mirror of
https://github.com/perstarkse/minne.git
synced 2026-03-26 19:31:32 +01:00
neo4j storage wip
This commit is contained in:
Per Stark
221
src/utils/llm.rs
221
src/utils/llm.rs
@@ -1,165 +1,98 @@
|
||||
use async_openai::types::ChatCompletionRequestSystemMessage;
|
||||
use async_openai::types::ChatCompletionRequestUserMessage;
|
||||
use async_openai::types::CreateChatCompletionRequestArgs;
|
||||
use tracing::debug;
|
||||
use tracing::info;
|
||||
use crate::models::text_content::ProcessingError;
|
||||
use serde_json::json;
|
||||
use crate::models::text_content::AnalysisResult;
|
||||
|
||||
/// Placeholder for your actual graph summary retrieval logic
|
||||
async fn get_graph_summary() -> Result<String, ProcessingError> {
|
||||
// Implement your logic to fetch and summarize the graph database
|
||||
Ok("Current graph contains documents related to AI, Rust programming, and asynchronous systems.".into())
|
||||
}
|
||||
|
||||
/// Sends text to an LLM for analysis with enhanced functionality.
|
||||
/// Sends text to an LLM for analysis.
|
||||
pub async fn create_json_ld(category: &str, instructions: &str, text: &str) -> Result<AnalysisResult, ProcessingError> {
|
||||
let client = async_openai::Client::new();
|
||||
|
||||
// Fetch the graph summary
|
||||
let graph_summary = get_graph_summary().await?;
|
||||
|
||||
let schema = json!({
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"knowledge_sources": {
|
||||
"type": "array",
|
||||
"items": {
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"id": {"type": "string"},
|
||||
"type": {"type": "string", "enum": ["Document", "Page", "TextSnippet"]},
|
||||
"title": {"type": "string"},
|
||||
"description": {"type": "string"},
|
||||
"relationships": {
|
||||
"type": "array",
|
||||
"items": {
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"type": {"type": "string", "enum": ["RelatedTo", "RelevantTo", "SimilarTo"]},
|
||||
"target": {"type": "string", "description": "ID of the related knowledge source"}
|
||||
},
|
||||
"required": ["type", "target"],
|
||||
"additionalProperties": false,
|
||||
}
|
||||
}
|
||||
},
|
||||
"required": ["id", "type", "title", "description", "relationships"],
|
||||
"additionalProperties": false,
|
||||
}
|
||||
let client = async_openai::Client::new();
|
||||
let schema = json!({
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"knowledge_source": {
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"id": {"type": "string"},
|
||||
"type": {"type": "string", "enum": ["Document", "Page", "TextSnippet"]},
|
||||
"title": {"type": "string"},
|
||||
"description": {"type": "string"},
|
||||
"relationships": {
|
||||
"type": "array",
|
||||
"items": {
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"type": {"type": "string", "enum": ["RelatedTo", "RelevantTo", "SimilarTo"]},
|
||||
"target": {"type": "string", "description": "ID of the related knowledge source"}
|
||||
},
|
||||
"required": ["type", "target"],
|
||||
"additionalProperties": false,
|
||||
}
|
||||
}
|
||||
},
|
||||
"required": ["id", "type", "title", "description", "relationships"],
|
||||
"additionalProperties": false,
|
||||
},
|
||||
"category": {"type": "string"},
|
||||
"instructions": {"type": "string"}
|
||||
},
|
||||
"required": ["knowledge_sources", "category", "instructions"],
|
||||
"additionalProperties": false
|
||||
});
|
||||
},
|
||||
"required": ["knowledge_source", "category", "instructions"],
|
||||
"additionalProperties": false
|
||||
});
|
||||
|
||||
let response_format = async_openai::types::ResponseFormat::JsonSchema {
|
||||
json_schema: async_openai::types::ResponseFormatJsonSchema {
|
||||
description: Some("Structured analysis of the submitted content".into()),
|
||||
name: "content_analysis".into(),
|
||||
schema: Some(schema),
|
||||
strict: Some(true),
|
||||
},
|
||||
};
|
||||
let response_format = async_openai::types::ResponseFormat::JsonSchema {
|
||||
json_schema: async_openai::types::ResponseFormatJsonSchema {
|
||||
description: Some("Structured analysis of the submitted content".into()),
|
||||
name: "content_analysis".into(),
|
||||
schema: Some(schema),
|
||||
strict: Some(true),
|
||||
},
|
||||
};
|
||||
|
||||
// Construct examples to guide the LLM
|
||||
let system_message = format!(
|
||||
"You are an expert document analyzer. You will receive a document's text content, user instructions, a category, and a summary of the current graph database. Your task is to provide a structured JSON-LD object representing the content, a moderately short description of the document, how it relates to the submitted category and any relevant instructions. You shall also include related objects. The goal is to insert your output into a graph database.
|
||||
// Construct the system and user messages
|
||||
let system_message = format!(
|
||||
"You are an expert document analyzer. You will receive a document's text content, along with user instructions and a category. Your task is to provide a structured JSON-LD object representing the content, a moderately short description of the document, how it relates to the submitted category and any relevant instructions. You shall also include related objects. The goal is to insert your output into a graph database."
