zulip_bots: Add a script for creating Zulip bots. #709
Conversation
| "name": "{bot_name}", | ||
| "version": {bot_name}.__version__, | ||
| "entry_points": {{ | ||
| "zulip_bots.registry": ["{bot_name}={bot_name}.{bot_name}"], |
There was a problem hiding this comment.
Does this working assume the generating directory (default .) is actually equal to the bot name that is supplied? If so, it seems cleaner to make the base folder and the nested one relative to the current location?
There was a problem hiding this comment.
The name of the base directory doesn't matter. We assume that the nested directory and the bot module have the same name as the bot. This is the same practice we have for bots under the zulip_bots/zulip_bots/bots directory.
There was a problem hiding this comment.
Right you are, this was from a quick read rather than test.
This removes the need to use -r flag and instead load the bot from the registry when no other method works.
Following support to running bots from entry points in zulip#708, we implement this `create-zulip-bot` tool to simplify the process of creating new bots. The user will be able to directly install the package with pip and run the bot with `zulip-run-bot`, or use it to quickly set up a git repository. Note that the boilerplate generated by this script does not contain `tests.py` yet. We need to figure out the right pattern for integrating unittests for such packaged bots.
|
Can you explain the motivation for the first commit? I can imagine a world where one would prefer to find a bot using a specific path, and don't have a clear idea of whether there's any potential situation where we might prefer semantics of "Please only use a bot from the registry". Maybe the right thing to do is add a docstring in the bot discovery function explaining clearly the sources it looks at the motivation for why they're prioritized the way they are? That seems like a good resource in general and in particular would make it a lot easier for me to review this change. |
|
@PIG208 This looks like it needs a rebase and tidy, but otherwise would progressing this be easier if the two commits were split between PRs? The first commit isn't necessary for the second, right? |
| bot_module_path = Path(bot_path, args.bot) | ||
|
|
||
| try: | ||
| os.mkdir(bot_path) |
There was a problem hiding this comment.
Reading this again, is there a reason not to use the pathlib functions instead, if we're already using Path? That is, can we drop import os?
|
Heads up @PIG208, we just merged some commits that conflict with the changes you made in this pull request! You can review this repository's recent commits to see where the conflicts occur. Please rebase your feature branch against the |
Following support to running bots from entry points in #708, we
implement this
create-zulip-bottool to simplify the process ofcreating new bots. The user will be able to directly install the package
with pip and run the bot with
zulip-run-bot, or use it to quickly setup a git repository.
Note that the boilerplate generated by this script does not contain
tests.pyyet. We need to figure out the right pattern for integratingunittests for such packaged bots.