forked from sheetjs/sheetjs
SheetJS
e1f8dbb863
- read BOM, handle UTF16LE-encoded XML - handle namespaces in [Content_Types].xml - parse workbook rels to determine sheet files - numbers OSX boolean support (apparently requires "0" or "1") - XLSX force "General" style to be serialized, omit implied cell type and style - updated SSF to 0.7.0 (h/t @sysarchitect) - updated jszip to 2.2.2 - removed old tests/files path, replaced with test_files - themes written - ignore potential existence of thumbnail when calculating relationship ids
178 lines
6.1 KiB
JavaScript
178 lines
6.1 KiB
JavaScript
/* 18.3 Worksheets */
|
|
function parse_ws_xml(data, opts, rels) {
|
|
if(!data) return data;
|
|
/* 18.3.1.99 worksheet CT_Worksheet */
|
|
var s = {}, mtch;
|
|
|
|
/* 18.3.1.35 dimension CT_SheetDimension ? */
|
|
var ref = data.match(/<dimension ref="([^"]*)"\s*\/>/);
|
|
if(ref && ref.length == 2 && ref[1].indexOf(":") !== -1) s["!ref"] = ref[1];
|
|
|
|
/* 18.3.1.55 mergeCells CT_MergeCells */
|
|
var mergecells = [];
|
|
if(data.match(/<\/mergeCells>/)) {
|
|
var merges = data.match(/<mergeCell ref="([A-Z0-9:]+)"\s*\/>/g);
|
|
mergecells = merges.map(function(range) {
|
|
return decode_range(/<mergeCell ref="([A-Z0-9:]+)"\s*\/>/.exec(range)[1]);
|
|
});
|
|
}
|
|
|
|
var refguess = {s: {r:1000000, c:1000000}, e: {r:0, c:0} };
|
|
var q = ["v","f"];
|
|
var sidx = 0;
|
|
|
|
/* 18.3.1.80 sheetData CT_SheetData ? */
|
|
if((mtch=data.match(/<(?:\w+:)?sheetData>([^\u2603]*)<\/(?:\w+:)?sheetData>/m))) mtch[1].split(/<\/(?:\w+:)?row>/).forEach(function(x) {
|
|
if(x === "" || x.trim() === "") return;
|
|
|
|
/* 18.3.1.73 row CT_Row */
|
|
var row = parsexmltag(x.match(/<(?:\w+:)?row[^>]*>/)[0]);
|
|
if(opts.sheetRows && opts.sheetRows < +row.r) return;
|
|
if(refguess.s.r > row.r - 1) refguess.s.r = row.r - 1;
|
|
if(refguess.e.r < row.r - 1) refguess.e.r = row.r - 1;
|
|
/* 18.3.1.4 c CT_Cell */
|
|
var cells = x.substr(x.indexOf('>')+1).split(/<(?:\w+:)?c /);
|
|
cells.forEach(function(c, idx) { if(c === "" || c.trim() === "") return;
|
|
var cref = c.match(/r=["']([^"']*)["']/);
|
|
c = "<c " + c;
|
|
if(cref && cref.length == 2) idx = decode_cell(cref[1]).c;
|
|
var cell = parsexmltag((c.match(/<c[^>]*>/)||[c])[0]); delete cell[0];
|
|
var d = c.substr(c.indexOf('>')+1);
|
|
var p = {};
|
|
q.forEach(function(f){var x=d.match(matchtag(f));if(x)p[f]=unescapexml(x[1]);});
|
|
/* SCHEMA IS ACTUALLY INCORRECT HERE. IF A CELL HAS NO T, EMIT "" */
|
|
if(cell.t === undefined && p.v === undefined) {
|
|
if(!opts.sheetStubs) return;
|
|
p.t = "str"; p.v = undefined;
|
|
}
|
|
else p.t = (cell.t ? cell.t : "n"); // default is "n" in schema
|
|
if(refguess.s.c > idx) refguess.s.c = idx;
|
|
if(refguess.e.c < idx) refguess.e.c = idx;
|
|
/* 18.18.11 t ST_CellType */
|
|
switch(p.t) {
|
|
case 'n': p.v = parseFloat(p.v); break;
|
|
case 's': {
|
|
sidx = parseInt(p.v, 10);
|
|
p.v = strs[sidx].t;
|
|
p.r = strs[sidx].r;
|
|
if(opts.cellHTML) p.h = strs[sidx].h;
|
|
} break;
|
|
case 'str': if(p.v) p.v = utf8read(p.v); break;
|
|
case 'inlineStr':
|
|
var is = d.match(/<is>(.*)<\/is>/);
|
|
is = is ? parse_si(is[1]) : {t:"",r:""};
|
|
p.t = 'str'; p.v = is.t;
|
|
break; // inline string
|
|
case 'b': if(typeof p.v !== 'boolean') p.v = parsexmlbool(p.v); break;
|
|
case 'd':
|
|
p.v = datenum(p.v);
|
|
p.t = 'n';
|
|
break;
|
|
/* in case of error, stick value in .raw */
|
|
case 'e': p.raw = RBErr[p.v]; break;
|
|
}
|
|
|
|
/* formatting */
|
|
var fmtid = 0;
|
|
if(cell.s && styles.CellXf) {
|
|
var cf = styles.CellXf[cell.s];
