nadeeshamedagama/RAG Milvus Vision Pipeline
Production RAG app using Azure OpenAI, Milvus, LangGraph, SOLID principles.
View on GitHubTrust Signals
- Scorecard Score
- not yet scored
- Maintenance Recency
- Stale
- License
- None
Inputs
| name | description | required | default |
|---|---|---|---|
| milvus-uri | Milvus Cloud cluster URI | yes | — |
| milvus-token | Milvus Cloud API token | yes | — |
| milvus-collection-name | Name of the Milvus collection | no | rag_collection |
| azure-openai-api-key | Azure OpenAI API key for embeddings | yes | — |
| azure-openai-endpoint | Azure OpenAI endpoint URL | yes | — |
| azure-openai-embedding-deployment | Azure OpenAI embedding deployment name | no | text-embedding-ada-002 |
| azure-openai-api-version | Azure OpenAI API version | no | 2024-02-01 |
| embedding-dimension | Dimension of the embedding vectors | no | 1536 |
| github-repo-url | GitHub repository URL to process (optional) | no | "" |
| data-directory | Path to local data directory with diagrams and documents | no | ./data/diagrams |
| google-vision-credentials-json | Google Vision API service account JSON (base64 encoded) | no | "" |
| google-vision-max-results | Maximum results for Google Vision API analysis | no | 10 |
| skip-existing | Skip documents that already exist in the vector store | no | true |
| force-reprocess | Force reprocessing of all documents | no | false |
| mode | Processing mode: ingest (run pipeline) or query (start API server) | no | ingest |
| query-text | Query text for retrieval (only used in query mode) | no | "" |
| top-k | Number of top results to return in query mode | no | 5 |
Outputs
| name | description |
|---|---|
| status | Pipeline execution status (success/failed) |
| documents-processed | Number of documents processed |
| embeddings-created | Number of embeddings created and stored |
| collection-name | Name of the Milvus collection used |
| query-results | Query results in JSON format (only in query mode) |