Remove duplicates in breadcrumbs

[drew2a]

Introduced a new method _format_breadcrumbs to handle the formatting of breadcrumbs in SentryReporter. This includes removing duplicates and ordering by timestamp. The logic for this has been extracted from the _before_send method.

An event example: https://sentry.tribler.org/organizations/tribler/issues/2658/events/b42dc2acdc184c709ab8f9a60367c23e/

This is a post-work of

• Add breadcrumb ordering #8034

[kozlovsky]

With this approach, the keys are ordered according to the order as they first appear in the list, but the item stored is the last item corresponding to this key. That is, the latest item with the same key overrides the first item but keeps the position of the first item. If we consider distinct_by a general purpose function, in some use cases, it can lead to a confusing order (depending on what is an "item" and what is a "key") because it can break expectations of the function's user, that probably expects that the first item is stored, not the last one, but for breadcrumbs, it should work fine.

[drew2a]

I know, but my reasoning is that a user should not rely on expectations about ordering if it is not explicitly written in the documentation or reflected in the function design. In our case, what is guaranteed is that the values will be unique based on the given key (getter).

However, we can use the following logic:

    distinct = {}
    for item in items:
        key = getter(item)
        if key not in distinct:
            distinct[key] = item
    return list(distinct.values())

[kozlovsky]

Looks good!

[drew2a]

@kozlovsky thank you for the review 🙏