|
||||
);
|
||||
|
||||
Here are examples of the desired output:
|
||||
let user_message = format!(
|
||||
"Category: {}\nInstructions: {}\nContent:\n{}",
|
||||
category, instructions, text
|
||||
);
|
||||
|
||||
Example 1:
|
||||
{{
|
||||
\"knowledge_sources\": [
|
||||
{{
|
||||
\"id\": \"ai_neural_networks\",
|
||||
\"type\": \"Document\",
|
||||
\"title\": \"Understanding Neural Networks\",
|
||||
\"description\": \"An in-depth analysis of neural networks and their applications in machine learning.\",
|
||||
\"relationships\": [
|
||||
{{
|
||||
\"type\": \"RelatedTo\",
|
||||
\"target\": \"ai_machine_learning\"
|
||||
}},
|
||||
{{
|
||||
\"type\": \"SimilarTo\",
|
||||
\"target\": \"ai_deep_learning\"
|
||||
}}
|
||||
]
|
||||
}}
|
||||
],
|
||||
\"category\": \"ai\",
|
||||
\"instructions\": \"Analyze the document and relate it to existing AI knowledge.\"
|
||||
}}
|
||||
// Build the chat completion request
|
||||
let request = CreateChatCompletionRequestArgs::default()
|
||||
.model("gpt-4o-mini")
|
||||
.max_tokens(2048u32)
|
||||
.messages([
|
||||
ChatCompletionRequestSystemMessage::from(system_message).into(),
|
||||
ChatCompletionRequestUserMessage::from(user_message).into(),
|
||||
])
|
||||
.response_format(response_format)
|
||||
.build().map_err(|e| ProcessingError::LLMError(e.to_string()))?;
|
||||
|
||||
Example 2:
|
||||
{{
|
||||
\"knowledge_sources\": [
|
||||
{{
|
||||
\"id\": \"rust_async_programming\",
|
||||
\"type\": \"Document\",
|
||||
\"title\": \"Asynchronous Programming in Rust\",
|
||||
\"description\": \"A comprehensive guide to writing asynchronous code in Rust using async/await syntax.\",
|
||||
\"relationships\": [
|
||||
{{
|
||||
\"type\": \"RelatedTo\",
|
||||
\"target\": \"rust_concurrency\"
|
||||
}},
|
||||
{{
|
||||
\"type\": \"SimilarTo\",
|
||||
\"target\": \"rust_multithreading\"
|
||||
}}
|
||||
]
|
||||
}}
|
||||
],
|
||||
\"category\": \"rust\",
|
||||
\"instructions\": \"Incorporate the document into the Rust programming knowledge base.\"
|
||||
}}
|
||||
// Send the request to OpenAI
|
||||
let response = client.chat().create(request).await.map_err(|e| {
|
||||
ProcessingError::LLMError(format!("OpenAI API request failed: {}", e.to_string()))
|
||||
})?;
|
||||
|
||||
Please ensure the IDs follow the format <category>_<short_description> using snake_case."
|
||||
);
|
||||
info!("{:?}", response);
|
||||
|
||||
let user_message = format!(
|
||||
"Graph Summary: {}\nCategory: {}\nInstructions: {}\nContent:\n{}",
|
||||
graph_summary, category, instructions, text
|
||||
);
|
||||
|
||||
// Build the chat completion request
|
||||
let request = CreateChatCompletionRequestArgs::default()
|
||||
.model("gpt-4o-mini") // Ensure this is the correct model identifier
|
||||
.max_tokens(2048u32)
|
||||
.messages([
|
||||
ChatCompletionRequestSystemMessage::from(system_message).into(),
|
||||
ChatCompletionRequestUserMessage::from(user_message).into(),
|
||||
])
|
||||
.response_format(response_format)
|
||||
.build()
|
||||
.map_err(|e| ProcessingError::LLMError(e.to_string()))?;
|
||||
|
||||
// Send the request to OpenAI
|
||||
let response = client.chat().create(request).await.map_err(|e| {
|
||||
ProcessingError::LLMError(format!("OpenAI API request failed: {}", e.to_string()))
|
||||
})?;
|
||||
|
||||
debug!("{:?}", response);
|
||||
|
||||
// Extract and parse the response
|
||||
for choice in response.choices {
|
||||
if let Some(content) = choice.message.content {
|
||||
let analysis: AnalysisResult = serde_json::from_str(&content).map_err(|e| {
|
||||
ProcessingError::LLMError(format!(
|
||||
"Failed to parse LLM response into LLMAnalysis: {}",
|
||||
e.to_string()
|
||||
))
|
||||
})?;
|
||||
return Ok(analysis);
|
||||
// Extract and parse the response
|
||||
for choice in response.choices {
|
||||
if let Some(content) = choice.message.content {
|
||||
let analysis: AnalysisResult = serde_json::from_str(&content).map_err(|e| {
|
||||
ProcessingError::LLMError(format!(
|
||||
"Failed to parse LLM response into LLMAnalysis: {}",
|
||||
e.to_string()
|
||||
))
|
||||
})?;
|
||||
return Ok(analysis);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
Err(ProcessingError::LLMError(
|
||||
"No content found in LLM response".into(),
|
||||
))
|
||||
}
|
||||
Err(ProcessingError::LLMError(
|
||||
"No content found in LLM response".into(),
|
||||
))
|
||||
}
|
||||
|
||||
Reference in New Issue
Block a user