NodeCode

function parseJSON(json, xThreshold, yThreshold) {
    const text = JSON.parse(json)['textAnnotations'];
    const response = {
        lines: []
    };
    let words = text.slice(1);
    words = words.sort((a, b) => compare(a, b, yThreshold));
    let curLineYCoord = (words[0].boundingPoly.vertices[0].y + words[0].boundingPoly.vertices[2].y) / 2;
    let curXEnd = words[0].boundingPoly.vertices[2].x;
    let curXStart = words[0].boundingPoly.vertices[0].x;
    let line = {
        y: curLineYCoord,
        words: []
    };
    let compledtedWord = words[0].description;
    for (let i = 1; i < words.length; i++) {
        const vertices = words[i].boundingPoly.vertices;
        const curWord = words[i].description;
        const yMid = (vertices[0].y + vertices[2].y) / 2
        const xStart = vertices[0].x;
        if (yMid <= curLineYCoord + yThreshold && yMid >= curLineYCoord - yThreshold) {
            if (xStart <= curXEnd + xThreshold) {
                compledtedWord += curWord;
                curXEnd = vertices[2].x;
            } else {
                line.words.push({
                    word: compledtedWord,
                    x1: curXStart,
                    x2: curXEnd
                });
                curXEnd = vertices[2].x;
                compledtedWord = words[i].description;
                curXStart = vertices[0].x;
            }
        } else {
            line.words.push({
                word: compledtedWord,
                x1: curXStart,
                x2: curXEnd
            });
            response.lines.push(line);
            curLineYCoord = (words[i].boundingPoly.vertices[0].y + words[i].boundingPoly.vertices[2].y) / 2;
            curXStart = vertices[0].x;
            curXEnd = vertices[2].x;
            compledtedWord = words[i].description;
            line = {
                y: curLineYCoord,
                words: []
            };
        }
    }
    console.log(JSON.stringify(response));
}

function compare(a, b, yThreshold) {
    const word1Vertices = a.boundingPoly.vertices, word2Vertices = b.boundingPoly.vertices;
    const x1 = word1Vertices[0].x, y1 = word1Vertices[0].y, x2 = word2Vertices[0].x, y2 = word2Vertices[0].y;
    const yMid1 = (y1 + word1Vertices[2].y) / 2, yMid2 = (y2 + word2Vertices[2].y) / 2;
    if (yMid2 <= yMid1 + yThreshold && yMid2 >= yMid1 - yThreshold) {
        if (x1 > x2) {
            return 1;
        } else if (x1 < x2) {
            return -1;
        } else {
            return 0;
        }
    } else if (y1 > y2) {
        return 1;
    } else if (y1 < y2) {
        return -1;
    } else if (x1 > x2) {
        return 1;
    } else if (x1 < x2) {
        return -1;
    } else {
        return 0;
    }
}

const fs = require('fs');
const file = fs.readFileSync('./chandrika_node_(coordinate_wise)_file.txt', 'UTF-8');
parseJSON(file, 5, 10);