PST Archive to RFC822 (*.eml) script

README.md

PST Archive to RFC822

This script extracts all emails from an Outlook PST archive and saves them into some output folder as individual RFC822 compliant *.eml files.

Installing the external dependency pypff may not be straight forward (it wasn't for me). I forked the original repository to make it work in Python 3. If you get errors, check their wiki pages for help or try my fork. Below are the steps that worked for me:

Clone https://github.com/libyal/libpff/tree/master/pypff

cd libpff/
./synclibs.sh
./autogen.sh
./configure --enable-python
make
sudo make install
python setup.py build
sudo python setup.py install

Now that everything is installed, you can execute the script as follows: python pst2eml.py /path/to/archive.pst /path/to/output/dir

Optionally, you can write the log into a file by adding --logfile=/path/to/log_dir to the command.

Full disclaimer: I was inspired by this script, but as you may see, I pretty much threw everything overboard and made my own thing. Only kept the logging and argparse really.

pst2eml.py

import os
import argparse
import logging
import re
import pypff


def process_folder(folder, path):
    folder_path = path + '/' + (folder.name or 'root')
    n_msg = folder.number_of_sub_messages
    logging.debug('Processing folder "{}" with {} sub-folders and {} messages; full path: "{}"'.format(
        folder.name, folder.number_of_sub_folders, n_msg, folder_path))

    safe_path = re.sub(r'[ /]', '_', re.sub(r'[^a-z0-9 /]', '', folder_path.lower()))
    for mi, message in enumerate(folder.sub_messages):
        logging.debug('{}/{} > Processing message by {} with subject: {}'.format(mi, n_msg,
                                                                                 message.sender_name, message.subject))
        msg = process_message(message)
        fname = os.path.join(output_directory, safe_path + '_' + str(mi) + '.eml')
        logging.debug('  -- saving as {}'.format(fname))
        with open(fname, 'w') as f:
            f.write(msg)

    for sub_folder in folder.sub_folders:
        process_folder(sub_folder, folder_path)


def get_body(msg):
    def prep(b):
        if type(b) == bytes:
            b = b.decode("utf-8")
        return b.strip() if b else None

    body = prep(msg.plain_text_body)
    if body:
        return 'plain-text', body

    body = prep(msg.html_body)
    if body:
        return 'html', body

    body = prep(msg.rtf_body)
    if body:
        return 'rtf', body

    return 'plain-text', ''


def process_message(message):
    msg = ''
    keys = []
    for hp in message.transport_headers.split('\n'):
        pts = re.findall(r'^([^:]+): (.+)\r$', hp)
        if pts:
            key = pts[0][0].capitalize()
            if key in keys:
                key = 'X-' + key
            keys.append(key)
            val = pts[0][1]

            if key == 'Date':
                val = ', '.join(val.split(',')[:2])

            msg += key + ': ' + val + '\r\n'

    msg += 'X-Sender-Name: ' + message.sender_name + '\r\n'
    msg += 'X-Delivery-Time: ' + str(message.delivery_time) + '\r\n'
    msg += 'X-Creation-Time: ' + str(message.creation_time) + '\r\n'
    msg += 'X-Client-Submit-Time: ' + str(message.client_submit_time) + '\r\n'
    msg += 'X-Subject: ' + message.subject + '\r\n'
    msg += 'X-Attachments: ' + str(message.number_of_attachments) + '\r\n'

    btype, body = get_body(message)
    msg += 'X-Body-Type: ' + btype + '\r\n'

    msg += '\r\n'
    msg += body
    return msg


if __name__ == "__main__":
    parser = argparse.ArgumentParser()
    parser.add_argument('PST_FILE', help="PST File Format from Microsoft Outlook")
    parser.add_argument('OUTPUT_DIR', help="Directory of output for temporary and report files.")
    parser.add_argument('--logfile', default=None, help='File path of log file.')
    args = parser.parse_args()

    output_directory = os.path.abspath(args.OUTPUT_DIR)

    if not os.path.exists(output_directory):
        os.makedirs(output_directory)

    if args.logfile:
        if not os.path.exists(args.logfile):
            os.makedirs(args.logfile)
        log_path = os.path.join(args.logfile, 'pst_indexer.log')
    else:
        log_path = None
    logging.basicConfig(level=logging.DEBUG, filename=log_path,
                        format='%(asctime)s | %(levelname)s | %(message)s', filemode='w')

    logging.info('Starting Script...')
    pst_file = args.PST_FILE
    try:
        pff_file = pypff.file()
        pff_file.open(pst_file)

        process_folder(pff_file.root_folder, os.path.basename(pst_file))
    except Exception as e:
        raise
    finally:
        pff_file.close()

    logging.info('Script Complete')

