Add logging
Browse files
worker.py
CHANGED
|
@@ -10,11 +10,19 @@ from typing import Tuple, Dict, Any
|
|
| 10 |
|
| 11 |
from mineru_single import Processor
|
| 12 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
|
| 14 |
class RabbitMQWorker:
|
| 15 |
def __init__(self, num_workers: int = 1):
|
| 16 |
self.num_workers = num_workers
|
| 17 |
self.rabbit_url = os.getenv("RABBITMQ_URL")
|
|
|
|
| 18 |
self.processor = Processor()
|
| 19 |
|
| 20 |
self.publisher_connection = None
|
|
@@ -23,6 +31,7 @@ class RabbitMQWorker:
|
|
| 23 |
|
| 24 |
def setup_publisher(self):
|
| 25 |
if not self.publisher_connection or self.publisher_connection.is_closed:
|
|
|
|
| 26 |
connection_params = pika.URLParameters(self.rabbit_url)
|
| 27 |
connection_params.heartbeat = 1000 # Match the consumer heartbeat
|
| 28 |
connection_params.blocked_connection_timeout = 500 # Increase timeout
|
|
@@ -30,6 +39,7 @@ class RabbitMQWorker:
|
|
| 30 |
self.publisher_connection = pika.BlockingConnection(connection_params)
|
| 31 |
self.publisher_channel = self.publisher_connection.channel()
|
| 32 |
self.publisher_channel.queue_declare(queue="ml_server", durable=True)
|
|
|
|
| 33 |
|
| 34 |
def publish_message(self, body_dict: dict, headers: dict):
|
| 35 |
"""Use persistent connection for publishing"""
|
|
@@ -48,9 +58,10 @@ class RabbitMQWorker:
|
|
| 48 |
headers=headers
|
| 49 |
)
|
| 50 |
)
|
|
|
|
| 51 |
return True
|
| 52 |
except Exception as e:
|
| 53 |
-
|
| 54 |
# Close failed connection
|
| 55 |
if self.publisher_connection and not self.publisher_connection.is_closed:
|
| 56 |
try:
|
|
@@ -61,7 +72,7 @@ class RabbitMQWorker:
|
|
| 61 |
self.publisher_channel = None
|
| 62 |
|
| 63 |
if attempt == max_retries - 1:
|
| 64 |
-
|
| 65 |
return False
|
| 66 |
time.sleep(2)
|
| 67 |
|
|
@@ -70,7 +81,7 @@ class RabbitMQWorker:
|
|
| 70 |
thread_id = threading.current_thread().name
|
| 71 |
headers = properties.headers or {}
|
| 72 |
|
| 73 |
-
|
| 74 |
|
| 75 |
try:
|
| 76 |
if headers.get("request_type") == "process_files":
|
|
@@ -81,6 +92,7 @@ class RabbitMQWorker:
|
|
| 81 |
# Process files
|
| 82 |
for file in body_dict.get("input_files", []):
|
| 83 |
try:
|
|
|
|
| 84 |
context = {"key": file["key"], "body": self.processor.process(file["url"], file["key"])}
|
| 85 |
contexts.append(context)
|
| 86 |
except Exception as e:
|
|
@@ -94,18 +106,18 @@ class RabbitMQWorker:
|
|
| 94 |
|
| 95 |
# Publish results
|
| 96 |
if self.publish_message(body_dict, headers):
|
| 97 |
-
|
| 98 |
else:
|
| 99 |
ch.basic_nack(delivery_tag=method.delivery_tag, requeue=True)
|
| 100 |
-
|
| 101 |
|
| 102 |
-
|
| 103 |
else:
|
| 104 |
ch.basic_ack(delivery_tag=method.delivery_tag)
|
| 105 |
-
|
| 106 |
|
| 107 |
except Exception as e:
|
