Add support for QFX files with bodies already in XML format (not need…

…ing SGML mangling)

ofx.js

@@ -1,5 +1,18 @@
 var xml2json = require('xml2json');
 
+function sgml2Xml(sgml) {
+    return sgml
+        .replace(/>\s+</g, '><')    // remove whitespace inbetween tag close/open
+        .replace(/\s+</g, '<')      // remove whitespace before a close tag
+        .replace(/>\s+/g, '>')      // remove whitespace after a close tag
+        .replace(/<([A-Z0-9_]*)+\.+([A-Z0-9_]*)>([^<]+)/g, '<\$1\$2>\$3' )
+        .replace(/<(\w+?)>([^<]+)/g, '<\$1>\$2</\$1>');
+}
+
+function parseXml(content) {
+    return JSON.parse(xml2json.toJson(content, { coerce: false }))
+}
+
 function parse(data) {
     // firstly, split into the header attributes and the footer sgml
     var ofx = data.split('<OFX>', 2);
@@ -13,16 +26,16 @@ function parse(data) {
     });
 
     // make the SGML and the XML
-    var sgml = '<OFX>' + ofx[1];
-    var xml = sgml
-        .replace(/>\s+</g, '><')
-        .replace(/\s+</g, '<')
-        .replace(/>\s+/g, '>')
-        .replace(/<([A-Z0-9_]*)+\.+([A-Z0-9_]*)>([^<]+)/g, '<\$1\$2>\$3' )
-        .replace(/<(\w+?)>([^<]+)/g, '<\$1>\$2</\$1>');
+    var content = '<OFX>' + ofx[1];
 
-    // parse the XML
-    var data = JSON.parse(xml2json.toJson(xml, { coerce: false }));
+    // Parse the XML/SGML portion of the file into an object
+    // Try as XML first, and if that fails do the SGML->XML mangling
+    var data = null;
+    try {
+        data = parseXml(content);
+    } catch (e) {
+        data = parseXml(sgml2Xml(content));
+    }
 
     // put the headers into the returned data
     data.header = header;

test/data/example-xml.qfx

@@ -0,0 +1,8 @@
+OFXHEADER:100
+DATA:OFXSGML
+VERSION:102
+SECURITY:NONE
+ENCODING:USASCII
+CHARSET:1252
+COMPRESSION:NONE
+<OFX><SIGNONMSGSRSV1><SONRS><STATUS><CODE>0</CODE><SEVERITY>INFO</SEVERITY></STATUS><DTSERVER>20170104225053</DTSERVER><LANGUAGE>ENG</LANGUAGE><FI><ORG>Ally</ORG><FID>6157</FID></FI><INTU.BID>6157</INTU.BID><INTU.USERID>XXXXXXXXXX</INTU.USERID></SONRS></SIGNONMSGSRSV1><BANKMSGSRSV1><STMTTRNRS><TRNUID>0</TRNUID><STATUS><CODE>0</CODE><SEVERITY>Info</SEVERITY></STATUS><STMTRS><CURDEF>USD</CURDEF><BANKACCTFROM><BANKID>XXXXXXXXX</BANKID><ACCTID>XXXXXXXXXX</ACCTID><ACCTTYPE>SAVINGS</ACCTTYPE></BANKACCTFROM><BANKTRANLIST><DTSTART>20161204000000</DTSTART><DTEND>20170104000000</DTEND><STMTTRN><TRNTYPE>INT</TRNTYPE><DTPOSTED>20161215000550</DTPOSTED><TRNAMT>0.84</TRNAMT><FITID>21172131531687</FITID><NAME>Interest Paid</NAME></STMTTRN></BANKTRANLIST><LEDGERBAL><BALAMT>1025.23</BALAMT><DTASOF>20170104225053</DTASOF></LEDGERBAL><AVAILBAL><BALAMT>1025.23</BALAMT><DTASOF>20170104225053</DTASOF></AVAILBAL></STMTRS></STMTTRNRS></BANKMSGSRSV1></OFX>

test/parse.js

@@ -24,6 +24,30 @@ test('parse', function (t) {
     t.end();
 });
 
+test('parse XML', function (t) {
+    var file = fs.readFileSync(__dirname + '/data/example-xml.qfx', 'utf8');
+    var data = ofx.parse(file);
+
+    // headers
+    t.equal(data.header.OFXHEADER, '100');
+    t.equal(data.header.ENCODING, 'USASCII');
+
+    var transaction = data.OFX.BANKMSGSRSV1.STMTTRNRS.STMTRS.BANKTRANLIST.STMTTRN;
+    t.same(transaction, {
+        TRNTYPE: 'INT',
+        DTPOSTED: '20161215000550',
+        TRNAMT: '0.84',
+        FITID: '21172131531687',
+        NAME: 'Interest Paid'
+    });
+
+    var status = data.OFX.SIGNONMSGSRSV1.SONRS.STATUS;
+    t.equal(status.CODE, '0');
+    t.equal(status.SEVERITY, 'INFO');
+
+    t.end();
+})
+
 // parse <-> serialize
 test('serialize', function(t) {
     var expected = fs.readFileSync(__dirname + '/data/example1.ofx', 'utf8');