holado 0.13.22__py3-none-any.whl

holado_rabbitmq/tools/rabbitmq/rabbitmq_blocking_client.py

@@ -0,0 +1,333 @@
+
+#################################################
+# HolAdo (Holistic Automation do)
+#
+# (C) Copyright 2021-2025 by Eric Klumpp
+# 
+# Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the “Software”), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
+# 
+# The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
+
+# The Software is provided “as is”, without warranty of any kind, express or implied, including but not limited to the warranties of merchantability, fitness for a particular purpose and noninfringement. In no event shall the authors or copyright holders be liable for any claim, damages or other liability, whether in an action of contract, tort or otherwise, arising from, out of or in connection with the software or the use or other dealings in the Software.
+#################################################
+
+import time
+import logging
+from holado_core.common.exceptions.technical_exception import TechnicalException
+from holado_rabbitmq.tools.rabbitmq.rabbitmq_client import RMQClient,\
+    RMQPublisher
+import threading
+import queue
+from holado_core.common.tools.tools import Tools
+from holado_python.standard_library.typing import Typing
+
+
+logger = logging.getLogger(__name__)
+
+
+if RMQClient.is_available():
+    import pika
+    from pika import exceptions
+    from pika.adapters.blocking_connection import _CallbackResult
+    
+    class RMQBlockingChannel(pika.adapters.blocking_connection.BlockingChannel):
+        """
+        Add features to pika.adapters.blocking_connection.BlockingChannel:
+          - Rapid close: possibility to rapidly close connection without data consistency verification.
+                         It is useful when using consumers for monitoring, like in testing context.
+          - Add time_limit parameter to start_consuming method.
+        """
+        def __init__(self, channel_impl, connection, rapid_close=False):
+            super().__init__(channel_impl, connection)
+            self.__rapid_close = rapid_close
+            self.__is_closing = False
+            self.__confirm_delivery = False
+            self.__is_consuming = False
+            self.__is_stopping_consuming = False
+        
+        def close(self, reply_code=0, reply_text="Normal shutdown"):
+            self.__is_closing = True
+            try:
+                super().close(reply_code=reply_code, reply_text=reply_text)
+            except ValueError as exc:
+                if self.__rapid_close and "file descriptor cannot be a negative integer" in str(exc):
+                    # This exception can appear during channel close with rapid close option
+                    pass
+                else:
+                    raise exc
+        
+        def _flush_output(self, *waiters):
+            if self.__rapid_close and self.__is_closing:
+                if Tools.do_log(logger, logging.DEBUG):
+                    logger.debug(f"Channel is closing, bypass output flush for rapid close")
+            else:
+                super()._flush_output(*waiters)
+            
+        def start_consuming(self, time_limit=None, time_limit_on_stop=None):
+            """
+            Override start_consuming method to add time_limit parameter.
+            """
+            if Tools.do_log(logger, logging.DEBUG):
+                logger.debug(f"[{self}] Start consuming...")
+            self.__is_consuming = True
+            try:
+                if not self.__rapid_close and time_limit is None and time_limit_on_stop is None:
+                    # Use pika implementation
+                    if Tools.do_log(logger, logging.DEBUG):
+                        logger.debug(f"[{self}] Start consuming with Pika implementation")
+                    return super().start_consuming()
+                else:
+                    if Tools.do_log(logger, logging.DEBUG):
+                        logger.debug(f"[{self}] Start consuming with HolAdo implementation")
+                    
+                # Check if called from the scope of an event dispatch callback
+                with self.connection._acquire_event_dispatch() as dispatch_allowed:
+                    if not dispatch_allowed:
+                        raise exceptions.ReentrancyError(
+                            'start_consuming may not be called from the scope of '
+                            'another BlockingConnection or BlockingChannel callback')
+                
+                self._impl._raise_if_not_open()
+                
+                # Process events as long as consumers exist on this channel
+                while self._consumer_infos:
+                    # This will raise ChannelClosed if channel is closed by broker
+                    if self.__is_stopping_consuming:
+                        t_limit = time_limit_on_stop
+                        if self.__rapid_close and t_limit is None:
+                            # To ensure a rapid close, time_limit shouldn't be None, set it to 0 while stopping consuming
+                            t_limit = 0
+                    else:
+                        t_limit = time_limit
+                    if t_limit is None and self.__rapid_close:
+                        # To ensure a rapid close, time_limit shouldn't be None, set it to 0.1s as a compromise between rapidity and resource cost 
+                        t_limit = 0.1
+                    
+                    if Tools.do_log(logger, logging.TRACE):  # @UndefinedVariable
+                        logger.trace(f"[{self}] Processing data events (time_limit={t_limit})")
+                    self._process_data_events(time_limit=t_limit)
+            finally:
+                self.__is_consuming = False
+                self.__is_stopping_consuming = False
+                if Tools.do_log(logger, logging.DEBUG):
+                    logger.debug(f"[{self}] Finished start consuming")
+        
+        def stop_consuming(self, consumer_tag=None):
+            self.__is_stopping_consuming = True
+            super().stop_consuming(consumer_tag=consumer_tag)
+        
+        @property
+        def has_confirm_delivery(self):
+            return self._delivery_confirmation
+        
+        def set_confirm_delivery(self):
+            self.confirm_delivery()
+            
+
+
+    class RMQBlockingConnection(pika.adapters.blocking_connection.BlockingConnection):
+        """
+        Add features to pika.adapters.blocking_connection.BlockingConnection:
+          - Rapid close: possibility to rapidly close connection without data consistency verification.