|
|
if(cf && cf.numFmtId) fmtid = cf.numFmtId;
|
|
}
|
|
try {
|
|
p.w = SSF.format(fmtid,p.v,_ssfopts);
|
|
if(opts.cellNF) p.z = SSF._table[fmtid];
|
|
} catch(e) { if(opts.WTF) throw e; }
|
|
s[cell.r] = p;
|
|
});
|
|
});
|
|
|
|
/* 18.3.1.48 hyperlinks CT_Hyperlinks */
|
|
if(data.match(/<\/hyperlinks>/)) data.match(/<hyperlink[^>]*\/>/g).forEach(function(h) {
|
|
var val = parsexmltag(h); delete val[0];
|
|
if(!val.ref) return;
|
|
var rel = rels['!id'][val.id];
|
|
if(rel) {
|
|
val.Target = rel.Target;
|
|
if(val.location) val.Target += "#"+val.location;
|
|
val.Rel = rel;
|
|
}
|
|
var rng = decode_range(val.ref);
|
|
for(var R=rng.s.r;R<=rng.e.r;++R) for(var C=rng.s.c;C<=rng.e.c;++C) {
|
|
var addr = encode_cell({c:C,r:R});
|
|
if(!s[addr]) s[addr] = {t:"str",v:undefined};
|
|
s[addr].l = val;
|
|
}
|
|
});
|
|
|
|
if(!s["!ref"] && refguess.e.c >= refguess.s.c && refguess.e.r >= refguess.s.r) s["!ref"] = encode_range(refguess);
|
|
if(opts.sheetRows && s["!ref"]) {
|
|
var tmpref = decode_range(s["!ref"]);
|
|
if(opts.sheetRows < +tmpref.e.r) {
|
|
tmpref.e.r = opts.sheetRows - 1;
|
|
if(tmpref.e.r > refguess.e.r) tmpref.e.r = refguess.e.r;
|
|
if(tmpref.e.r < tmpref.s.r) tmpref.s.r = tmpref.e.r;
|
|
if(tmpref.e.c > refguess.e.c) tmpref.e.c = refguess.e.c;
|
|
if(tmpref.e.c < tmpref.s.c) tmpref.s.c = tmpref.e.c;
|
|
s["!fullref"] = s["!ref"];
|
|
s["!ref"] = encode_range(tmpref);
|
|
}
|
|
}
|
|
if(mergecells.length > 0) s["!merges"] = mergecells;
|
|
return s;
|
|
}
|
|
|
|
var WS_XML_ROOT = writextag('worksheet', null, {
|
|
'xmlns': XMLNS.main[0],
|
|
'xmlns:r': XMLNS.r
|
|
});
|
|
|
|
var write_ws_xml_cell = function(cell, ref, ws, opts, idx, wb) {
|
|
var vv = cell.v; if(cell.t == 'b') vv = cell.v ? "1" : "0";
|
|
var v = writextag('v', escapexml(String(vv))), o = {r:ref};
|
|
o.s = get_cell_style(opts.cellXfs, cell, opts);
|
|
if(o.s === 0 || o.s === "0") delete o.s;
|
|
/* TODO: cell style */
|
|
if(typeof cell.v === 'undefined') return "";
|
|
switch(cell.t) {
|
|
case 's': case 'str': {
|
|
if(opts.bookSST) {
|
|
v = writextag('v', String(get_sst_id(opts.Strings, cell.v)));
|
|
o.t = "s"; return writextag('c', v, o);
|
|
} else { o.t = "str"; return writextag('c', v, o); }
|
|
} break;
|
|
case 'n': delete o.t; return writextag('c', v, o);
|
|
case 'b': o.t = "b"; return writextag('c', v, o);
|
|
case 'e': o.t = "e"; return writextag('c', v, o);
|
|
}
|
|
};
|
|
|
|
var write_ws_xml_data = function(ws, opts, idx, wb) {
|
|
var o = [], r = [], range = utils.decode_range(ws['!ref']), cell, ref;
|
|
for(var R = range.s.r; R <= range.e.r; ++R) {
|
|
r = [];
|
|
for(var C = range.s.c; C <= range.e.c; ++C) {
|
|
ref = utils.encode_cell({c:C, r:R});
|
|
if(!ws[ref]) continue;
|
|
if((cell = write_ws_xml_cell(ws[ref], ref, ws, opts, idx, wb))) r.push(cell);
|
|
}
|
|
if(r.length) o.push(writextag('row', r.join(""), {r:encode_row(R)}));
|
|
}
|
|
return o.join("");
|
|
};
|
|
|
|
var write_ws_xml = function(idx, opts, wb) {
|
|
var o = [], s = wb.SheetNames[idx], ws = wb.Sheets[s] || {}, sidx = 0, rdata = "";
|
|
o.push(XML_HEADER);
|
|
o.push(WS_XML_ROOT);
|
|
o.push(writextag('dimension', null, {'ref': ws['!ref'] || 'A1'}));
|
|
|
|
sidx = o.length;
|
|
o.push(writextag('sheetData', null));
|
|
if(ws['!ref']) rdata = write_ws_xml_data(ws, opts, idx, wb);
|
|
if(rdata.length) o.push(rdata);
|
|
if(o.length>sidx+1){ o.push('</sheetData>'); o[sidx]=o[sidx].replace("/>",">"); }
|
|
|
|
if(o.length>2){ o.push('</worksheet>'); o[1]=o[1].replace("/>",">"); }
|
|
return o.join("");
|
|
};
|