| 108 |
-
|
| 109 |
ch.basic_nack(delivery_tag=method.delivery_tag, requeue=True)
|
| 110 |
|
| 111 |
def connect_to_rabbitmq(self):
|
|
@@ -113,6 +125,8 @@ class RabbitMQWorker:
|
|
| 113 |
connection_params = pika.URLParameters(self.rabbit_url)
|
| 114 |
connection_params.heartbeat = 1000
|
| 115 |
connection_params.blocked_connection_timeout = 500
|
|
|
|
|
|
|
| 116 |
|
| 117 |
connection = pika.BlockingConnection(connection_params)
|
| 118 |
channel = connection.channel()
|
|
@@ -123,21 +137,23 @@ class RabbitMQWorker:
|
|
| 123 |
queue="gpu_server",
|
| 124 |
on_message_callback=self.callback
|
| 125 |
)
|
|
|
|
| 126 |
return connection, channel
|
| 127 |
|
| 128 |
def worker(self, channel):
|
| 129 |
"""Worker function"""
|
| 130 |
-
|
| 131 |
try:
|
| 132 |
channel.start_consuming()
|
| 133 |
except Exception as e:
|
| 134 |
-
|
| 135 |
finally:
|
|
|
|
| 136 |
channel.close()
|
| 137 |
|
| 138 |
def start(self):
|
| 139 |
"""Start the worker threads"""
|
| 140 |
-
|
| 141 |
while True:
|
| 142 |
try:
|
| 143 |
with ThreadPoolExecutor(max_workers=self.num_workers) as executor:
|
|
@@ -145,7 +161,7 @@ class RabbitMQWorker:
|
|
| 145 |
connection, channel = self.connect_to_rabbitmq()
|
| 146 |
executor.submit(self.worker, channel)
|
| 147 |
except Exception as e:
|
| 148 |
-
|
| 149 |
time.sleep(5) # Wait before reconnecting
|
| 150 |
|
| 151 |
def main():
|
|
|
|
| 10 |
|
| 11 |
from mineru_single import Processor
|
| 12 |
|
| 13 |
+
import logging
|
| 14 |
+
|
| 15 |
+
logging.basicConfig(
|
| 16 |
+
level=logging.INFO,
|
| 17 |
+
format="%(asctime)s [%(levelname)s] %(name)s - %(message)s"
|
| 18 |
+
)
|
| 19 |
+
logger = logging.getLogger(__name__)
|
| 20 |
|
| 21 |
class RabbitMQWorker:
|
| 22 |
def __init__(self, num_workers: int = 1):
|
| 23 |
self.num_workers = num_workers
|
| 24 |
self.rabbit_url = os.getenv("RABBITMQ_URL")
|
| 25 |
+
logger.info("Initializing RabbitMQWorker")
|
| 26 |
self.processor = Processor()
|
| 27 |
|
| 28 |
self.publisher_connection = None
|
|
|
|
| 31 |
|
| 32 |
def setup_publisher(self):
|
| 33 |
if not self.publisher_connection or self.publisher_connection.is_closed:
|
| 34 |
+
logger.info("Setting up publisher connection to RabbitMQ.")
|
| 35 |
connection_params = pika.URLParameters(self.rabbit_url)
|
| 36 |
connection_params.heartbeat = 1000 # Match the consumer heartbeat
|
| 37 |
connection_params.blocked_connection_timeout = 500 # Increase timeout
|
|
|
|
| 39 |
self.publisher_connection = pika.BlockingConnection(connection_params)
|
| 40 |
self.publisher_channel = self.publisher_connection.channel()
|
| 41 |
self.publisher_channel.queue_declare(queue="ml_server", durable=True)
|
| 42 |
+
logger.info("Publisher connection/channel established successfully.")
