mindnet/scripts/import_markdown.py
Lars 80a7db0e54
Some checks failed
Deploy mindnet to llm-node / deploy (push) Failing after 1s
scripts/import_markdown.py aktualisiert
2025-09-06 15:42:10 +02:00

323 lines
12 KiB
Python

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Script: scripts/import_markdown.py
Version: 0.7.1 (2025-09-06)
Autor: mindnet / Architektur Datenimporte & Sync
Kurzbeschreibung
---------------
Importiert Markdown-Notizen aus einem Obsidian-ähnlichen Vault in Qdrant:
- Validiert Frontmatter / Note-Payload (gegen note.schema.json).
- Chunking + Embeddings.
- Leitet Edges direkt beim Import aus [[Wikilinks]] ab:
- 'references' (Note→Note)
- 'references_at' (Chunk→Note)
- 'backlink' (Note←Note) (symmetrisch zu 'references')
- Idempotente Upserts (deterministische IDs über qdrant_points).
Neu in 0.7.1
------------
- Korrekte Änderungsdetektion via SHA-256 über die **komplette Datei** (Frontmatter+Body):
- Feld: payload.hash_fulltext
- Vergleicht neuen Hash gegen bestehenden Hash in Qdrant.
- Nur bei Änderung → Verarbeitung/Upsert; sonst "skip".
- `--purge-before-upsert` wird **nur** ausgeführt, wenn sich die Note **wirklich geändert** hat.
- Robuste Qdrant-Scroll-Kompatibilität (2- oder 3-Tupel Rückgaben).
Aufrufbeispiele
---------------
Dry-Run (nur prüfen, nichts schreiben):
python3 -m scripts.import_markdown --vault ./vault
Nur eine spezifische Note:
python3 -m scripts.import_markdown --vault ./vault --note-id 20250821-foo
Apply (schreiben) mit Purge (nur geänderte Noten werden bereinigt + neu geschrieben):
python3 -m scripts.import_markdown --vault ./vault --apply --purge-before-upsert
Parameter
---------
--vault PATH : Pflicht. Root-Verzeichnis des Vaults.
--apply : Wenn gesetzt, werden Upserts durchgeführt (sonst Dry-Run).
--purge-before-upsert : Vor Upsert alte Chunks/Edges der **geänderten** Note löschen.
--note-id ID : Optional, verarbeitet nur diese eine Note.
Umgebungsvariablen (.env)
-------------------------
QDRANT_URL, QDRANT_API_KEY, COLLECTION_PREFIX, VECTOR_DIM
Defaults: url=http://127.0.0.1:6333, prefix=mindnet, dim=384
Kompatibilität
--------------
- Bestehende Kernmodule werden weiterverwendet:
app.core.parser (read_markdown, normalize_frontmatter, validate_required_frontmatter)
app.core.validate_note (validate_note_payload)
app.core.chunker (assemble_chunks)
app.core.chunk_payload (make_chunk_payloads)
app.core.embed (embed_texts)
app.core.qdrant (QdrantConfig, ensure_collections, get_client, collection_names)
app.core.qdrant_points (points_for_note, points_for_chunks, points_for_edges, upsert_batch)
app.core.derive_edges (build_note_index, derive_wikilink_edges)
Hinweise
--------
- Bitte im aktivierten venv laufen lassen: source .venv/bin/activate
"""
from __future__ import annotations
import argparse
import glob
import hashlib
import json
import os
import sys
from typing import List, Dict, Tuple, Optional
from dotenv import load_dotenv
from qdrant_client.http import models as rest
from qdrant_client import QdrantClient
# Kern-Bausteine (aus eurem Projekt)
from app.core.parser import (
read_markdown,
normalize_frontmatter,
validate_required_frontmatter,
)
from app.core.validate_note import validate_note_payload
from app.core.chunker import assemble_chunks
from app.core.chunk_payload import make_chunk_payloads
from app.core.embed import embed_texts
from app.core.qdrant import QdrantConfig, ensure_collections, get_client, collection_names
from app.core.qdrant_points import (
points_for_note,
points_for_chunks,
points_for_edges,
upsert_batch,
)
from app.core.derive_edges import build_note_index, derive_wikilink_edges
# -------------------------------------------------
# Hilfsfunktionen
# -------------------------------------------------
def iter_md(root: str) -> List[str]:
patterns = ["**/*.md", "*.md"]
out: List[str] = []
for p in patterns:
out.extend(glob.glob(os.path.join(root, p), recursive=True))
# de-dupe + sort
return sorted(list(dict.fromkeys(out)))
def file_sha256(path: str) -> str:
"""SHA256 über die **Rohdatei** (Frontmatter + Body)."""
h = hashlib.sha256()
with open(path, "rb") as f:
for chunk in iter(lambda: f.read(8192), b""):
h.update(chunk)
return h.hexdigest()
def qdrant_scroll(client: QdrantClient, **kwargs) -> Tuple[List, Optional[str]]:
"""
Wrapper, der sowohl (points, next_offset) als auch (points, next_page, _) Signaturen abdeckt.
"""
res = client.scroll(**kwargs)
if isinstance(res, tuple):
if len(res) == 2:
return res[0], res[1]
if len(res) >= 3:
return res[0], res[1]
return res, None
def fetch_existing_note_hash(client: QdrantClient, prefix: str, note_id: str) -> Optional[str]:
"""Liest hash_fulltext aus mindnet_notes.payload für eine Note."""
notes_col, _, _ = collection_names(prefix)
f = rest.Filter(must=[rest.FieldCondition(key="note_id", match=rest.MatchValue(value=note_id))])
pts, _ = qdrant_scroll(
client,
collection_name=notes_col,
scroll_filter=f,
with_payload=True,
with_vectors=False,
limit=1
)
if not pts:
return None
pl = getattr(pts[0], "payload", {}) or {}
return pl.get("hash_fulltext")
def make_note_stub(abs_path: str, vault_root: str) -> Dict:
"""Minimaler Stub (id/title/path) für build_note_index."""
parsed = read_markdown(abs_path)
fm = normalize_frontmatter(parsed.frontmatter or {})
if "id" not in fm or not fm["id"]:
raise ValueError(f"Missing id in frontmatter: {abs_path}")
rel = os.path.relpath(abs_path, vault_root)
return {"note_id": fm["id"], "title": fm.get("title"), "path": rel}
def build_vault_index(vault_root: str) -> tuple[Dict, Dict, Dict]:
"""Index für robuste Wikilink-Auflösung über alle Noten im Vault."""
stubs = []
for p in iter_md(vault_root):
try:
stubs.append(make_note_stub(p, vault_root))
except Exception:
continue
return build_note_index(stubs)
def purge_for_note(client: QdrantClient, prefix: str, note_id: str, chunk_ids: List[str]) -> None:
"""
Selektives Purge der alten Daten für **diese** Note:
- löscht Chunks (payload.note_id == note_id)
- löscht Edges, deren source_id == note_id ODER in chunk_ids
Notes selbst werden nicht gelöscht (Upsert reicht).
