pydantic-cacheable-model 1.0.0

Small Pydantic helpers for cacheable models

pydantic-cacheable-model

Small helpers to make your Pydantic v2 models trivially cacheable to disk.

---

Installation

pip install pydantic-cacheable-model

Supports Python 3.11+

---

Usage

from __future__ import annotations
from datetime import datetime
from pydantic_cacheable_model import CacheableModel, CacheId

class UserModel(CacheableModel):
    email: CacheId[str]  # <- used as the cache identifier
    joined: datetime

# Create and cache to disk
u = UserModel(email="[e-mail]", joined=datetime.now())
u.cache()

# Load a single item by identifier
u2 = UserModel.load(cache_id="[e-mail]")
assert u2 == u

# Optionally handle missing caches or validation mismatches
maybe_user = UserModel.load(cache_id="[e-mail]", not_found_ok=True)

# List everything cached
all_users = UserModel.load_all_cached()

---

Configuration

Below are the configuration knobs you can use. This single block shows all options; pick the ones you need.

from __future__ import annotations
from datetime import datetime
from pydantic_cacheable_model import CacheableModel, CacheId

# 1) Customize cache root and directory name
class UserModel(CacheableModel):
    # Change root directory (default: ".cache")
    CACHE_ROOT = ".data-cache"

    # Force a specific subdirectory name (otherwise derived from class name)
    CACHE_DIRNAME = "users"

    email: CacheId[str]
    joined: datetime


# 2) Customize filename mapping (avoid if ids contain unsafe characters)
class FriendlyFilenames(CacheableModel):
    slug: CacheId[str]

    @classmethod
    def cache_id_to_filename(cls, *, cache_id: str) -> str:
        # Default uses a SHA-256 hash; this stores plain ids instead
        return f"{cache_id}.json"


# 3) Custom identifier logic without using CacheId
class Document(CacheableModel):
    kind: str
    slug: str

    @property
    def cache_id(self) -> str:
        return f"{self.kind}/{self.slug}"

---

How It Works

• Cache location: ./.cache/<model-dir>/ by default. <model-dir> is derived from the class name, removing a trailing Model and converting CamelCase to kebab-case (e.g., UserModel → user, LongNameModel → long-name).
• File naming: sha256(cache_id).json for safe, stable filenames.
• Data format: pretty-printed JSON. Enum values and datetime objects serialize automatically.
• Choosing the identifier: mark one field with CacheId[T]. If you need custom logic, override the cache_id property instead.

---

API

• CACHE_ROOT: str — class var; default ".cache".
• CACHE_DIRNAME: str | None — class var; override to force subdirectory name. If None, derived from the class name.
• cache_dir_path() -> str — directory path where this model caches files.
• cache_id_to_filename(*, cache_id: str) -> str — classmethod; maps an id to a filename (default: SHA-256 hex + .json).
• get_cache_path(*, cache_id: str) -> str — full path to the cache file for a given id.
• load(*, cache_id: str, not_found_ok: bool = False, warn_mismatch: bool = True) -> Self | None — load and validate. If not_found_ok=True, returns None on missing or, when validation fails, returns None and optionally warns.
• load_all_cached() -> list[Self] — load all cached instances for the model.
• cache() -> None — write the instance to disk.
• cache_path: str — the path where this instance is cached.

---

License

MIT — free to use and modify.