From 504f59e38e57b543ac87f73df53fe3a6ba34e26d Mon Sep 17 00:00:00 2001 From: Michael Williamson Date: Fri, 29 Mar 2013 18:57:02 +0000 Subject: Very basic conversion from docx to HTML --- lib/index.js | 49 +++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 49 insertions(+) create mode 100644 lib/index.js (limited to 'lib/index.js') diff --git a/lib/index.js b/lib/index.js new file mode 100644 index 0000000..11364b3 --- /dev/null +++ b/lib/index.js @@ -0,0 +1,49 @@ +var q = require("q"); +var xmlreader = require("./xmlreader"); + +exports.Converter = Converter; + +function Converter() { +} + +Converter.prototype.convertToHtml = function(docxFile) { + return docxFile.read("word/document.xml").then(function(documentXmlString) { + return parseXml(documentXmlString); + }).then(function(documentXml) { + var body = documentXml.root.first("w:body"); + + var html = body.children.map(elementToHtml).join(""); + + return { + html: html + }; + }); +} + +function parseXml(string) { + return xmlreader.read(string); +} + +function elementToHtml(element) { + if (element.type === "element") { + var handler = elementConverters[element.name]; + if (handler) { + return handler(element); + } else { + return ""; + } + } +} + +var elementConverters = { + "w:p": function(element) { + var innerHtml = element.children.map(elementToHtml).join(""); + return "

" + innerHtml + "

"; + }, + "w:r": function(element) { + return element.children.map(elementToHtml).join(""); + }, + "w:t": function(element) { + return element.children[0].value; + } +}; -- cgit v1.2.3