Last active
May 16, 2017 03:39
-
-
Save shinaisan/8fe33f6d179a3deef6114c2f8f1b483d to your computer and use it in GitHub Desktop.
Node.js xmldom whitespace stripping sample
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
node_modules | |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
<table> | |
<tr><td>Alpha<!-- COMMENT --></td><td>Α</td><td>U+0391</td></tr> | |
<tr><td>Beta</td><td>Β</td><td>U+0392</td></tr> | |
<tr><td>Gamma</td><td>Γ</td><td>U+0393</td></tr> | |
<tr><td>Delta</td><td>Δ</td><td>U+0394</td></tr> | |
<tr><td>Epsilon</td><td>Ε</td><td>U+0395</td></tr> | |
<tr><td>Zeta</td><td>Ζ</td><td>U+0396</td></tr> | |
<tr><td>Eta</td><td>Η</td><td>U+0397</td></tr> | |
<tr><td>Theta</td><td>Θ</td><td>U+0398</td></tr> | |
<tr><td>Iota</td><td>Ι</td><td>U+0399</td></tr> | |
<tr><td>Kappa</td><td>Κ</td><td>U+039A</td></tr> | |
<tr><td>Lambda</td><td>Λ</td><td>U+039B</td></tr> | |
<tr><td>Mu</td><td>Μ</td><td>U+039C</td></tr> | |
<tr><td>Nu</td><td>Ν</td><td>U+039D</td></tr> | |
<tr><td>Xi</td><td>Ξ</td><td>U+039E</td></tr> | |
<tr><td>Omicron</td><td>Ο</td><td>U+039F</td></tr> | |
<tr><td>Pi</td><td>Π</td><td>U+03A0</td></tr> | |
<tr><td>Rho</td><td>Ρ</td><td>U+03A1</td></tr> | |
<tr><td>Sigma</td><td>Σ</td><td>U+03A3</td></tr> | |
<tr><td>Tau</td><td>Τ</td><td>U+03A4</td></tr> | |
<tr><td>Upsilon</td><td>Υ</td><td>U+03A5</td></tr> | |
<tr><td>Phi</td><td>Φ</td><td>U+03A6</td></tr> | |
<tr><td>Chi</td><td>Χ</td><td>U+03A7</td></tr> | |
<tr><td>Psi</td><td>Ψ</td><td>U+03A8</td></tr> | |
<tr><td>Omega</td><td>Ω</td><td>U+03A9</td></tr> | |
<tr><td>alpha</td><td>α</td><td>U+03B1</td></tr> | |
<tr><td>beta</td><td>β</td><td>U+03B2</td></tr> | |
<tr><td>gamma</td><td>γ</td><td>U+03B3</td></tr> | |
<tr><td>delta</td><td>δ</td><td>U+03B4</td></tr> | |
<tr><td>epsilon</td><td>ε</td><td>U+03B5</td></tr> | |
<tr><td>zeta</td><td>ζ</td><td>U+03B6</td></tr> | |
<tr><td>eta</td><td>η</td><td>U+03B7</td></tr> | |
<tr><td>theta</td><td>θ</td><td>U+03B8</td></tr> | |
<tr><td>iota</td><td>ι</td><td>U+03B9</td></tr> | |
<tr><td>kappa</td><td>κ</td><td>U+03BA</td></tr> | |
<tr><td>lambda</td><td>λ</td><td>U+03BB</td></tr> | |
<tr><td>mu</td><td>μ</td><td>U+03BC</td></tr> | |
<tr><td>nu</td><td>ν</td><td>U+03BD</td></tr> | |
<tr><td>xi</td><td>ξ</td><td>U+03BE</td></tr> | |
<tr><td>omicron</td><td>ο</td><td>U+03BF</td></tr> | |
<tr><td>pi</td><td>π</td><td>U+03C0</td></tr> | |
<tr><td>rho</td><td>ρ</td><td>U+03C1</td></tr> | |
<tr><td>sigmaf</td><td>ς</td><td>U+03C2</td></tr> | |
<tr><td>sigma</td><td>σ</td><td>U+03C3</td></tr> | |
<tr><td>tau</td><td>τ</td><td>U+03C4</td></tr> | |
<tr><td>upsilon</td><td>υ</td><td>U+03C5</td></tr> | |
<tr><td>phi</td><td>φ</td><td>U+03C6</td></tr> | |
<tr><td>chi</td><td>χ</td><td>U+03C7</td></tr> | |
<tr><td>psi</td><td>ψ</td><td>U+03C8</td></tr> | |
<tr><td>omega</td><td>ω</td><td>U+03C9</td></tr> | |
<tr><td>thetasym</td><td>ϑ</td><td>U+03D1</td></tr> | |
<tr><td>upsih</td><td>ϒ</td><td>U+03D2</td></tr> | |
<tr><td>piv</td><td>ϖ</td><td>U+03D6</td></tr> | |
</table> | |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
var xml = require('xmldom'); | |
var fs = require('fs'); | |
function main(outputType, fileName) { | |
var s = fs.readFileSync(fileName).toString(); | |
var parser = new xml.DOMParser(); | |
var ser = new xml.XMLSerializer(); | |
var dom = parser.parseFromString(s); | |
wsstrip(dom.firstChild); | |
if (outputType === 'json') { | |
var j = toJson(dom.firstChild); | |
console.log(JSON.stringify(j, null, 2)); | |
} else { | |
var t = ser.serializeToString(dom); | |
console.log(t); | |
} | |
} | |
function wsstrip(node) { | |
var wsre = /[ \t\n]+/; | |
if (node.nodeType == node.TEXT_NODE) { | |
if (wsre.test(node.nodeValue)) { | |
return null; | |
} | |
} | |
if (node.childNodes) { | |
var wsnodes = []; | |
for (var i = 0; i < node.childNodes.length; i++) { | |
var ch = node.childNodes[i] | |
var c = wsstrip(ch); | |
if (!c) { | |
wsnodes.push(ch); | |
} | |
} | |
for (var i = 0; i < wsnodes.length; i++) { | |
node.removeChild(wsnodes[i]); | |
} | |
} | |
return node; | |
} | |
function toJson(node) { | |
var children = []; | |
if (node.childNodes) { | |
for (var i = 0; i < node.childNodes.length; i++) { | |
var ch = node.childNodes[i]; | |
children.push(toJson(ch)); | |
} | |
} | |
return { | |
tagName: node.tagName, | |
nodeName: node.nodeName, | |
nodeType: node.nodeType, | |
nodeValue: node.nodeValue, | |
data: node.data, | |
children: children | |
}; | |
} | |
if (require.main === module) { | |
main(process.argv[2], process.argv[3]); | |
} | |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment