[geaflow/ai-memory] Define provenance model for facts, chunks, and evidence

[kitalkuyo-gita]

Priority: P0
Difficulty: Starter

Context: Citation-aware answers and deletion propagation require source lineage. Existing subgraph verbalization does not provide claim-level provenance.

Scope:

• Add ProvenanceRef and SourceSpan contracts.
• Link extracted facts to document, chunk, extractor version, and schema version.
• Add JSON examples.

Constraints:

• Do not implement answer generation.
• Do not store full raw private text inside provenance objects.

Acceptance Criteria:

• Every extracted fact fixture can point back to source document and chunk.
• Missing provenance fails validation for extracted facts.

Suggested paths:

• geaflow-ai/src/main/java/org/apache/geaflow/ai/provenance

[E2ern1ty]

Hi, I'd like to pick up this issue.

Plan:

• Add ProvenanceRef and SourceSpan contracts under geaflow-ai/src/main/java/org/apache/geaflow/ai/provenance, linking facts to document, chunk, extractor version and schema version
• Add JSON examples showing how extracted facts point back to their source document and chunk
• Add validation so that extracted facts missing provenance fail explicitly, without embedding raw document text in provenance objects

I'll open a PR soon. Happy to adjust the contract shape if reviewers prefer a different structure.

[E2ern1ty]

Submitted an initial implementation in #843. Feedback welcome - happy to adjust the contract shape if reviewers prefer a different structure.

[kitalkuyo-gita]

Submitted an initial implementation in #843. Feedback welcome - happy to adjust the contract shape if reviewers prefer a different structure.

Okay, I'm taking a weekend off. I'm glad you're willing to take on this issue; I'll review your PR in the next day or two.