Standalone MCP (#346)

* ADB/Readme

added ADB support.
Fixed Readme to run correctly the mcp/inspector

* fix to ensure reachability

* fix OCI model getting issue

* Fixup OCI

---------

Co-authored-by: corradodebari <[email protected]>

Author: gotsysdba

src/client/mcp/rag/.python-version

@@ -4,9 +4,9 @@
 **Version:** *Developer preview*
 
 ## Introduction
-This document describe how to re-use the configuration tested in the **AI Optimizer & Toolkit** an expose it as an MCP tool to a local **Claude Desktop** and how to setup as a remote MCP server. This early draft implementation utilizes the `stdio` and `sse` to interact between the agent dashboard, represented by the **Claude Desktop**, and the tool. 
+This document describe how to re-use the configuration tested in the **AI Optimizer & Toolkit** an expose it as an MCP tool to a local **Claude Desktop** and how to setup as a remote MCP server. This early draft implementation utilizes the `stdio`, `sse` and, by default, the `streamable-http` to interact between the agent dashboard, represented by the **Claude Desktop**, and the tool. 
 
-**NOTICE**: Only `Ollama` or `OpenAI` configurations are currently supported. Full support will come.
+**NOTICE**: Only `Ollama`, `OpenAI`, `hosted_vllm` configurations are currently supported. Full support will come. `DB23ai`, `26ai` and `Autonomous DB` with wallet configuration is also supported. 
 
 ## Pre-requisites.
 You need:
@@ -34,12 +34,14 @@ In the **AI Optimizer & Toolkit** web interface, after tested a configuration, i
 * press button `Download LangchainMCP` to download an VectorSearch MCP Agent built on current configuration.
 * unzip the file in a `<PROJECT_DIR>` dir.
 
+**NOTICE**: if you want to run the application in another server, remember to change in the `optimizer_settings.json` any reference no more local, like hostname for LLM servers, Database, wallet dir and so on.
+
 
 ## Standalone client
 There is a client that you can run without MCP via command-line to test it:
 
 ```bash
-uv run rag_base_optimizer_config.py "[YOUR_QUESTION]"
+uv run rag_base_optimizer_config_direct.py "[YOUR_QUESTION]"
 ```
 In `rag_base_optimizer_config_mcp.py`:
 
@@ -117,15 +119,15 @@ uv run rag_base_optimizer_config_mcp.py
    		}
 	}
 	```
-	* Set `Local` with `Remote client` line in `<PROJECT_DIR>/rag_base_optimizer_config_mcp.py`:
+	* Set `Remote client` line of code with `Local` in `<PROJECT_DIR>/rag_base_optimizer_config_mcp.py`:
 
 	```python
-	#mcp.run(transport='stdio')
-    #mcp.run(transport='sse')
-    mcp.run(transport='streamable-http')
+	# Initialize FastMCP server
+	#mcp = FastMCP("rag", port=9090) #Remote client
+	mcp = FastMCP("rag") #Local
 	```
 
-	* Substitute `stdio` with `streamable-http` line of code:
+	* Substitute  line of code `streamable-http` with  `stdio`:
 
 	```python
 	mcp.run(transport='stdio')
@@ -144,7 +146,7 @@ uv run rag_base_optimizer_config_mcp.py
 * Run the inspector:
 
 ```bash
-npx @modelcontextprotocol/inspector 
+npx @modelcontextprotocol/inspector@0.15.0
 ```
 
 * connect the browser to `http://127.0.0.1:6274` 
@@ -158,37 +160,37 @@ npx @modelcontextprotocol/inspector
 
