MMS participants' numbers not parsed correctly

[mwhdc]

Hi, thanks for publishing this very useful script. I used it to convert a large (~15K messages) Textra database and noticed that in convos with 2+ participants, messages are not always attributed to the correct address, but to a single address per convo or a concatenated string of all participating addresses ("+1aaabbbcccc+1dddeeeffff").

I would suggest:

• When eliminating '^' in convos.lookup_key, first replace instances of '^^' with a delimiter indicating separate addresses, then replace remaining '^' with ''
• For convos with multiple addresses in lookup_key, assign individual messages to message.originator

Let me know what you think - apologies for not just submitting a PR, I'm rusty in Python.

[biqqles]

Hi. I'm glad you found the script useful. When writing it I gave no consideration to MMS messages (whether text or multimedia) simply because they're rather costly on my network and since no one I know really uses them either they weren't present in my database. I probably didn't make that clear enough in the documentation.

However, I'm happy to improve the script so that it handles MMS texts properly. Could you provide me with a (redacted, obviously) sample of SMS Backup and Restore's output for a conversation with multiple participants so I'm clear on what the correct behaviour should be? And am I correct in thinking that your first bullet point means that for these conversations convos.lookup_key looks like ^first^^second^ and so on?