basic backend

rag-chat-backend/src/common_packages/dashboard_logging.py

@@ -0,0 +1,102 @@
+""" This module defines the logging component for the LLM interactions.
+The data is displayed in the OpenSearch Dashboard.
+"""
+
+import os
+import uuid
+import logging
+
+from opensearch_logger import OpenSearchHandler
+from opensearchpy import RequestsHttpConnection
+
+
+class DashboardLogger:
+    """Logger instance for OpenSearch dashboard"""
+
+    def __init__(self, logger_name=None):
+        if logger_name is None:
+            # Generate a unique logger name
+            logger_name = "rag-logs-" + str(uuid.uuid4())
+        self.logger_instance = self._create_os_logger(logger_name=logger_name)
+        self.logger_info = {}
+
+    def add_information(self, label: str, value):
+        """Here you can add information to a given process.
+        Each information consists of a label and its given value.
+
+        Args:
+            label (str): label
+            value (any): value
+        """
+        self.logger_info[label] = value
+
+    def close_logging(self):
+        """Close logging in the final process."""
+
+        self.logger_info = self._roll_out_json(original_json=self.logger_info)
+
+        if type(self.logger_info) == list:
+            for logger_json in self.logger_info:
+                self.logger_instance.info("Logging information", extra=logger_json)
+        else:
+            self.logger_instance.info("Logging information", extra=self.logger_info)
+
+    def _create_os_logger(self, logger_name):
+        """Create a logger which logs on OpenSearch.
+
+        Args:
+            logger_name (str): use some logger name.
+        Returns:
+            logger: OpenSearch logger instance.
+        """
+
+        logger = logging.getLogger(logger_name)
+        logger.setLevel(logging.INFO)
+
+        opensearch_host = os.getenv("VECTOR_STORE_ENDPOINT", "localhost")
+        opensearch_port = os.getenv("VECTOR_STORE_PORT", "9200")
+
+        handler = OpenSearchHandler(
+            index_name="rag-logs",
+            hosts=[f"http://{opensearch_host}:{opensearch_port}"],
+            http_auth=("admin", "admin"),
+            use_ssl=os.getenv("OPENSEARCH_USE_SSL", "False").lower()
+            in ["true", "1", "yes", "y"],
+            verify_certs=False,
+            connection_class=RequestsHttpConnection,
+        )
+
+        logger.addHandler(handler)
+
+        return logger
+
+    def _roll_out_json(self, original_json):
+        """Roll out JSON with lists into individual JSONs with additional attributes.
+
+        Args:
+            original_json: The original JSON with lists in 'sources' and 'passages' attributes.
+        Return
+            rolled_out_jsons: List of rolled out JSONs.
+        """
+
+        rolled_out_jsons = []
+
+        # Iterate through each item in the 'sources' and 'passages' lists
+        for rank, (source, passage) in enumerate(
+            zip(original_json["sources"], original_json["passages"]), start=1
+        ):
+            # Create a new JSON for each pair of source and passage
+            new_json = {
+                "query": original_json["query"],
+                "source": source["source"],
+                "page": source["page"],
+                # NOTE we need to cut. Otherwise the visualization wont work with these many chars
+                "passage": passage[:250],
+                "answer": original_json["answer"][:250],
+                "language": original_json["language"],
+                "model": original_json["model"],
+                "rank": rank,  # Assign the rank
+            }
+            rolled_out_jsons.append(new_json)
+
+        return rolled_out_jsons

rag-chat-backend/src/common_packages/logging.py

@@ -0,0 +1,46 @@
+""" This module defines the logging component."""
+
+import logging
+
+
+def create_logger(log_level: str, logger_name: str = "custom_logger"):
+    """Create a logging based on logger.
+
+    Args:
+        log_level (str): Kind of logging
+        logger_name (str, optional): Name of logger
+
+    Returns:
+        logger: returns logger
+    """
+    logger = logging.getLogger(logger_name)
+    logger.setLevel(logging.DEBUG)  # Set the base logging level to the lowest (DEBUG)
+
+    # If logger already has handlers, don't add a new one
+    if logger.hasHandlers():
+        logger.handlers.clear()
+
+    # Create a console handler and set the level based on the input
+    console_handler = logging.StreamHandler()
+    if log_level == "DEBUG":
+        console_handler.setLevel(logging.DEBUG)
+    elif log_level == "INFO":
+        console_handler.setLevel(logging.INFO)
+    elif log_level == "WARNING":
+        console_handler.setLevel(logging.WARNING)
+    elif log_level == "ERROR":
+        console_handler.setLevel(logging.ERROR)
+    else:
+        raise ValueError("Invalid log level provided")
+
+    # Create a formatter and set it for the console handler
+    formatter = logging.Formatter(
+        "%(asctime)s - %(levelname)s [%(name)s] - %(message)s",
+        datefmt="%Y-%m-%d %H:%M:%S",
+    )
+    console_handler.setFormatter(formatter)
+
+    # Add the console handler to the logger
+    logger.addHandler(console_handler)
+
+    return logger