mindnet/app/core/note_payload.py

"""
note_payload.py — Mindnet core payload builder (v0.5, 2025-11-08)

Purpose
-------
Builds a **JSON-serializable** payload dict for a single note to be stored in
the `<prefix>_notes` collection. The function is defensive against both
attribute- and dict-like ParsedNote inputs, unknown kwargs, and ensures
`retriever_weight` is always present as a float.

Key guarantees
--------------
- Accepts extra positional/keyword args without error (for importer compatibility).
- Tolerant of attribute vs dict access for ParsedNote.
- Always sets 'retriever_weight' in the payload (float).
- Never includes non-serializable objects (functions, PosixPath, datetime, etc.).

Public API
----------
make_note_payload(parsed_note, *args, retriever_weight=None, vault_root=None, type_defaults=None, **kwargs) -> dict
"""

from __future__ import annotations
from pathlib import Path
from typing import Any, Dict, Optional, Union, Iterable, Mapping
import datetime, math

Json = Union[None, bool, int, float, str, list, dict]

# ------------------------- helpers -------------------------

def _is_mapping(x: Any) -> bool:
    return isinstance(x, Mapping)

def _get(obj: Any, *names: str, default: Any=None) -> Any:
    """Try attribute lookup, then mapping (dict) lookup, first hit wins."""
    for n in names:
        if hasattr(obj, n):
            try:
                return getattr(obj, n)
            except Exception:
                pass
        if _is_mapping(obj) and n in obj:
            try:
                return obj[n]
            except Exception:
                pass
    return default

def _to_float(x: Any, default: float=1.0) -> float:
    if x is None:
        return float(default)
    if isinstance(x, (int, float)) and math.isfinite(x):
        return float(x)
    try:
        s = str(x).strip().replace(',', '.')
        return float(s)
    except Exception:
        return float(default)

def _ensure_list(x: Any) -> list:
    if x is None:
        return []
    if isinstance(x, list):
        return x
    if isinstance(x, (set, tuple)):
        return list(x)
    return [x]

def _sanitize(obj: Any) -> Json:
    """Recursively convert to JSON-serializable primitives; drop callables."""
    if obj is None or isinstance(obj, (bool, int, float, str)):
        return obj
    if callable(obj):
        return None
    if isinstance(obj, (list, tuple, set)):
        return [_sanitize(v) for v in obj]
    if isinstance(obj, dict):
        out = {}
        for k, v in obj.items():
            if callable(v):
                continue
            out[str(k)] = _sanitize(v)
        return out
    if isinstance(obj, Path):
        return str(obj)
    if isinstance(obj, datetime.datetime):
        return obj.isoformat()
    if hasattr(obj, "__str__"):
        try:
            return str(obj)
        except Exception:
            return None
    return None

def _compute_retriever_weight(explicit: Any, frontmatter: dict, type_defaults: Optional[dict], note_type: Optional[str]) -> float:
    if explicit is not None:
        return _to_float(explicit, 1.0)
    # common frontmatter keys
    for key in ("retriever_weight", "retriever.weight", "retrieverWeight"):
        if key in frontmatter:
            return _to_float(frontmatter.get(key), 1.0)
    # type defaults map like: {"concept": {"retriever_weight": 0.9}, ...}
    if type_defaults and note_type:
        tdef = type_defaults.get(note_type) or {}
        for key in ("retriever_weight", "retriever.weight", "retrieverWeight"):
            if key in tdef:
                return _to_float(tdef.get(key), 1.0)
    return 1.0

# ------------------------- public API -------------------------

def make_note_payload(parsed_note: Any,
                      *args,
                      retriever_weight: Optional[float]=None,
                      vault_root: Optional[str]=None,
                      type_defaults: Optional[dict]=None,
                      **kwargs) -> Dict[str, Json]:
    """
    Build a JSON-safe payload for the note.

    Parameters (tolerant; unknown args are ignored)
    ----------
    parsed_note : object or dict
        Expected fields/keys (best-effort): note_id/id, title, type, path/rel_path,
        frontmatter, tags, aliases, chunk_profile.
    retriever_weight : float|None
        Overrides frontmatter/type-defaults if provided.
    vault_root : str|None
        Optional; used to produce a normalized relative path.
    type_defaults : dict|None
        Optional map for per-type defaults.

    Returns
    -------
    dict suitable for Qdrant payload
    """
    fm = _get(parsed_note, "frontmatter", "fm", default={})
    if not isinstance(fm, dict):
        fm = {}

    note_id = _get(parsed_note, "note_id", "id", default=fm.get("id"))
    title = _get(parsed_note, "title", default=fm.get("title"))
    ntype = _get(parsed_note, "type", default=fm.get("type"))
    raw_path = _get(parsed_note, "path", "rel_path", "relpath", default=fm.get("path"))
    tags = _ensure_list(_get(parsed_note, "tags", default=fm.get("tags")))
    aliases = _ensure_list(_get(parsed_note, "aliases", default=fm.get("aliases")))
    chunk_profile = _get(parsed_note, "chunk_profile", "profile", default=fm.get("chunk_profile"))
    created = _get(parsed_note, "created", default=fm.get("created"))
    updated = _get(parsed_note, "updated", default=fm.get("updated"))

    # normalize path relative to vault root if both available
    rel_path = raw_path
    if raw_path and vault_root:
        try:
            rel_path = str(Path(raw_path)).replace(str(Path(vault_root)), "").lstrip("/\\")
        except Exception:
            rel_path = str(raw_path)

    rw = _compute_retriever_weight(retriever_weight, fm, type_defaults, ntype)

    payload = {
        "note_id": note_id,
        "title": title,
        "type": ntype,
        "path": rel_path or raw_path,
        "tags": tags,
        "aliases": aliases,
        "chunk_profile": chunk_profile,
        "created": created,
        "updated": updated,
        "retriever_weight": float(rw),
    }

    # Add selected FM fields if present (safe subset)
    for key in ("status", "priority", "owner", "source"):
        if key in fm:
            payload[key] = fm.get(key)

    return _sanitize(payload)