will/swarm-master
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat(rag): switch Obsidian endpoint to NPU embeddings

Browse Source
This commit is contained in:
William Valentin
2026-06-03 20:04:43 -07:00
parent bcc652e5aa
commit 1a674e854e
3 changed files with 43 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files
scripts/obsidian-reindex-server.py
+23 -6
View File
@@ -26,14 +26,20 @@ from urllib.parse import parse_qs, urlparse
PORT = int(os.environ.get("PORT", 18810))
REINDEX_TIMEOUT = int(os.environ.get("REINDEX_TIMEOUT", "1800"))
RAG_COLLECTION = os.environ.get("RAG_COLLECTION", "obsidian").strip() or "obsidian"
RAG_EMBED_MODEL = os.environ.get("RAG_EMBED_MODEL", "nomic-embed-text").strip() or "nomic-embed-text"
OLLAMA_BASE_URL = (os.environ.get("OLLAMA_BASE_URL") or "http://127.0.0.1:18807").rstrip("/")
REINDEX_SCRIPT = str(
Path.home()
/ ".hermes/skills/note-taking/rag-search/scripts/reindex_obsidian.sh"
)
STATE_FILE = (
Path.home() / ".hermes/data/rag-search/obsidian_index_state.json"
)
STATE_FILE = Path(
os.environ.get("RAG_STATE_FILE")
or Path.home() / ".hermes/data/rag-search" / (
"obsidian_index_state.json" if RAG_COLLECTION == "obsidian" else f"{RAG_COLLECTION}_index_state.json"
)
).expanduser()
SEARCH_SCRIPT = str(Path.home() / ".hermes/skills/note-taking/rag-search/scripts/search.py")
VENV_PYTHON = str(Path.home() / ".hermes/skills/note-taking/rag-search/venv/bin/python")
@@ -50,11 +56,16 @@ def run_reindex(full: bool = False) -> dict:
cmd = [REINDEX_SCRIPT]
if full:
cmd.append("--full")
env = os.environ.copy()
env.setdefault("RAG_COLLECTION", RAG_COLLECTION)
env.setdefault("RAG_EMBED_MODEL", RAG_EMBED_MODEL)
env.setdefault("OLLAMA_BASE_URL", OLLAMA_BASE_URL)
result = subprocess.run(
cmd,
capture_output=True,
text=True,
timeout=REINDEX_TIMEOUT,
env=env,
)
if result.returncode != 0:
return {
@@ -97,12 +108,16 @@ def run_semantic_search(query: str, top_k: int = 5) -> dict:
if not query:
return {"ok": False, "error": "query is required", "results": []}
top_k = max(1, min(int(top_k or 5), 20))
env = os.environ.copy()
env.setdefault("RAG_COLLECTION", RAG_COLLECTION)
env.setdefault("RAG_EMBED_MODEL", RAG_EMBED_MODEL)
env.setdefault("OLLAMA_BASE_URL", OLLAMA_BASE_URL)
result = subprocess.run(
[
VENV_PYTHON if Path(VENV_PYTHON).exists() else sys.executable,
SEARCH_SCRIPT,
"--index",
"obsidian",
RAG_COLLECTION,
"--top-k",
str(top_k),
"--raw",
@@ -111,6 +126,7 @@ def run_semantic_search(query: str, top_k: int = 5) -> dict:
capture_output=True,
text=True,
timeout=90,
env=env,
)
if result.returncode != 0:
return {
@@ -125,7 +141,7 @@ def run_semantic_search(query: str, top_k: int = 5) -> dict:
return {
"ok": True,
"query": query,
"index": payload.get("index", "obsidian"),
"index": payload.get("index", RAG_COLLECTION),
"top_k": top_k,
"result_count": len(results),
"results": results,
@@ -144,7 +160,8 @@ def semantic_health() -> dict:
"note_count",
"vector_count",
"collection",
"chroma_path",
"embedding_backend",
"embedding_model",
"last_full_index",
"last_incremental_index",
)