reprospect.tools.cacher module

class reprospect.tools.cacher.Cacher(directory: str | Path)View on GitHub

Bases: ABC

Cache results.

The cacher stores the files as follows:

~/.<cache_dir>/
├── <hash>/
│   ├── file_A
│   ├── file_B
│   └── ...
└── cache.db

Available hashes are stored in a SQLITE database (cache.db) that also stores the timestamp at which the entry was populated.

Inspired by ccache, see also https://ccache.dev/manual/4.12.1.html#_how_ccache_works.

class Entry(cached: 'bool', digest: 'str', timestamp: 'datetime.datetime', directory: 'pathlib.Path')View on GitHub

Bases: object

__init__(cached: bool, digest: str, timestamp: datetime, directory: Path) → None

cached: bool

Set to True if the result was served from cache.

digest: str

Digest of the entry.

directory: Path

timestamp: datetime

Timestamp at which the entry was populated.

TABLE: ClassVar[str]

Name of the table.

__enter__() → SelfView on GitHub

Connect to the database.

__exit__(*args, **kwargs) → NoneView on GitHub

__init__(directory: str | Path) → NoneView on GitHub

Parameters:

directory – Where the cacher stores all files.

create_db(file: Path) → ConnectionView on GitHub

Create the SQLITE database.

get(**kwargs) → EntryView on GitHub

Serve cached entry on cache hit, populate on cache miss.

hash(**kwargs) → blake3View on GitHub

This method is intended to be overridden in a child class with the specifics of your case.

abstractmethod populate(directory: Path, **kwargs) → NoneView on GitHub