Pablan, as it stands

Self-hosted knowledge management for SMEs: a split-screen Markdown editor
whose sections an LLM refines while you write, and RAG question answering
over the documents that result. FastAPI + Postgres/pgvector on the back,
SvelteKit on the front, everything OpenAI-compatible and self-hostable.

Squashed into a single commit; the development history stays local.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CA43ZJda8Rbp2hKXNy8f6b

backend/tests/test_models.py

@@ -0,0 +1,74 @@
+import pytest
+from sqlalchemy import select, text
+from sqlalchemy.exc import IntegrityError
+from sqlalchemy.ext.asyncio import AsyncSession
+
+from app.models import (
+    EMBEDDING_DIM,
+    Chunk,
+    Document,
+    DocumentStatus,
+)
+
+
+async def _make_document(db: AsyncSession, content: str) -> Document:
+    document = Document(
+        title="Server maintenance",
+        status=DocumentStatus.published,
+        content_md=content,
+    )
+    db.add(document)
+    await db.flush()
+    return document
+
+
+async def test_chunk_tsv_is_generated_with_german_config(db: AsyncSession) -> None:
+    document = await _make_document(db, "# Maintenance")
+    chunk = Chunk(
+        document_id=document.id,
+        chunk_index=0,
+        content="The servers are maintained and checked regularly.",
+        embedding=[0.1] * EMBEDDING_DIM,
+    )
+    db.add(chunk)
+    await db.commit()
+
+    # Same word in content and query stems identically under any config;
+    # real German retrieval assertions come with the M4 fixture corpus.
+    matches = (
+        await db.execute(
+            select(Chunk.id).where(
+                text("tsv @@ websearch_to_tsquery('german', 'maintained')")
+            )
+        )
+    ).all()
+    assert len(matches) == 1
+
+
+async def test_chunk_index_unique_per_document(db: AsyncSession) -> None:
+    document = await _make_document(db, "# Duplicate")
+    for _ in range(2):
+        db.add(
+            Chunk(
+                document_id=document.id,
+                chunk_index=0,
+                content="same index",
+                embedding=[0.0] * EMBEDDING_DIM,
+            )
+        )
+    with pytest.raises(IntegrityError):
+        await db.commit()
+
+
+async def test_embedding_dimension_enforced(db: AsyncSession) -> None:
+    document = await _make_document(db, "# Dimension")
+    db.add(
+        Chunk(
+            document_id=document.id,
+            chunk_index=0,
+            content="wrong dimension",
+            embedding=[0.0] * (EMBEDDING_DIM - 1),
+        )
+    )
+    with pytest.raises(Exception, match="expected 1024 dimensions"):
+        await db.commit()