"""
notes_col, chunks_col, edges_col = collection_names(prefix)
# Chunks löschen
f_chunks = rest.Filter(must=[rest.FieldCondition(key="note_id", match=rest.MatchValue(value=note_id))])
client.delete(collection_name=chunks_col, points_selector=f_chunks, wait=True)
# Edges löschen (OR über source_id Werte)
should_conds = [rest.FieldCondition(key="source_id", match=rest.MatchValue(value=note_id))]
for cid in chunk_ids:
should_conds.append(rest.FieldCondition(key="source_id", match=rest.MatchValue(value=cid)))
if should_conds:
f_edges = rest.Filter(should=should_conds)
client.delete(collection_name=edges_col, points_selector=f_edges, wait=True)
# -------------------------------------------------
# Main
# -------------------------------------------------
def main():
ap = argparse.ArgumentParser()
ap.add_argument("--vault", required=True, help="Pfad zum Vault-Root")
ap.add_argument("--apply", action="store_true", help="Schreibt in Qdrant (sonst Dry-Run)")
ap.add_argument(
"--purge-before-upsert",
action="store_true",
help="Vor Upsert alte Chunks/Edges **nur für geänderte** Noten löschen.",
)
ap.add_argument("--note-id", help="Optional: nur diese Note verarbeiten")
args = ap.parse_args()
load_dotenv()
cfg = QdrantConfig(
url=os.getenv("QDRANT_URL", "http://127.0.0.1:6333"),
api_key=os.getenv("QDRANT_API_KEY", None),
prefix=os.getenv("COLLECTION_PREFIX", "mindnet"),
dim=int(os.getenv("VECTOR_DIM", "384")),
)
client = get_client(cfg)
ensure_collections(client, cfg.prefix, cfg.dim)
vault_root = os.path.abspath(args.vault)
files = iter_md(vault_root)
if not files:
print("Keine Markdown-Dateien gefunden.", file=sys.stderr)
sys.exit(2)
# Index einmal bauen (für Linkauflösung bei geänderten Noten)
note_index = build_vault_index(vault_root)
processed = 0
for abs_path in files:
parsed = read_markdown(abs_path)
fm = normalize_frontmatter(parsed.frontmatter or {})
try:
validate_required_frontmatter(fm)
except Exception:
# unvollständige Note überspringen
continue
if args.note_id and fm.get("id") != args.note_id:
continue
processed += 1
# Änderungsdetektion (Datei-Hash vs. Qdrant)
new_hash = file_sha256(abs_path)
old_hash = fetch_existing_note_hash(client, cfg.prefix, fm["id"])
changed = (old_hash != new_hash)
decision = "skip"
if changed:
decision = "apply" if args.apply else "dry-run"
# Bei "skip" kein teures Chunking/Embedding/Edges nötig
if not changed:
print(json.dumps({
"note_id": fm["id"],
"title": fm.get("title"),
"changed": False,
"decision": "skip",
"path": os.path.relpath(abs_path, vault_root),
}, ensure_ascii=False))
continue
# --- Ab hier: Nur für geänderte Noten ---
# Note-Payload erzeugen
from app.core.note_payload import make_note_payload # lazy import
note_pl = make_note_payload(parsed, vault_root=vault_root)
# Hash im Payload mitschreiben (Schema erlaubt hash_fulltext)
note_pl["hash_fulltext"] = new_hash
validate_note_payload(note_pl)
# Chunking & Payloads
chunks = assemble_chunks(fm["id"], parsed.body, fm.get("type", "concept"))
chunk_pls = make_chunk_payloads(fm, note_pl["path"], chunks)
# Embeddings
texts = [c.get("text") or c.get("content") or "" for c in chunk_pls]
vectors = embed_texts(texts)
# Edges ableiten (Note-/Chunk-Level)
edges = derive_wikilink_edges(note_pl, chunk_pls, note_index)
# Purge (nur wenn apply + Option gesetzt)
if args.apply and args.purge_before_upsert:
# Chunk-IDs bestimmen (für Edge-Purge by source_id)
chunk_ids = []
for i, ch in enumerate(chunk_pls, start=1):
cid = ch.get("chunk_id") or ch.get("id") or f"{fm['id']}#{i}"
ch["chunk_id"] = cid # sicherstellen
chunk_ids.append(cid)
purge_for_note(client, cfg.prefix, fm["id"], chunk_ids)
# Upserts (nur Apply)
if args.apply:
notes_col, note_pts = points_for_note(cfg.prefix, note_pl, note_vec=None, dim=cfg.dim)
upsert_batch(client, notes_col, note_pts)
chunks_col, chunk_pts = points_for_chunks(cfg.prefix, chunk_pls, vectors)
upsert_batch(client, chunks_col, chunk_pts)
edges_col, edge_pts = points_for_edges(cfg.prefix, edges)
upsert_batch(client, edges_col, edge_pts)
# Logging geänderte Note
print(json.dumps({
"note_id": fm["id"],
"title": fm.get("title"),
"chunks": len(chunk_pls),
"edges": len(edges),
"changed": True,
"decision": decision,
"path": note_pl["path"],
}, ensure_ascii=False))
print(f"Done. Processed notes: {processed}")
if __name__ == "__main__":
main()