Improve prompts and context

src/pinecone.rs

@@ -34,7 +34,7 @@ pub struct AssistantContext {
 
 #[derive(Debug, Deserialize)]
 pub struct AssistantContextResponse {
- pub snippets: serde_json::Value,
+ pub snippets: Vec<serde_json::Value>,
  pub usage: serde_json::Value,
 }
 
@@ -108,7 +108,7 @@ mod tests {
  .mock("POST", "/assistant/chat/test-assistant/context")
  .with_status(200)
  .with_header("content-type", "application/json")
- .with_body(r#"{"snippets": [{"text": "This is a test response"}], "usage": {"total_tokens": 100}}"#)
+ .with_body(r#"{"snippets": [{"text": "snippet 1"}, {"text": "snippet 2"}], "usage": {"total_tokens": 100}}"#)
  .create();
 
  let client = PineconeClient::new("test-api-key".to_string(), server.url());
@@ -119,7 +119,8 @@ mod tests {
 
  mock.assert();
  let response = result.unwrap();
- assert_eq!(response.snippets[0]["text"], "This is a test response");
+ assert_eq!(response.snippets[0]["text"], "snippet 1");
+ assert_eq!(response.snippets[1]["text"], "snippet 2");
  }
 
  #[tokio::test]

src/router.rs

@@ -1,10 +1,10 @@
 use crate::config::Config;
 use crate::pinecone::{PineconeClient, PineconeError};
+use mcp_server::router::CapabilitiesBuilder;
 use mcp_spec::content::Content;
 use mcp_spec::handler::{PromptError, ResourceError, ToolError};
 use mcp_spec::prompt::Prompt;
 use mcp_spec::{protocol::ServerCapabilities, resource::Resource, tool::Tool};
-use mcp_server::router::CapabilitiesBuilder;
 use serde_json::Value;
 use std::future::Future;
 use std::pin::Pin;
@@ -52,21 +52,24 @@ impl PineconeAssistantRouter {
  client,
  tools: vec![Tool::new(
  TOOL_ASSISTANT_CONTEXT.to_string(),
- "Retrieve context snippets from a Pinecone assistant knowledge base".to_string(),
+ "Retrieves relevant document snippets from your Pinecone Assistant knowledge base. \
+ Returns an array of text snippets from the most relevant documents. \
+ You can use the 'top_k' parameter to control result count (default: 15). \
+ Recommended top_k: a few (5-8) for simple/narrow queries, 10-20 for complex/broad topics.".to_string(),
  serde_json::json!({
  "type": "object",
  "properties": {
  PARAM_ASSISTANT_NAME: {
  "type": "string",
- "description": "Name of the Pinecone assistant"
+ "description": "Name of an existing Pinecone assistant"
  },
  PARAM_QUERY: {
  "type": "string",
- "description": "The query to use for generating context."
+ "description": "The query to retrieve context for."
  },
  PARAM_TOP_K: {
  "type": "integer",
- "description": "Number of context snippets to return. Default is 15."
+ "description": "The number of context snippets to retrieve. Defaults to 15."
  }
  },
  "required": [PARAM_ASSISTANT_NAME, PARAM_QUERY]
@@ -100,7 +103,11 @@ impl PineconeAssistantRouter {
  .await?;
 
  tracing::info!("Successfully received response from Pinecone API");
- Ok(vec![Content::text(response.snippets.to_string())])
+ Ok(response
+ .snippets
+ .iter()
+ .map(|snippet| Content::text(snippet.to_string()))
+ .collect())
  }
 }
 
@@ -111,14 +118,9 @@ impl mcp_server::Router for PineconeAssistantRouter {
 
  fn instructions(&self) -> String {
  format!(
- "This server provides tools to interact with Pinecone's Assistant API. \
- The {TOOL_ASSISTANT_CONTEXT} tool allows you to retrieve relevant context snippets from given knowledge base assistants. \
- Returned value structure: \
- An array of relevant document snippets, each containing: \
- - content: The text content of the snippet \
- - score: A relevance score (float) \
- - reference: Source information including document ID and location \
- You can tune the number of returned snippets by setting the `top_k` parameter."
+ "This server connects to an existing Pinecone Assistant,\
+ a RAG system for retrieving relevant document snippets. \
+ Use the {TOOL_ASSISTANT_CONTEXT} tool to access contextual information from its knowledge base"
  )
  }