✨ feat: Add duplicate LLM request prevention

[cxyfer]

Summary

• Implement request tracking to prevent duplicate LLM API calls
• Prevent users from clicking translation/inspiration buttons multiple times
• Add proper cleanup with try/finally blocks

Test plan

• Test rapid clicking on translation button
• Test rapid clicking on inspiration button
• Verify cached responses still work correctly
• Check error handling doesn't break cleanup

🤖 Generated with Claude Code

[Copilot]

Pull Request Overview

This PR introduces a mechanism to track and prevent duplicate LLM requests in the Discord bot and adds comprehensive tests and documentation for this feature.

• Add ongoing_llm_requests set with an asyncio.Lock to serialize check-and-add in handlers
• Augment translation and inspiration branches with duplicate-request checks and proper cleanup in finally blocks
• New async tests (test_interaction_handler.py) and updated tests/README.md for running interaction handler tests

Reviewed Changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 3 comments.

File	Description
tests/test_interaction_handler.py	New async test suite covering duplicate prevention, cleanup, and concurrency
tests/README.md	Document new tests and commands for running interaction handler tests
cogs/interaction_handler_cog.py	Implement duplicate-request lock, prevention logic, and cleanup

Comments suppressed due to low confidence (2)

tests/README.md:104

• [nitpick] The command uv run pytest seems incorrect; consider updating it to a standard pytest tests/test_interaction_handler.py -v invocation for clarity.

uv run pytest tests/test_interaction_handler.py -v

cogs/interaction_handler_cog.py:152

• Consider adding a unit test that verifies _cleanup_request is called when serving cached translations, ensuring that the request is removed from tracking after a cached response.

                        await self._cleanup_request(request_key)

[Copilot]

This inline cleanup call is redundant since the final finally block performs cleanup as well; consider removing branch-specific cleanup to simplify the flow.

Suggested change

	await interaction.followup.send(translation, ephemeral=True)
	# Remove from ongoing requests when returning cached result
	await self._cleanup_request(request_key)
	return
	await interaction.followup.send(translation, ephemeral=True)
	return

[Copilot]

The patch import is not used in this file; consider removing it to clean up unused imports.

Suggested change

	from unittest.mock import MagicMock, AsyncMock, patch
	from unittest.mock import MagicMock, AsyncMock

[Copilot]

[nitpick] Imports could be reorganized to follow PEP8 ordering (e.g., place import asyncio before import time) for consistency.

Suggested change

	import discord
	from discord.ext import commands
	import time
	import asyncio
	from leetcode import html_to_text # 確保這個 import 存在
	# from utils.logger import get_logger # 使用 bot.logger
	import asyncio
	import time
	import discord
	from discord.ext import commands
	from leetcode import html_to_text # 確保這個 import 存在
	# from utils.logger import get_logger # 使用 bot.logger