Formatting helpers

[MKRhere]

Ref: related conversation

Formatting is often troublesome, especially if you have many characters that need escaping. Escaping is messy, and it is much cleaner to create entities instead, but there were no utils to help with this in Telegraf so far. This PR adds support for telegraf/format. To be used like this:

import { fmt, bold, italic, underline, mention, link } from "telegraf/format";

ctx.reply(fmt`
Ground control to ${mention("Major Tom", 10000000)}
${bold`Lock your Soyuz hatch`} and ${italic`put your helmet on`}

— ${link(underline`David Bowie`), "https://en.wikipedia.org/wiki/David_Bowie")}
`);

This also just works with captions!

ctx.replyWithPhoto(
  file.id,
  { caption: fmt`${bold`File name:`} ${file.name}` },
);

For the future: emoji() may be a nice addition to formatting; however emojibase is quite large and would not be a suitable dependency for telegraf. Creating a separate @telegraf/format-emoji package could work.

[Kamrulhasan12345]

Looks ok.

[KnorpelSenf]

So effectively you're slowly integrating the most useful plugins into the core lib? That sounds like a better strategy than hoping for permission to publish packages under @telegraf/ :D

[MKRhere]

@KnorpelSenf I have access to npm @telegraf/ now :) I don't have access to the GitHub org to create repositories yet. But this is the strategy for telegraf now - everything that's very useful and is lightweight (~100 loc) goes in the core.

I started this PR thinking of format being a tiny function in Markup.fmt which sounded very appropriate, but it's grown into a full file :D This is mostly because we don't want to clutter the main import space, but still want to import { bold, italic, ... } as top-level items with import suggestions.

The benefit of in-core here is that user doesn't have to mess with Context/API types or implementation - even external plugins (like a future @telegraf/format-emoji) can use the FmtString contract and telegraf will know what to do.

[Viiprogrammer]

Interesting idea, but what about localization?

Many bots use it. I think the use of such solutions complicates the localization and translating of the text, even if i18n was not used now, may be needed in the future.

About escaping - telegram-escape

[MKRhere]

@Viiprogrammer That's a very good point - at the moment I don't think there's any way to reconcile fmt and i18n.

may be needed in the future.

Any solution in this regard will need to be rewritten when i18n is needed - so not many ways to be prematurely prepared for it. I think the two needs are orthographic.