rainyrhy

#!/usr/bin/python
#
# ------------------------------------------------------------------------
#   quick hack for importing google notebook notes onto evernote.
#   some parts are from samples from my test export.
#   written by rainyrhy on 23 Oct 2008
# ---------------
#   as for now, it opens read.xml in the current directory and outputs as
#   write.enex
# ---------------
#   this script works successfully most of the time.
#   known problems:
#       - exporting - source url in title
#       - evernote can't sync some notes upstream, said that content
#         data is incorrect
#       - tag not supported.
# ------------------------------------------------------------------------
from xml import sax
from datetime import datetime
import pprint

class ParseGoogleReader(sax.handler.ContentHandler):
    def startDocument(self):
        self.notelist = []
        self.curnote = None

        self.receive = None # characters receive status, enumuration of the 3 below
        self.recv_date = 'date'
        self.recv_title = 'title'
        self.recv_content = 'content'

    def startElement(self, name, attrs):
        self.receive = None # in case it skipped characters callback
                            # cause there is none

        if name == 'entry':
            print 'new note'
            self.curnote = {}

        if self.curnote is None: return

        if name == 'updated':
            print 'taking date'
            self.receive = self.recv_date

        if name == 'title':
            print 'taking title'
            self.receive = self.recv_title

        if name == 'content':
            print 'taking content'
            self.receive = self.recv_content

        if name == 'link':
            content = attrs.getValue('href')
            print 'link', content
            self.curnote['link'] = content

    def endElement(self, name):
        if name == 'entry':
            print 'end note'
            self.notelist.append(self.curnote)
            self.curnote = None;

    def characters(self, content):
        if self.receive is self.recv_date:
            print 'receiving date', content
            content = content[ : content.rfind('.')]
            content = content.replace('T', ' ')
            date = datetime.strptime(content, "%Y-%m-%d %H:%M:%S")
            self.curnote['date'] = date

        if self.receive is self.recv_title:
            try:
                print 'receiving title', content
            except UnicodeEncodeError:
                pass

            self.curnote['title'] = content

        if self.receive is self.recv_content:
            print 'receiving content', len(content)
            if 'content' in self.curnote:
                self.curnote['content'] += content
            else:
                self.curnote['content'] = content

# -- some strings taken from sample. -----
export_header = '''<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE en-export SYSTEM "http://xml.evernote.com/pub/evernote-export.dtd">
<en-export export-date="20081011T080456Z" application="Evernote" version="Mac 1.1.5 (36338)">
'''

export_close = '''</en-export>'''

export_content_start = '''<content><![CDATA[<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE en-note SYSTEM "http://xml.evernote.com/pub/enml.dtd">
<en-note>'''

export_content_end = '''</en-note>]]></content>'''

# string if there's no link. make it look like a manual created note.
export_deflink = '''<note-attributes><subject-date>19700101T000000Z</subject-date><source-application>Evernote Web; 36564; Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.0.3) Gecko/2008092417 Firefox/3.0.3</source-application></note-attributes>'''

# string if there's a link. make it look like a web clip.
export_link_start = '''<note-attributes><subject-date>19700101T000000Z</subject-date><source>web.clip</source><source-url>'''
export_link_end = '''</source-url></note-attributes>'''

def writenode(writeout, nodename, data):
    ''' write data in the nodename as tagname to writeout '''
    writeout.write( '<%s>' % nodename )
    writeout.write(data)
    writeout.write( '</%s>' % nodename )

def writenote(writeout, note):
    ''' func for 1 note'''
    writeout.write('<note>')

    if 'title' not in note:
        note['title'] = ''
    if 'content' not in note:
        note['content'] = ''

    title = note['title'].encode('utf-8')
    content = note['content'].encode('utf-8')

    writenode(writeout,'title', title)

    writeout.write(export_content_start)
    writeout.write(content)
    writeout.write(export_content_end)

    # weird way of date string from evernote.
    datefoo = note['date'].strftime('%Y%m%d')
    datebar = note['date'].strftime('%H%M%S')
    datestr = datefoo + 'T' + datebar + 'Z'

    writenode(writeout, 'created', datestr)
    writenode(writeout, 'updated', datestr)

    if 'link' in note:
        writeout.write(export_link_start)
        link = sax.saxutils.escape(note['link'])
        writeout.write(link)
        writeout.write(export_link_end)
    else:
        writeout.write(export_deflink)

    writeout.write('</note>')

def main():
    read = ParseGoogleReader();
    sax.parse('read.xml', read)

    print len(read.notelist)
    pp = pprint.PrettyPrinter(width=80)
    #pp.pprint(read.notelist)

    writeout = open('write.enex','w')
    writeout.write(export_header)
    for note in read.notelist:
        writenote(writeout,note)
    writeout.write(export_close)
    writeout.flush()
    writeout.close()

if __name__ == "__main__":
    main()