+                         It is useful when using consumers for monitoring, like in testing context.
+        """
+        def __init__(self, parameters=None, _impl_class=None, rapid_close=False):
+            super().__init__(parameters, _impl_class)
+            self.__rapid_close = rapid_close
+            self.__is_closing = False
+
+        def channel(self, channel_number=None):
+            """
+            Override channel method to use RMQBlockingChannel instead of BlockingChannel.
+            The implementation is a copy of BlockingConnection.channel.
+            """
+            with _CallbackResult(self._OnChannelOpenedArgs) as opened_args:
+                impl_channel = self._impl.channel(
+                    channel_number=channel_number,
+                    on_open_callback=opened_args.set_value_once)
+    
+                # Create our proxy channel
+                channel = RMQBlockingChannel(impl_channel, self, rapid_close=self.__rapid_close)
+    
+                # Link implementation channel with our proxy channel
+                impl_channel._set_cookie(channel)
+    
+                # Drive I/O until Channel.Open-ok
+                channel._flush_output(opened_args.is_ready)
+            
+            return channel
+
+        def close(self, reply_code=200, reply_text='Normal shutdown'):
+            self.__is_closing = True
+            super().close(reply_code=reply_code, reply_text=reply_text)
+            
+        def _flush_output(self, *waiters):
+            if self.__rapid_close and self.__is_closing:
+                if Tools.do_log(logger, logging.DEBUG):
+                    logger.debug(f"Connection is closing, bypass output flush for rapid close")
+            else:
+                super()._flush_output(*waiters)
+                
+
+class RMQBlockingClient(RMQClient):
+    USE_RMQBlockingPublisher = True
+    
+    def __init__(self):
+        super().__init__('RMQBlockingClient')
+    
+    def close(self):
+        super().close()
+        
+        # Close all consumers connections
+        for obj_kwargs in self._consumer_and_kwargs:
+            connection = obj_kwargs.object.connection
+            if connection and connection.is_open:
+                self._close_connection(connection)
+            
+        for obj_kwargs in self._buffer_consumer_and_kwargs:
+            connection = obj_kwargs.object.connection
+            if connection and connection.is_open:
+                self._close_connection(connection)
+    
+    def _new_connection(self):
+        res = None
+        connection_parameters = self._new_connection_parameters()
+        
+        last_exception = None
+        for _ in range(300):
+            try:
+                res = RMQBlockingConnection(connection_parameters, rapid_close=self.rapid_close)
+                break
+            except pika.exceptions.IncompatibleProtocolError as exc:
+                is_same_exception = str(last_exception) == str(exc) if last_exception else False
+                if not last_exception or not is_same_exception:
+                    logger.exception("Retry after possible temporary incompatible protocol errors while trying to connect with a blocking connection")
+                    last_exception = exc
+                elif is_same_exception:
+                    logger.warning("Retry after same exception")
+                time.sleep(0.1)
+            except Exception as exc:
+                is_same_exception = str(last_exception) == str(exc) if last_exception else False
+                if not last_exception or not is_same_exception:
+                    logger.exception("Retry after unexpected exception while trying to connect with a blocking connection")
+                    last_exception = exc
+                elif is_same_exception:
+                    logger.warning("Retry after same exception")
+                time.sleep(0.1)
+                
+        if res is None:
+            raise TechnicalException(f"Failed to connect with a blocking connection and parameters [{connection_parameters}]. Last error was (more details in report): {last_exception} (type: {Typing.get_object_class_fullname(last_exception)})")
+        return res
+    
+    def _new_publisher(self, connection, channel, exchange, routing_key, nb_runners):
+        if self.USE_RMQBlockingPublisher:
+            res = RMQBlockingPublisher(self, connection, channel, exchange, routing_key, nb_runners)
+        else:
+            res = RMQSimpleBlockingPublisher(self, connection, channel, exchange, routing_key)
+        return res
+        
+    def new_consumer(self, queue, message_callback, queue_args=None, exchange="", exchange_args=None, bind_args=None):
+        if not self.has_consumer and not self.has_buffer_consumer:
+            # Use default connection for first consumer or buffer consumer
+            return super().new_consumer(queue, message_callback, queue_args=queue_args, exchange=exchange, exchange_args=exchange_args, bind_args=bind_args)
+        
+        # Note: A new blocking connection is needed for each consumer due to multitasking possibilities of holado_rabbitmq.
