feat(embedding): new embedding process for documents

[rmolinamir]

Summary

Centralized Document Embedding Architecture

Overview

This PR introduces a major refactor that consolidates the embedding system around a unified document-based approach, replacing separate embedding schemas for journals and pages with a centralized DocumentEmbedding system.

Key Changes

Architecture:

• Unified Schema: Replaced separate JournalEmbedding and PageEmbedding schemas with a single DocumentEmbedding schema
• Document: All content (journals, pages) now flows through the unified Document entity for embedding
• tRPC: Split API into apiRouter (user-facing) and embedderRouter (internal embedding operations)

File Changes Summary

• Removed: Old webhook handlers, separate embedding schemas
• Added: New unified embedding system, webhook processor, embedder router
• Refactored: API structure, search tools, and type definitions

Benefits

1. Simplified Architecture: Single embedding table reduces complexity and maintenance overhead
2. Better Performance: Unified schema enables more efficient queries and better indexing
3. Enhanced Reliability: Webhook-driven processing with retry logic and failure handling
4. Improved AI: Richer metadata and semantic chunking enhance search quality
5. Future-Proof: Document-centric approach supports easier extension for new content types

[rmolinamir]

Closes #40.