`[cco-service]` Implement Background Worker for Metadata Enrichment

[clnsmth]

Description
We need to implement the asynchronous worker to handle high-latency tasks. As defined in ADR-001, the system must support processing large batches (1000+) of citations without blocking the main API thread or timing out the user interface.

Context
When a user uploads a raw list of 1,000 DOIs (via the IdentifierListIngestor from Issue #3), the system creates "hollow" objects. We need a background process to iterate through these IDs, query external APIs (Crossref/DataCite) for metadata (Title, Author, Year), and "enrich" the CCO in the background.

Requirements

• Queue Infrastructure: Set up a task queue (recommend Celery or RQ) that connects to the existing Redis instance.
• Worker Container: Add a service to docker-compose.yml that runs the worker process.
• Enrichment Task: Implement a task (e.g., enrich_cco_metadata(session_id)) that:

1. Retrieves the current CCO from Redis.
2. Iterates through AggregatedResources that are missing metadata.
3. Fetches metadata from public APIs (Crossref/DataCite).
4. Updates the CCO object in Redis.

• Status Tracking: The task should update a status key in Redis (e.g., processing_status: { percent: 50, status: "running" }) so the Frontend can poll for progress.

Technical Scope (Suggested)

• Use the same cco-core library model definitions to ensure data consistency.
• Important: Ensure the worker handles API rate limits (e.g., add rudimentary sleeps or retries if hitting Crossref too hard).
• The API (Issue [cco-core] Implement Repository Adapter Pattern (Zenodo Integration) #4) will need an endpoint to trigger this task, but this issue focuses on the worker infrastructure and the task logic itself.

Acceptance Criteria

• Worker container is running and connected to Redis.
• A "dummy task" can be triggered that writes to logs (to prove infrastructure works).
• The enrich_cco_metadata task successfully fetches a title for a given DOI and saves it back to the Redis CCO object.
• Basic error handling (e.g., what happens if a DOI is invalid?) is implemented so the worker doesn't crash.

References

• ADR-001 Section 2: Container Design (Worker)
• ADR-001 Section 6: Consequences (Async UX)
• Requirements Section 2: Scalability (1000+ citations)