Convert olm to csv

parse-olm.py

import zipfile
from lxml import etree
from email.mime.text import MIMEText
from email.mime.multipart import MIMEMultipart
import pandas as pd
import datetime


def load_attachment(zip, name):
    fh = zip.open(name)
    return fh


def get_id(email):
    tag_id = email.find('.//OPFMessageCopyMessageID')
    if tag_id is None:
        tag_id = email.find('.//OPFMessageCopyExchangeConversationId')
    return tag_id.text.strip()


def get_date(email):
    tag = email.find('.//OPFMessageCopySentTime')
    if tag is None:
        tag = email.find('.//OPFMessageCopyReceivedTime')
    date = datetime.datetime.strptime(tag.text.strip(), '%Y-%m-%dT%H:%M:%S')
    return date.isoformat()


def get_body(email):
    has_html = email.find('.//OPFMessageGetHasHTML')
    # has_rich = email.find('.//OPFMessageGetHasRichText')
    tag_body = email.find('.//OPFMessageCopyBody')
    mime_type = 'text/plain'
    if has_html is not None:
        html = has_html.text.replace('E0', '')
        if html == '1':
            tag_body = email.find('.//OPFMessageCopyHTMLBody')
            mime_type = 'text/html'

    body = None
    if tag_body is not None:
        body = tag_body.get('text')
        if body is not None:
            # There might be no body if it's a calender reply or something.
            # Calendar replies still have subject lines and addressees and stuff
            # though so probably worth keeping.
            body = body.strip().encode('utf-8')

    return {mime_type: body}


def get_attachments(zip, email):
    attachments = []
    tag_attachments = email.find('.//OPFMessageCopyAttachmentList')
    if tag_attachments is not None:
        for attachment in tag_attachments.findall('.//messageAttachment'):
            name = attachment.get('OPFAttachmentName')
            mime_type = attachment.get('OPFAttachmentContentType')
            # extension = attachment.get('OPFAttachmentContentExtension')
            # id = attachment.get('OPFAttachmentContentID')
            file = {
                'file_name': name,
                'mime_type': mime_type
            }
            url = attachment.get('OPFAttachmentURL')
            if url is not None:
                fh = load_attachment(zip, url)
                file['file_path'] = url
                file['file_handle'] = fh
            attachments.append(file)
    return attachments


def get_addresses(email):
    tag_from = email.find('.//OPFMessageCopyFromAddresses')
    tag_sender = email.find('.//OPFMessageCopySenderAddress')
    tag_to = email.find('.//OPFMessageCopyToAddresses')
    tag_cc = email.find('.//OPFMessageCopyCCAddresses')
    tag_bcc = email.find('.//OPFMessageCopyBCCAddresses')

    from_names, from_emails = get_contacts(tag_from)
    sender_names, sender_emails = get_contacts(tag_sender)
    to_names, to_emails = get_contacts(tag_to)
    cc_names, cc_emails = get_contacts(tag_cc)
    bcc_names, bcc_emails = get_contacts(tag_bcc)

    names = to_names + from_names + cc_names + bcc_names + sender_names
    emails = to_emails + from_emails + cc_emails + bcc_emails + sender_emails

    frm = from_emails + sender_emails
    author = from_names + sender_names

    return names, emails, author, frm, to_emails, cc_emails, bcc_emails


def get_contacts(addresses):
    names = []
    emails = []
    if addresses is not None:
        for address in addresses.findall('.//emailAddress'):
            email = address.get('OPFContactEmailAddressAddress')
            if email is not None:
                emails.append(email)
            name = address.get('OPFContactEmailAddressName')
            if name is not None and name != email:
                names.append(name)

            # etype = address.get('OPFContactEmailAddressType')

    return names, emails


def parse_message(zip, name):
    headers = {
        'From': None,
        'To': None,
        'Subject': None,
        'Message-ID': None,
        'CC': None,
        'BCC': None,
        'Date': None,
    }
    body = None
    attachments = []
    names = []
    emails = []
    title = None
    author = None

    doc = None
    fh = zip.open(name)
    try:
        doc = etree.parse(fh)
    except etree.XMLSyntaxError:
        p = etree.XMLParser(huge_tree=True)
        try:
            doc = etree.parse(fh, p)
        except etree.XMLSyntaxError:
            # probably corrupt
            pass

    if doc is None:
        return

    for email in doc.findall('//email'):

        headers['Message-ID'] = get_id(email)
        headers['Date'] = get_date(email)

        tag_subject = email.find('.//OPFMessageCopySubject')
        # OPFMessageCopyThreadTopic
        if tag_subject is not None:
            headers['Subject'] = title = tag_subject.text.strip()

        names, emails, author, frm, to, cc, bcc = get_addresses(email)
        headers['To'] = to
        headers['From'] = frm
        headers['CC'] = cc
        headers['BCC'] = bcc

        body = get_body(email)
        attachments = get_attachments(zip, email)

        return {
            'headers': headers,
            'body': body,
            'attachments': attachments,
            'names': names,
            'emails': emails,
            'title': title,
            'author': author,
            'created_at': headers['Date']
        }


def make_email(headers, body, attachments):
    msg = MIMEMultipart()
    for header in headers.keys():
        if isinstance(headers[header], str):
            msg[header] = headers[header]
        elif isinstance(headers[header], list):
            msg[header] = ', '.join(headers[header])

    if body is not None:
        if 'text/html' in body.keys():
            email = MIMEText(body['text/html'], 'html')
        elif 'text/plain' in body.keys():
            email = MIMEText(body['text/plain'], 'plain')
        msg.attach(email)
    return msg


def format_json(msg):
    headers = msg.pop('headers')
    attachments = [file.get('file_name') for file in msg.pop('attachments')]
    return {**msg, **headers, 'attachments': attachments}


def parse(file_path):
    archive = zipfile.ZipFile(file_path, 'r')
    parsed_messages = list()
    for info in archive.namelist():
        if 'com.microsoft.__Attachments' in info:
            continue
        if 'message_' not in info:
            continue

        parsed_messages.append(format_json(parse_message(archive, info)))
    return parsed_messages


if __name__ == '__main__':
    parsed = parse('export.olm')
    pd.DataFrame(parsed).to_csv('export.csv')