NVIDIA / NVIDIA/NeMo-Retriever
[FEA]: Add Task handler for storing metadata outputs to a persistent object database
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 3k
- Forks
- 349
- Avg merge
- 1d 23h
- Merged PRs (30d)
- 116
Description
Is this a new feature, an improvement, or a change to existing functionality?
New Feature
How would you describe the priority of this feature request
Currently preventing usage
Please provide a clear description of problem this feature solves
We have large JSON formatted metadata objects that are produced by our pipeline. To improve our debugging and tracing abilities, we need to add a configuration option to submitted jobs that allows all resulting metadata objects to be pushed to a storage database. This task should occur out of band, after the pipeline sink.
Describe the feature, and optionally a solution or implementation and any alternatives
Requirements
-
Configuration Option:
- Add a new task to enable storing metadata objects a database.
-
Database Integration:
- Integrate the pipeline with a sample object database.
-
Metadata Handling:
- Handle exporting all payload artifacts as independent items to the database
- JSON Metadata should be stored as a JSON field
-
Error Handling and Logging:
- Ensure that storage tasks occur prior to metric/telemetry offload and are traced correctly
- Implement robust error handling to manage database connection issues and data storage failures.
Additional context
Pipeline Flow
graph TD
A[Job Submission] --> B[Pipeline Processing]
B --> C[Generate Metadata]
C --> D[Pipeline Sink]
D --> E{Configuration Option}
E -->|Enabled| F[Store in Object DB]
E -->|Disabled| G[Skip Object DB storage]
F --> H[Telemetry Export]
G --> H[Telemetry Export]
H --> J[End of Pipeline]
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
No files, tests, or entry points are named. Start by locating the job submission, pipeline sink, metadata generation, and telemetry export stages, then trace how configuration is passed through them. Done means an optional storage task exports payload artifacts and JSON metadata to a chosen object database before telemetry, with connection and storage failures handled and traced.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- backend
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100