 
 **Optional:** run with local **stdio** protocol
-* Set as shown before the protolo to run locally in `<PROJECT_DIR>/rag_base_optimizer_config_mcp.py`:
+* Set as shown before the protocol to run locally in `<PROJECT_DIR>/rag_base_optimizer_config_mcp.py`:
 
-	```
 	* Set `Local` with `Remote client` line:
 
 	```python
 	#mcp = FastMCP("rag", port=9090) #Remote client
 	mcp = FastMCP("rag") #Local
 	```
 
-	* Substitute `stdio` with `sse` line of code:
+	* Substitute `streamable-http` with `stdio` line of code:
 	```python
 	mcp.run(transport='stdio')
-	#mcp.run(transport='sse')
+    #mcp.run(transport='sse')
+    #mcp.run(transport='streamable-http')
 	```
 
 * Run the inspector:
 
-```bash
-npx @modelcontextprotocol/inspector uv run rag_base_optimizer_config_mcp.py
-```
+	```bash
+	npx @modelcontextprotocol/inspector@0.15.0 uv run rag_base_optimizer_config_mcp.py
+	```
 
 * connect to the port `http://localhost:6274/` with your browser on the link printed, like in the following example:
-```bash
-..
-Open inspector with token pre-filled:
-http://localhost:6274/?MCP_PROXY_AUTH_TOKEN=cb2ef7521aaf2050ad9620bfb5e5df42dc958889e6e99ce4e9b18003eb93fffd
-..
-```
+	```bash
+	..
+	Open inspector with token pre-filled:
+	http://localhost:6274/?MCP_PROXY_AUTH_TOKEN=cb2ef7521aaf2050ad9620bfb5e5df42dc958889e6e99ce4e9b18003eb93fffd
+	..
+	```
 
-* setup the `Inspector Proxy Address` with `http://127.0.0.1:6277` 
+* setup the Transport Type: `STDIO`
 * test the tool developed.
 
 

src/client/mcp/rag/README.md

@@ -97,9 +97,20 @@ def get_vectorstore(data, embeddings):
     user=db_config["user"]
     password=db_config["password"]
     dsn=db_config["dsn"]
+    
+    # ADB connection with wallet 
+
+    wallet_pwd= db_config["wallet_password"]
+    wallet_location = db_config["wallet_location"]
 
     logger.info(f"{db_table}: {user} - {dsn}")
-    conn23c = oracledb.connect(user=user, password=password, dsn=dsn)
+    
+    if wallet_pwd and wallet_location:
+        logger.info("ADB connection starting..")
+        conn23c = oracledb.connect(user=user, password=password, dsn=dsn,wallet_location=wallet_location,wallet_password=wallet_pwd)
+    else:
+        conn23c = oracledb.connect(user=user, password=password, dsn=dsn)
+
 
     logger.info("DB Connection successful!")
     metric = data["client_settings"]["vector_search"]["distance_metric"]
@@ -109,7 +120,7 @@ def get_vectorstore(data, embeddings):
         dist_strategy = DistanceStrategy.COSINE
     elif metric == "EUCLIDEAN":
         dist_strategy = DistanceStrategy.EUCLIDEAN
-
+    
     logger.info(embeddings)
     knowledge_base = OracleVS(client=conn23c,table_name=db_table, embedding_function=embeddings, distance_strategy=dist_strategy)
 

src/client/mcp/rag/optimizer_utils/config.py

@@ -0,0 +1,57 @@
+"""
+Copyright (c) 2024, 2025, Oracle and/or its affiliates.
+Licensed under the Universal Permissive License v1.0 as shown at http://oss.oracle.com/licenses/upl.
+"""
+import sys
+from typing import List
+import os
+from dotenv import load_dotenv
+import logging
+logger = logging.getLogger(__name__)
+
+logging.basicConfig(
+    level=logging.INFO,
+    format="%(name)s - %(levelname)s - %(message)s"
+)
+
+
+logger.info("Successfully imported libraries and modules")
+
+from optimizer_utils import config
+
+from optimizer_utils import rag
+
+CHUNKS_DIR = "chunks_temp"
+data = {}
+
+def similarity_search(question: str, max_results: int = 5) -> List[str]:
+    """
+    Use this tool to get the top similar information to any question that may benefit from up-to-date or domain-specific information.
+    
+    Args:
+        question: The topic to search for
+        max_results: Maximum number of results to retrieve (default: 5)
+        
+    Returns:
+        List of information related to the question
+    """
+    
+    logger.info(f"Results provided for question: {question} with top {max_results}")
+    chunks=["first chunk", "second chunk"]
+    
+    return chunks
+
+if __name__ == "__main__":
+    # Initialize and run the server
+    # Load JSON file
+    file_path = os.path.join(os.getcwd(), "optimizer_settings.json")
+    logger.info(file_path)
+    rag.set_optimizer_settings_path(file_path)
+
+    if len(sys.argv) > 1:
+        question = sys.argv[1]
+        print(question)
+        logger.info(f"Question: {sys.argv[1]}")
+        logger.info(f"\n\nAgent Answer: {rag.rag_tool_base(question)}")
+    else:
+        logger.info("No question provided.")

src/client/mcp/rag/pyproject.toml

@@ -29,7 +29,7 @@
 data = {}
 
 # Initialize FastMCP server
-mcp = FastMCP("rag", port=9090) #Remote client
+mcp = FastMCP("rag",host="0.0.0.0", port=9090) #Remote client
 #mcp = FastMCP("rag") #Local