version 1.2.1

[yedidyakfir]

No description provided.

[Copilot]

Pull request overview

This pull request bumps the version from 1.2.0 to 1.2.1 and introduces a global rapyer.apipeline() context manager for batching Redis operations across multiple models, along with several important fixes for NOSCRIPT error recovery, nested pipeline support, and inherited field assignment in pipelines.

Changes:

• Added global rapyer.apipeline() function for batching operations across multiple models without requiring a specific model instance
• Implemented NOSCRIPT error recovery in arun_sha() to automatically re-register Lua scripts after Redis restarts
• Fixed __setattr__ to properly handle inherited model fields by checking model_fields instead of __annotations__

Reviewed changes

Copilot reviewed 26 out of 26 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
rapyer/base.py	Added global apipeline() function, updated __setattr__ to use model_fields, added client property, refactored model's apipeline() to use global function
rapyer/scripts/registry.py	Added NOSCRIPT error recovery to arun_sha() with handle_noscript_error(), updated to use variant constants
rapyer/types/dct.py	Updated apop() and apopitem() to pass self.Meta to arun_sha() for proper error recovery
rapyer/scripts/constants.py	Added REDIS_VARIANT and FAKEREDIS_VARIANT constants
rapyer/scripts/loader.py	Updated to use variant constants instead of string literals
rapyer/context.py	Removed unused _context_xx_pipe variable
rapyer/init.py	Exported new apipeline function
pyproject.toml	Version bump to 1.2.1
CHANGELOG.md	Documented all changes for version 1.2.1
docs/documentation/atomic-actions.md	Added documentation for global pipeline feature
docs/api/rapyer-functions.md	Added API documentation for apipeline() function
tests/integration/conftest.py	Added flush_scripts and disable_noscript_recovery fixtures
tests/integration/pipeline/test_pipeline_noscript_recovery.py	Updated tests to use new fixtures, removed inline script flushing
tests/integration/pipeline/test_pipeline_asave_batching.py	New test file for batched asave operations
tests/integration/pipeline/test_apipeline_field_assignment.py	New test file for field assignment in pipelines
tests/integration/dct/test_redis_dict.py	Added test for persistent NOSCRIPT error
tests/unit/test_scripts.py	Added test for NOSCRIPT handling with FakeRedis
tests/unit/types/test_dict_lua_scripts_with_fakeredis.py	Added tests for dict operations NOSCRIPT recovery
tests/integration/pipeline/test_pipeline_context_manager_atomic_operations.py	Moved import to top of file
.planning/codebase/*	Added new planning documentation files (TESTING.md, STRUCTURE.md, STACK.md, etc.)

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Critical bug: The pipeline context manager exits immediately after line 698, before yielding the pipeline object. This causes the redis-py pipeline context manager to auto-execute the (empty) pipeline when exiting the context. The pipeline object is then yielded on line 700, but it's no longer in an active context manager state.

The entire try-finally block (lines 699-741) should be inside the async with redis.pipeline(transaction=True) as pipe: block. The correct structure should have the async with encompass everything up to the finally block.

Suggested change

	try:
	yield pipe
	commands_backup = list(pipe.command_stack)
	noscript_on_first_attempt = False
	noscript_on_retry = False
	try:
	await pipe.execute()
	except NoScriptError:
	noscript_on_first_attempt = True
	except ResponseError as exc:
	if ignore_redis_error:
	logger.warning(
	"Swallowed ResponseError during pipeline.execute() with "
	"ignore_redis_error=True: %s",
	exc,
	)
	else:
	raise
	if noscript_on_first_attempt:
	await scripts_registry.handle_noscript_error(redis, _meta)
	evalsha_commands = [
	(args, options)
	for args, options in commands_backup
	if args[0] == "EVALSHA"
	]
	# Retry execute the pipeline actions
	async with redis.pipeline(transaction=True) as retry_pipe:
	for args, options in evalsha_commands:
	retry_pipe.execute_command(*args, **options)
	try:
	await retry_pipe.execute()
	except NoScriptError:
	noscript_on_retry = True
	if noscript_on_retry:
	raise PersistentNoScriptError(
	"NOSCRIPT error persisted after re-registering scripts. "
	"This indicates a server-side problem with Redis."
	)
	finally:
	_context_var.reset(pipe_prev)
	try:
	yield pipe
	commands_backup = list(pipe.command_stack)
	noscript_on_first_attempt = False
	noscript_on_retry = False
	try:
	await pipe.execute()
	except NoScriptError:
	noscript_on_first_attempt = True
	except ResponseError as exc:
	if ignore_redis_error:
	logger.warning(
	"Swallowed ResponseError during pipeline.execute() with "
	"ignore_redis_error=True: %s",
	exc,
	)
	else:
	raise
	if noscript_on_first_attempt:
	await scripts_registry.handle_noscript_error(redis, _meta)
	evalsha_commands = [
	(args, options)
	for args, options in commands_backup
	if args[0] == "EVALSHA"
	]
	# Retry execute the pipeline actions
	async with redis.pipeline(transaction=True) as retry_pipe:
	for args, options in evalsha_commands:
	retry_pipe.execute_command(*args, **options)
	try:
	await retry_pipe.execute()
	except NoScriptError:
	noscript_on_retry = True
	if noscript_on_retry:
	raise PersistentNoScriptError(
	"NOSCRIPT error persisted after re-registering scripts. "
	"This indicates a server-side problem with Redis."
	)
	finally:
	_context_var.reset(pipe_prev)