iHashmap

Wrapper for key-value based storage systems. Provides convenient way to organize data for quick searching.

Installation

1. Using pip: pip install ihashmap

2. Building from source: make install

How to use

Firstly you need to register methods:

from ihashmap.cache import Cache

Cache.register_get_method(YOUR_GET_METHOD)
Cache.register_set_method(YOUR_SET_METHOD)
Cache.register_update_method(YOUR_UPDATE_METHOD)
Cache.register_delete_method(YOUR_DELETE_METHOD)

NOTE: Methods signature MUST match their placeholders signature

GET_METHOD = lambda cache, name, key, default=None: None  # noqa: E731
SET_METHOD = lambda cache, name, key, value: None  # noqa: E731
UPDATE_METHOD = lambda cache, name, key, value: None  # noqa: E731
DELETE_METHOD = lambda cache, name, key: None  # noqa: E731
"""METHODS placeholders. You should register yours."""

Now you are all set up to use Cache.search

How it works

In default setup Cache creates and maintains indexes based on Cache.primary_key.

So every object save in cache MUST have such key. (By default its _id)

On every called action for example Cache.update Cache looks in pipeline Cache.PIPELINE.update for middlewares to run before and after main function execution. For example in current situation after .update function execution indexing middleware will check if documents fields matching its keys were changed. If so it will get index data, look for old values in value.__shadow_copy__ remove such index data and create new record with updated values.

Adding middlewares

Adding new action is easy:

from ihashmap.cache import Cache, PipelineContext

@Cache.PIPELINE.set.before()
def add_my_field(ctx: PipelineContext):

    key, value = ctx.args
    value["my_field"] = 1

Now every cache value saved with Cache.set will be added 'my_field' before main function execution.

Custom Indexes

To create custom index you need to simply create new subclass of Index.

from ihashmap.index import Index

class IndexByModel(Index):
    keys = ["_id", "model"]

NOTE: After that all values MUST have fields _id AND model

NOTE: Primary key MUST ALWAYS be in keys

Searching

After all required indexes created - searching will be as quick as possible.

from ihashmap.cache import Cache
from ihashmap.index import Index

class IndexByModel(Index):
    keys = ["_id", "model"]

cache = Cache()
cache.search("my_cache", {"model": "1.0"})

When .search is called it will firstly check for indexes containing search fields. After finding best index, it will get index data and find matching primary keys. Now searching is as easy as getting values by their key.