapping = { '1': ['1','l'], 'l': ['1','l'], '2': ['2','z'], 'z': ['2','z'], '5': ['5','s'], 's': ['5','s'] }; const buckets = { pattern1: /^[az]{4,5}$/, pattern2: /^[ag]{3}[0-9]{1,2}$/, pattern3: /^[0-9][az]{4}$/ }; const input = 'abclz'; console.log('input:', input); let params = input.split('').map(c => charMapping[c] || [c]); let toCompare = cartesian(...params).map(arr => arr.join('')); console.log('toCompare:', toCompare); let potentialMatches = toCompare.flatMap(str => { return Object.keys(buckets).map(pattern => { let match = buckets[pattern].test(str); console.log(str, pattern + ':', match); return match? str: null; }).filter(Boolean); }); console.log('potentialMatches:', potentialMatches);

Output:

input: abclz
toCompare: [
  "abc12",
  "abc1z",
  "abcl2",
  "abclz"
]
abc12 pattern1: false
abc12 pattern2: true
abc12 pattern3: false
abc1z pattern1: false
abc1z pattern2: false
abc1z pattern3: false
abcl2 pattern1: false
abcl2 pattern2: false
abcl2 pattern3: false
abclz pattern1: true
abclz pattern2: false
abclz pattern3: false
potentialMatches: [
  "abc12",
  "abclz"
]