https://github.com/hantmac/langchaingo-ollama-rag
https://github.com/hantmac/langchaingo-ollama-rag
Last synced: 3 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/hantmac/langchaingo-ollama-rag
- Owner: hantmac
- License: apache-2.0
- Created: 2024-05-02T16:38:32.000Z (about 1 year ago)
- Default Branch: main
- Last Pushed: 2024-05-03T02:08:27.000Z (about 1 year ago)
- Last Synced: 2024-05-04T02:52:54.878Z (about 1 year ago)
- Language: Go
- Size: 48.8 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
[
](https://api.gitsponsors.com/api/badge/link?p=H9OBR3Ex9c17MY0dU4Jiu7yYYYRLRj2YhmvqsFSO6LE4DxqYz7afUc045lQHZy97)
# Ollama RAG
Project for adding RAG to ollama models using :
- [langchaingo](https://github.com/tmc/langchaingo)
- [qdrant vector database](https://github.com/qdrant/qdrant)
- [nomic-embed-text](https://ollama.com/library/nomic-embed-text)
- [qwen1.8B](https://github.com/QwenLM/Qwen1.5)
## Installation :
Download packages :
```bash
go mod download
```
Pull Ollama models :
```
ollama pull nomic-embed-text
ollama pull qwen1.8B
```
Install qdrant :
```bash
docker pull qdrant/qdrant
docker run -p 6333:6333 qdrant/qdrant
```
Create Collection for qdrant
use any http client for make a PUT request like example blow for creating a Collection
```bash
curl -X PUT http://localhost:6333/collections/romeo \
-H 'Content-Type: application/json' \
--data-raw '{
"vectors": {
"size": 768,
"distance": "Dot"
}
}'
```
Delete a collection
```bash
curl --location --request DELETE 'http://localhost:6333/collections/romeo'
```
## Using
put you texts in text.txt and run :
``` go run main.go getanswer
```
then ask anything from the text you provided
## Reference
Thanks to https://github.com/shayanfarzi/ollama-rag