Telegram webhook receiver

[Tijani-Dia]

PR to add Telegram webhook receiver.

Current limitations
When a message with multiple images is edited on Telegram, we only receive the edited message and the first photo of the message. This means that the other images of the message will be removed from the corresponding live post.

See #80 to find how to setup TelegramWebhookReceiver.

[codecov-commenter]

Codecov Report

Merging #85 (6b61833) into main (f43c0c1) will increase coverage by 0.68%.
The diff coverage is 99.32%.

@@            Coverage Diff             @@
##             main      #85      +/-   ##
==========================================
+ Coverage   98.11%   98.79%   +0.68%     
==========================================
  Files          17       20       +3     
  Lines         689      833     +144     
  Branches       61       81      +20     
==========================================
+ Hits          676      823     +147     
+ Misses          8        5       -3     
  Partials        5        5              

Impacted Files	Coverage Δ
src/wagtail_live/receivers.py	98.47% <50.00%> (+2.20%)	⬆️
src/wagtail_live/adapters/telegram/__init__.py	100.00% <100.00%> (ø)
src/wagtail_live/adapters/telegram/receiver.py	100.00% <100.00%> (ø)
src/wagtail_live/adapters/telegram/utils.py	100.00% <100.00%> (ø)
src/wagtail_live/utils.py	95.83% <100.00%> (+0.18%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update f43c0c1...6b61833. Read the comment docs.

[lucasmoeskops]

Great work! Lots of functionality :-) I tested it and it works very well!

I didn't really understand what was going on in the process_text method, so I looked at it and made a proposal to try and make it simpler / more readable.

[lucasmoeskops]

I think this manual algorithm is a bit too complicated which makes it hard to understand what happens. After some thinking I concluded that no extra variables are needed if the entities are process backwards, because the text will then only be edited at locations which are no longer matches by indices in the entity objects.

I propose the following algorithm:

# Process the entities in reversed order to be able to edit the text in place.
for entity in reversed(entities):
    url = ""
    start = entity['offset']
    end = start + entity['length']

    if entity["type"] == "url":
        url = description = text[start:end]

        if is_embed(url):
            # Check if this can match an embed block, if so no conversion happens.
            # It matches an embed block if it has a line in the text for itself.
            if end == len_text or text[end] == "\n":
                if start == 0 or text[start - 1] == "\n":
                    # This is an embed block, skip to the next entity
                    continue

    if entity["type"] == "text_link":
        url = entity["url"]
        description = text[start:end]

    if url:
        link = f'<a href="{url}">{description}</a>'
        text = text[:start] + link + text[end:]

[Tijani-Dia]

Very nice algorithm. Thanks!

[lucasmoeskops]

Unfortunate that Telegram doesn't send an update when the user deletes a message. Would it be nice to think of a work-around for this, or would this be out of scope?

I'm thinking of some string that would mark the message as deleted if the user puts it at the start, for example by editing the message "Hello world!" to become "DELETED Hello world!".

(It creates a bit of a weird case when the user then removes the "DELETED " part again then, so that makes it a bit harder.)

[Tijani-Dia]

Thanks for your nice review @lucasmoeskops .

Yes, indeed it's unfortunate that they don't send updates when a message is deleted, I'm not sure but I think it's due to the fact that after a certain amount of time you can no longer delete messages in Telegram (Again I'm not 100% sure).

It would be really nice to have a work-around for that. We can maybe fill an issue for it so we can give it a look later?