Add dashAI document data structure and database model

[Irozuku]

Summary

First of a stacked series that turns the Datasets module into a Data module holding two kinds of data: datasets and documents. This branch only adds the document data structure and its database record; nothing in the UI changes yet.

• DashAIData is a new common base for DashAIDataset and the new DashAIDocument.
• DashAIDocument holds parsed content split by DashAI type (Text, Image) in Arrow tables, keeping the reading order (position) and the source page of every item, so tools that already restrict themselves by DashAI type can work on documents.
• Documents are saved to and loaded from disk (one Arrow file per content type plus document.json), and every path into a document validates its contents.
• A data_document table and the DataDocument model store uploaded documents, with a DocumentStatus enum for their parsing job.

---

Type of Change

Check all that apply like this [x]:

• Backend change
• Frontend change
• CI / Workflow change
• Build / Packaging change
• Bug fix
• Documentation

---

Changes (by file)

• DashAI/back/dataloaders/classes/dashai_data.py: new DashAIData base class shared by datasets and documents (data_kind()).
• DashAI/back/dataloaders/classes/dashai_dataset.py: DashAIDataset now extends DashAIData.
• DashAI/back/dataloaders/classes/dashai_document.py: DocumentItem, DashAIDocument (from_items, items, types, count, as_dataset, with_content), save_document / load_document with atomic writes and strict validation (types, non null content, image bytes, pages from 1, positions unique across types).
• DashAI/back/dependencies/database/models.py: DataDocument model (name, file info and hash, dataloader and params, content counts, pages, folder, job id, status).
• DashAI/back/core/enums/status.py: DocumentStatus.
• DashAI/alembic/versions/a7c3e91b5d20_add_data_document.py: migration creating data_document.
• tests/back/dataloaders/test_dashai_data.py, test_dashai_document.py, test_document_persistence.py: document structure, reading order and save/load round trips.
• tests/back/dependencies/test_data_document_model.py, tests/back/migrations/test_add_data_document.py: model and migration (upgrade, downgrade, ORM parity).

Notes (optional)

• Base branch: develop. The next PR, feat/document-dataloaders, builds on this one.
• Adds migration a7c3e91b5d20; starting the app upgrades the local database.
• The legacy RAG document table is untouched here; it is replaced in feat/rag-data-documents.