+        #       When a client hosts many consumers and start_consuming is called on it (RMQClient.start_consuming), 
+        #       the method start_consuming is called in a separated thread for each consumer. But current implementation
+        #       of BlockingChannel supposes that all pika methods are called in the same thread.
+        connection = self._new_connection()
+        return self._new_consumer_with_connection(connection, queue, message_callback, queue_args, exchange, exchange_args, bind_args)
+    
+    def new_buffer_consumer(self, queue, queue_args=None, exchange="", exchange_args=None, bind_args=None):
+        if not self.has_consumer and not self.has_buffer_consumer:
+            # Use default connection for first consumer or buffer consumer
+            return super().new_buffer_consumer(queue, queue_args=queue_args, exchange=exchange, exchange_args=exchange_args, bind_args=bind_args)
+        
+        # Note: A new blocking connection is needed for each consumer due to multitasking possibilities of holado_rabbitmq.
+        #       When a client hosts many consumers and start_consuming is called on it (RMQClient.start_consuming), 
+        #       the method start_consuming is called in a separated thread for each consumer. But current implementation
+        #       of BlockingChannel supposes that all pika methods are called in the same thread.
+        connection = self._new_connection()
+        return self._new_buffer_consumer_with_connection(connection, queue, queue_args=queue_args, exchange=exchange, exchange_args=exchange_args, bind_args=bind_args)
+    
+class RMQBlockingPublisher(RMQPublisher):
+    """
+    Implementation of a publisher for long and intensive publishes.
+    """
+    
+    def __init__(self, client, connection, channel, exchange, routing_key, nb_runners=None):
+        super().__init__(client, connection, channel, exchange, routing_key)
+
+        self.__publish_lock = threading.Lock()
+        self.__publish_nb_runners = nb_runners if nb_runners is not None else 10
+        self.__publish_running_nb = 0
+        self.__publish_queue = queue.Queue(maxsize=self.__publish_nb_runners * 2)
+        self.__publish_counter = 0
+        self.__published_counter = 0
+
+    def flush(self):
+        if Tools.do_log(logger, logging.DEBUG):
+            logger.debug(f"[{self.name}] flushing...")
+        self.__publish_queue.join()
+        while self.__published_counter < self.__publish_counter:
+            time.sleep(0.01)
+        if Tools.do_log(logger, logging.DEBUG):
+            logger.debug(f"[{self.name}] flushed")
+
+    def __publish_start(self):
+        with self.__publish_lock:
+            if self.__publish_running_nb < self.__publish_nb_runners:
+                self.__publish_running_nb += 1
+            else:
+                return
+        
+        self.__publish_next()
+
+    def __publish_next(self):
+        try:
+            msg_item = self.__publish_queue.get(block=False)
+        except queue.Empty:
+            with self.__publish_lock:
+                self.__publish_running_nb -= 1
+            return
+        
+        # Note: following code doesn't seem to be more efficient, thus it is commented
+        # if self.client.is_in_connection_thread:
+        #     self.client.connection.call_later(0, lambda: self.__publish_message(msg_item[0], **msg_item[1]))
+        # else:
+        #     self.client.connection.add_callback_threadsafe(
+        #         lambda: self.__publish_message(msg_item[0], **msg_item[1])
+        #         )
+        self.client.connection.add_callback_threadsafe(
+            lambda: self.__publish_message(msg_item[0], **msg_item[1]) )
+        
+        self.__publish_queue.task_done()
+
+    def __publish_message(self, body, **kwargs):
+        # Publish message
+        RMQPublisher.publish(self, body, **kwargs)
+        self.__published_counter += 1
+
+        # Manage to publish next message
+        self.__publish_next()
+
+    def publish(self, body, **kwargs):
+        # Start data events processing if needed
+        self.client.start_process_data_events(raise_if_already_started=False)
+        
+        # Add message to publish in queue
+        self.__publish_queue.put((body, kwargs))
+        self.__publish_counter += 1
+
+        # Start publishing if it is not already running
+        self.__publish_start()
+
+class RMQSimpleBlockingPublisher(RMQPublisher):
+    """
+    Simple implementation of a publisher
+    """
+    def __init__(self, client, connection, channel, exchange, routing_key):
+        super().__init__(client, connection, channel, exchange, routing_key)
+        
+
+