Add Ollama LLM Client Support #62
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This commit introduces support for the Ollama LLM provider, including configuration options in the environment variables, constants, and the LLM manager. The Ollama client is implemented to handle requests and responses, ensuring proper JSON formatting and schema compliance. Additionally, the necessary constants and prompts for Ollama are defined to facilitate its integration with the existing system.
This commit enhances the MySQL prompt with detailed instructions and rules for generating SQL queries, ensuring compliance with schema requirements and improving user interaction through structured JSON responses.
This commit enhances the Ollama prompts by adding support for YugabyteDB and MySQL, including specific response schemas and optimization guidelines tailored for each database type. The changes improve the assistant's ability to generate accurate and efficient SQL queries while ensuring compliance with the defined schema.
This commit adds support for Clickhouse and MongoDB in the Ollama prompts, including specific response schemas and detailed guidelines for generating SQL and MongoDB queries. The enhancements improve the assistant's ability to provide accurate and efficient responses while ensuring compliance with the defined schema for both database types.
This commit updates the README.md to accurately reflect the current support status for the Ollama LLM client, removing it from the "Planned to be supported LLM Clients" section and adding it to the "Supported LLM Clients" section. Additionally, a minor grammatical correction was made in the contributing guidelines.
|
@SumonRayy is attempting to deploy a commit to the Aurivance Technologies Team on Vercel. A member of the Team first needs to authorize it. |
Owner
|
Hi @SumonRayy I will review this soon. |
…and YugabyteDB system prompts This commit extends the Ollama prompt functionality by adding support for PostgreSQL, MySQL, Clickhouse, MongoDB, and YugabyteDB. Each database type now has a corresponding prompt, enhancing the assistant's ability to generate accurate and efficient SQL queries while ensuring compliance with the defined schemas.
bhaskarblur
reviewed
May 28, 2025
| ollamaMessages := make([]OllamaMessage, 0) | ||
|
|
||
| // Add system message first with explicit JSON formatting instruction | ||
| systemPrompt = systemPrompt + "\n\nCRITICAL INSTRUCTION: You MUST respond with ONLY a valid JSON object that strictly follows the schema above. Your response MUST include all required fields: assistantMessage, queries (array), and optionally actionButtons. Do not include any other text, markdown, or HTML in your response. Your entire response must be a single JSON object starting with { and ending with }. Do not include any explanations or additional text." |
Owner
There was a problem hiding this comment.
Please inject this hardcode prompt from constants of ollama only rather than writing here.
bhaskarblur
reviewed
May 28, 2025
| // Add a final instruction message to reinforce JSON formatting | ||
| ollamaMessages = append(ollamaMessages, OllamaMessage{ | ||
| Role: "system", | ||
| Content: "Remember: Your response must be ONLY a valid JSON object with all required fields: assistantMessage, queries (array), and optionally actionButtons. Do not include any other text or explanations.", |
Owner
There was a problem hiding this comment.
Same comment, pick from constants
bhaskarblur
reviewed
May 28, 2025
| Stream: false, | ||
| Options: OllamaOptions{ | ||
| Temperature: c.temperature, | ||
| NumPredict: c.maxCompletionTokens, |
Owner
There was a problem hiding this comment.
Is NumPredict = maxCompletionTokens?
Is this correct?
bhaskarblur
reviewed
May 28, 2025
| for _, q := range queries { | ||
| if queryStr, ok := q.(string); ok { | ||
| queryInfos = append(queryInfos, constants.QueryInfo{ | ||
| Query: queryStr, |
Owner
There was a problem hiding this comment.
why hardcoded these values and not from llm response? Also many properties are missing in this append.
Owner
|
Please also update the docker-compose folder files too like env, docker compose files for these changes configs. |
Owner
|
Hi @SumonRayy Any news on this? |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
This PR adds support for the Ollama LLM client as one of the supported language models in NeoBase. This enhancement enables users to run NeoBase queries locally without requiring external API keys, supporting offline and secure environments.
Changes
Database-Specific Optimizations
Clickhouse
YugabyteDB
MongoDB
Benefits
Testing
Related Issues
Resolves #43
Security Considerations