hashdict

Dict that only stores hashes of the keys, and not the key value. Good for large keys.

Install

This isn't on PyPI because

requests.exceptions.HTTPError: 400 Client Error: The name 'hashdict' is too similar to an existing project. See https://pypi.org/help/#project-name for more information. for url: https://upload.pypi.org/legacy/

So you can install it from GitHub which has namespaces:

pip install https://github.com/cjrh/hashdict/releases/download/0.1.0/hashdict-0.1.0-py3-none-any.whl 

Overview

HashDict is a Python package that provides a dictionary-like class designed to save memory when keys are large. Instead of storing the full text, it only stores a hash of the text, reducing memory usage significantly.

The main drawback is that the original key values are not recoverable from the HashDict instance, but this is also by design.

Example Usage

Here's a quick example to get you started with HashDict:

from hashdict import HashDict

# Create a new HashDict instance
hash_dict = HashDict()

# Add a large text key
s = "some large text " * 1000
hash_dict[s] = 42

# Retrieve the value using the same large text key
print(hash_dict[s])  # Output: 42

# Save the HashDict to a file
hash_dict.save('hash_dict.pkl')

# Load the HashDict from a file
loaded_hash_dict = HashDict.load('hash_dict.pkl')
print(loaded_hash_dict[s])  # Output: 42

Extra features

HashDict also supports the following features:

Persistence

You can save a HashDict instance to a file and load it back later:

from hashdict import HashDict

# Create a new HashDict instance
hash_dict = HashDict()

# Add entries
hash_dict["example text"] = 100
hash_dict["another example"] = 200

# Save to file
hash_dict.save('my_hash_dict.pkl')

# Load from file
loaded_hash_dict = HashDict.load('my_hash_dict.pkl')
print(loaded_hash_dict["example text"])  # Output: 100

Internally, the save method writes to a temporary file and then renames it to the final file name, to ensure that the file at the original path is not corrupted if the process is interrupted.