Can not connect with NeptuneML graph databases from outside AWS.

[msasso-hellohippo]

Checked other resources

• I added a very descriptive title to this question.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.

Commit to Help

• I commit to help with one of those options 👆

Example Code

from langchain_community.graphs.neptune_graph import NeptuneGraph

kg = NeptuneGraph(
    host=os.environ["NEPTUNEML_HOST"],
    port=os.environ["NEPTUNEML_PORT"],   
)

Description

I just created a NeptuneML graph in AWS as this video shows :https://www.youtube.com/watch?v=Xkts629WM6s

And I want to connect with AWS externally from my python project using Langchain.

This is the error I receive:

langchain_community.graphs.neptune_graph.NeptuneQueryException: {'message': 'Could not get schema for Neptune database', 'detail': '{\'message\': \'Summary API is not available for this instance of Neptune,ensure the engine version is >=1.2.1.0\', \'details\': \'Connect timeout on endpoint URL: "https://<my-neptune-host>:<my-neptune-port>/propertygraph/statistics/summary"\'}'}

System Info

boto3==1.34.149
botocore==1.34.149
langchain==0.2.6
langchain-chroma==0.1.2
langchain-community==0.2.6
langchain-core==0.2.11
langchain-openai==0.1.14
langchain-text-splitters==0.2.2
langgraph==0.1.5
langsmith==0.1.90

[dosubot[bot]]

To resolve the connection issue to your NeptuneML graph database from outside AWS, ensure that your Neptune instance supports the Summary API. The error message indicates that the Summary API is not available, suggesting that the engine version might be outdated. Here are the steps to resolve this:

1. Check the Neptune Engine Version: Ensure that your Neptune instance is running an engine version that supports the Summary API (>=1.2.1.0).

2. Upgrade the Neptune Engine Version: If your engine version is outdated, upgrade it to a version that supports the Summary API. You can do this through the Amazon Neptune console by selecting your DB cluster, choosing "Modify," and selecting the appropriate engine version.

3. Verify API Availability: After upgrading, verify that the Summary API is available by running a test query. You can use the AWS CLI command get-propertygraph-summary to check this:

   aws neptune get-propertygraph-summary --endpoint <your-neptune-endpoint>

4. Update Your Code: Ensure your connection settings are correct. Here is an example of how to initialize the NeptuneRdfGraph class and handle the connection:

   from langchain_community.graphs.neptune_rdf_graph import NeptuneRdfGraph
   
   # Initialize the NeptuneRdfGraph instance
   graph = NeptuneRdfGraph(
       host=os.environ["NEPTUNEML_HOST"],
       port=int(os.environ["NEPTUNEML_PORT"]),
       use_https=True,
       credentials_profile_name='my-aws-profile',
       region_name='us-west-2'
   )
   
   # Attempt to refresh the schema to check for Summary API availability
   try:
       graph._refresh_schema()
   except NeptuneQueryException as e:
       print(f"Error: {e.get_message()}")
       print(f"Details: {e.get_details()}")

Ensure that your AWS credentials and network configurations are correctly set up to allow connections from outside AWS ^[1][2].

To continue talking to Dosu, mention @dosu.

[msasso-hellohippo]

Engine version is 1.3.2.1 so here it is ok.

The command aws neptune get-propertygraph-summary is not valid

[dosubot[bot]]

To resolve the issue of connecting to your NeptuneML graph database from outside AWS, given that your engine version is 1.3.2.1, follow these steps:

1. Ensure Correct Connection Parameters:
   Make sure your connection parameters and credentials are correctly set up. Here is a sample code snippet to help you set up the connection:

   from langchain_community.graphs.neptune_graph import NeptuneGraph
   
   # Replace with your actual values
   host = '<your-neptune-host>'
   port = 8182
   use_https = True
   credentials_profile_name = '<your-aws-profile>'
   region_name = '<your-region>'
   
   try:
       graph = NeptuneGraph(
           host=host,
           port=port,
           use_https=use_https,
           credentials_profile_name=credentials_profile_name,
           region_name=region_name
       )
       # Example query
       query = "MATCH (n) RETURN n LIMIT 10"
       results = graph.query(query)
       print(results)
   except NeptuneQueryException as e:
       print(f"Query failed: {e.get_message()} - {e.get_details()}")
   except Exception as e:
       print(f"An error occurred: {str(e)}")

2. Verify API Availability:
   Since the aws neptune get-propertygraph-summary command is not valid, you can verify the Summary API availability by using the _get_summary method in your code:

   def verify_summary_api(graph):
       try:
           summary = graph._get_summary()
           print("Summary API is enabled. Summary:", summary)
       except NeptuneQueryException as e:
           print("Summary API is not enabled or an error occurred:", e.get_message(), e.get_details())
   
   # Verify the Summary API
   verify_summary_api(graph)

3. Check Network Configurations:
   Ensure that your network configurations and security group settings allow external access to the NeptuneML graph database. This includes:

   • Allowing inbound traffic on the necessary port (default is 8182).
   • Ensuring that the security group associated with your Neptune instance allows connections from your external IP address.

4. AWS Credentials and Permissions:
   Ensure that your AWS credentials are correctly configured and have the necessary permissions. You can configure your AWS credentials using the AWS CLI or environment variables.

By following these steps, you should be able to connect to your NeptuneML graph database from outside AWS and verify that the Summary API is enabled ^[1].

To continue talking to Dosu, mention @dosu.

[eschaffn]

this bot stinks

[parvathysarat]

@msasso-hellohippo did you end up fixing this issue? I have the same error and my engine is already upgraded. TIA!

[Boopalanoptisol]

@parvathysarat did u end up fixing the issue?. I am also have the same error.

[msasso-hellohippo]

No I gave up on Neptune ML and I am using NEO4J

[moriokaj]

I use llama-index, but like you, I was getting an error when I ran the API to get the summary.
While researching the aws neptune api, I came across the following statement: “Statistics generation is disabled on t3 and t4 instance types (that is, on db.t3.medium and db.t4g.medium instance types) to conserve memory. As a result, graph summary data is not available either on those instance types.” Changing the instance type of the cluster resolved the error for me.
I hope you will try it too.
https://docs.aws.amazon.com/neptune/latest/userguide/neptune-graph-summary.html

[recovere-hima]

is it possible to load the neptune database from the local system. @moriokaj