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/embedding_stub.py

@@ -0,0 +1,24 @@
+"""Deterministic fake embeddings for indexing/retrieval tests.
+
+Identical text maps to an identical unit vector (cosine distance 0);
+unrelated texts land near-orthogonal. Semantics are exercised by the evals
+against the real endpoint — these tests cover the SQL plumbing.
+"""
+
+import hashlib
+import math
+import random
+
+from app.models import EMBEDDING_DIM
+
+
+def deterministic_embedding(text: str) -> list[float]:
+    seed = hashlib.sha256(text.encode()).digest()
+    rng = random.Random(seed)
+    vector = [rng.uniform(-1.0, 1.0) for _ in range(EMBEDDING_DIM)]
+    norm = math.sqrt(sum(value * value for value in vector))
+    return [value / norm for value in vector]
+
+
+async def fake_embed(texts: list[str], *, role: str = "embedding") -> list[list[float]]:
+    return [deterministic_embedding(text) for text in texts]