| Commit message (Collapse) | Author | Lines |
|
|
|
* Mention possibility of timing out as a reason for aborting a sync
|
|
|
|
|
|
|
|
It doesn't play along well with NamedTuple due to metaclass conflicts.
The workaround involved created a NamedTuple-only base class, which
does work but at the cost of confusing some static type checkers.
Since Diff is now an internal data structure, it no longer really needs
to have precise type annotations. Therefore, a normal namedtuple is
adequate.
|
|
The interface was becoming cumbersome to work with so it was all moved
to a single location. Now just calling Syncer.sync() will take care of
everything.
* Remove Optional type annotation from Diff attributes
* _confirm() can edit the original message and use it as the prompt
* Calculate the total diff and compare it against the max before
sending a confirmation prompt
* Remove abort message from sync(); _confirm() will handle that
|
|
The confirmation prompt will be sent to the dev-core channel or the
specified context. Confirmation is done via reactions and waits 5
minutes before timing out.
* Add name property to Syncers
* Make _get_diff private; only sync() needs to be called now
* Change spelling of synchronize to synchronise
* Update docstrings
|
|
Replaces the functions with a class for each syncer. The classes
inherit from a Syncer base class. A NamedTuple was also created to
replace the tuple of the object differences that was previously being
returned.
* Use namedtuple._asdict to simplify converting namedtuples to JSON
|
|
|
|
|
|
|
|
* Use ID from constants directly instead of SYNC_SERVER_ID
* Use f-strings instead of %s for logging
* Fit into margin of 100
* Invert condition to reduce nesting
* Use Any instead of incorrect function annotation for JSON values
|
|
Reduces code redundancy.
|
|
This seems to have been the intent of the original implementation.
|
|
|
|
It has a much better guarantee that the cache will be available.
|
|
|
|
This coroutine waits until the configured guild is available and
ensures the cache is present.
The on_ready event is inadequate because it only waits 2 seconds for a
GUILD_CREATE gateway event before giving up and thus not populating the
cache for unavailable guilds.
|
|
Fixes #746.
|
|
It could be confused with the delete reaction. Clearing reactions
manually is rarely a useful feature anyway.
|
|
|
|
|
|
|
|
After the wiki pages adjustment to the resources page, these two urls are needing to be updated to point to the new correct locations for each page.
Tools will be under resources, and Asking Good Questions is a guide, so will be under Guides.
|
|
- Added regex back to sub and split by non-alphabet.
- Now use two pointers to move from words to words.
|
|
|
|
- Since it is returning as soon as there are suggestions found for a threshold, this will give a better reflection of what the bot thinks user is searching for.
|
|
- Added a regex to remove non-alphabet ( `[^a-z]` with `re.IGNORECASE` )
|
|
- Matching scores will be calculated once now and stored in the dict `scores`.
- Allow `_get_suggestions()` to go through a list of score threshold and return the first list of matching tags that's not empty and above the threshold. This avoid calling the function multiple time like before ( `self._get_suggestions(tag_name, 100) or self._get_suggestions(tag_name, 80)` for example, is calling this function twice, and is inefficient )
- Deleted commented line.
- Added `typing` module for more typehints.
|
|
- Changed type of `self._last_fetch` to `float` and give it the initial value of `0.0` instead of `None`
- Assigned `time.time()` to `time_now` to avoid calling this function twice.
- Added `self._last_fetch = time_now` after calling the api call.
|
|
Previously used a leading underscore, but that's usually meant for non-used names.
|
|
|
|
|
|
|
|
|
|
|