Hi,
I'm getting this error:

  File "pst2eml.py", line 110, in <module>
    process_folder(pff_file.root_folder, os.path.basename(pst_file))
  File "pst2eml.py", line 27, in process_folder
    process_folder(sub_folder, folder_path)
  File "pst2eml.py", line 27, in process_folder
    process_folder(sub_folder, folder_path)
  File "pst2eml.py", line 20, in process_folder
    msg = process_message(message)
  File "pst2eml.py", line 75, in process_message
    btype, body = get_body(message)
  File "pst2eml.py", line 36, in get_body
    body = prep(msg.plain_text_body)
  File "pst2eml.py", line 34, in prep
    return b.strip()
AttributeError: 'NoneType' object has no attribute 'strip'

Hi @thenktor, thanks for the hint. Apparently there is an email that can't be parsed or has no message text. A quick workaround would be to return an empty string. This way you at least get something, but it might not be complete due to an error in pypff.
I updated the script above.

Hi @TimRepke, thanks for the answer. Now I have a lot of encoding problems: UnicodeDecodeError: 'utf8' codec can't decode byte 0xe4 in position 2164: invalid continuation byte
It turns out, that the mails in the PST file have different encodings. So far, this is my fix in the get_body function:

def get_body(msg):
    def prep(b):
        if type(b) == bytes:
            try:
                b = b.decode("utf8")
            except UnicodeDecodeError:
                logging.info('Is not utf8...')
            else:
                logging.info('Is utf8...')
                return b.strip() if b else None
            try:
                b = b.decode("cp1252")
            except UnicodeDecodeError:
                logging.info('Is not cp1252...')
                logging.error('Could not decode body!')
                return None
            else:
                logging.info('Is cp1252...')
                return b.strip() if b else None
        return b.strip() if b else None

    body = prep(msg.plain_text_body)
    if body:
        return 'plain-text', body

    try:
        body = prep(msg.html_body)
    except OSError:
        logging.error('A PyPff error!')
        return 'plain-text', ''

    if body:
        return 'html', body

    body = prep(msg.rtf_body)
    if body:
        return 'rtf', body

    return 'plain-text', ''

Next error: Some mails are missing specific headers, e.g. an Exchange appointment invitation mail apparantly misses message.transport_headers. Or an e-mail draft was missing message.sender_name.
Also there are a lot of item, that are missing a message body (vCards and stuff).
Therefore I've changed the process_message function as follows:

def process_message(message):
    msg = ''
    keys = []
    if message.transport_headers:
        for hp in message.transport_headers.split('\n'):
            pts = re.findall(r'^([^:]+): (.+)\r$', hp)
            if pts:
                key = pts[0][0].capitalize()
                if key in keys:
                    key = 'X-' + key
                keys.append(key)
                val = pts[0][1]
    
                if key == 'Date':
                    val = ', '.join(val.split(',')[:2])
    
                msg += key + ': ' + val + '\r\n'
    if message.sender_name:
        msg += 'X-Sender-Name: ' + message.sender_name + '\r\n'
    if message.delivery_time:
        msg += 'X-Delivery-Time: ' + str(message.delivery_time) + '\r\n'
    if message.creation_time:
        msg += 'X-Creation-Time: ' + str(message.creation_time) + '\r\n'
    if message.client_submit_time:
        msg += 'X-Client-Submit-Time: ' + str(message.client_submit_time) + '\r\n'
    if message.subject:
        msg += 'X-Subject: ' + message.subject + '\r\n'
    if message.number_of_attachments:
        msg += 'X-Attachments: ' + str(message.number_of_attachments) + '\r\n'

    btype, body = get_body(message)
    msg += 'X-Body-Type: ' + btype + '\r\n'

    msg += '\r\n'
    if body:
        msg += body
    else:
        logging.debug('Missing message body!')
        return None
    
    return msg

In process_folder function I've changed as follows to discard non mail items without body:

        msg = process_message(message)
        if msg:
            fname = os.path.join(output_directory, safe_path + '_' + str(mi) + '.eml')
            logging.debug('  -- saving as {}'.format(fname))
            with open(fname, 'w') as f:
                f.write(msg)

Now the script is working on my test.pst file, but the attachments are missing from the generated *.eml files.

In case anyone else ends up here, see libyal/libpff#2 to check the status of python bindings for attachments.