[geaflow/ai-memory] Preserve pre-structured graph import as a bypass path

Priority: P1
Difficulty: Starter

Context: geaflow-ai already supports CSV/JSON-like pre-structured graph loading. Phase 1 should not force every user through LLM extraction.

Scope:

- Document and test a bypass path where structured vertices/edges skip extraction but still produce provenance and idempotency metadata.
- Add a small fixture that imports pre-structured graph facts and raw document chunks in the same run.

Constraints:

- Do not rewrite existing CSV reader.
- Do not allow bypassed facts to skip schema validation.

Acceptance Criteria:

- Structured import and extracted facts share the same validation contract.
- Bypass facts include source reference and import mode.

Suggested paths:

- `geaflow-ai/src/main/java/org/apache/geaflow/ai/graph/io`
- `geaflow-ai/src/test/resources/import`

Start by reading the graph I/O code under geaflow-ai/src/main/java/org/apache/geaflow/ai/graph/io and inspect the import resources under geaflow-ai/src/test/resources/import. Add a small mixed-run fixture and tests showing that pre-structured facts use the same schema validation while retaining provenance, source reference, idempotency metadata, and import mode.

Written by the indexing model from the issue text.