|
| 43 |
|
| 44 |
def publish_message(self, body_dict: dict, headers: dict):
|
| 45 |
"""Use persistent connection for publishing"""
|
|
|
|
| 58 |
headers=headers
|
| 59 |
)
|
| 60 |
)
|
| 61 |
+
logger.info("Published message to ml_server queue (attempt=%d).", attempt + 1)
|
| 62 |
return True
|
| 63 |
except Exception as e:
|
| 64 |
+
logger.error("Publish attempt %d failed: %s", attempt + 1, e)
|
| 65 |
# Close failed connection
|
| 66 |
if self.publisher_connection and not self.publisher_connection.is_closed:
|
| 67 |
try:
|
|
|
|
| 72 |
self.publisher_channel = None
|
| 73 |
|
| 74 |
if attempt == max_retries - 1:
|
| 75 |
+
logger.error("Failed to publish after %d attempts", max_retries)
|
| 76 |
return False
|
| 77 |
time.sleep(2)
|
| 78 |
|
|
|
|
| 81 |
thread_id = threading.current_thread().name
|
| 82 |
headers = properties.headers or {}
|
| 83 |
|
| 84 |
+
logger.info("[Worker %s] Received message: %s, headers: %s", thread_id, body, headers)
|
| 85 |
|
| 86 |
try:
|
| 87 |
if headers.get("request_type") == "process_files":
|
|
|
|
| 92 |
# Process files
|
| 93 |
for file in body_dict.get("input_files", []):
|
| 94 |
try:
|
| 95 |
+
logger.info("[Worker %s] Processing %d input files.", thread_id, len(file))
|
| 96 |
context = {"key": file["key"], "body": self.processor.process(file["url"], file["key"])}
|
| 97 |
contexts.append(context)
|
| 98 |
except Exception as e:
|
|
|
|
| 106 |
|
| 107 |
# Publish results
|
| 108 |
if self.publish_message(body_dict, headers):
|
| 109 |
+
logger.info("[Worker %s] Successfully published results to ml_server.", thread_id)
|
| 110 |
else:
|
| 111 |
ch.basic_nack(delivery_tag=method.delivery_tag, requeue=True)
|
| 112 |
+
logger.error("[Worker %s] Failed to publish results.", thread_id)
|
| 113 |
|
| 114 |
+
logger.info("[Worker %s] Contexts: %s", thread_id, contexts)
|
| 115 |
else:
|
| 116 |
ch.basic_ack(delivery_tag=method.delivery_tag)
|
| 117 |
+
logger.warning("[Worker %s] Unknown process type in headers: %s", thread_id, headers.get("request_type"))
|
| 118 |
|
| 119 |
except Exception as e:
|
| 120 |
+
logger.error("Error in callback: %s", e)
|
| 121 |
ch.basic_nack(delivery_tag=method.delivery_tag, requeue=True)
|
| 122 |
|
| 123 |
def connect_to_rabbitmq(self):
|
|
|
|
| 125 |
connection_params = pika.URLParameters(self.rabbit_url)
|
| 126 |
connection_params.heartbeat = 1000
|
| 127 |
connection_params.blocked_connection_timeout = 500
|
| 128 |
+
logger.info("Connecting to RabbitMQ for consumer with heartbeat=1000.")
|
| 129 |
+
|
| 130 |
|
| 131 |
connection = pika.BlockingConnection(connection_params)
|
| 132 |
channel = connection.channel()
|
|
|
|
| 137 |
queue="gpu_server",
|
| 138 |
on_message_callback=self.callback
|
| 139 |
)
|
| 140 |
+
logger.info("Consumer connected. Listening on queue='gpu_server'...")
|
| 141 |
return connection, channel
|
| 142 |
|
| 143 |
def worker(self, channel):
|
| 144 |
"""Worker function"""
|
| 145 |
+
logger.info("Worker thread started. Beginning consuming...")
|
| 146 |
try:
|
| 147 |
channel.start_consuming()
|
| 148 |
except Exception as e:
|
| 149 |
+
logger.error("Worker thread encountered an error: %s", e)
|
| 150 |
finally:
|
| 151 |
+
logger.info("Worker thread shutting down. Closing channel.")
|
| 152 |
channel.close()
|
| 153 |
|
| 154 |
def start(self):
|
| 155 |
"""Start the worker threads"""
|
| 156 |
+
logger.info("Starting %d workers in a ThreadPoolExecutor.", self.num_workers)
|
| 157 |
while True:
|
| 158 |
try:
|
| 159 |
with ThreadPoolExecutor(max_workers=self.num_workers) as executor:
|
|
|
|
| 161 |
connection, channel = self.connect_to_rabbitmq()
|
| 162 |
executor.submit(self.worker, channel)
|
| 163 |
except Exception as e:
|
| 164 |
+
logger.error("Connection lost, reconnecting... Error: %s", e)
|
| 165 |
time.sleep(5) # Wait before reconnecting
|
| 166 |
|
| 167 |
def main():
|