Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/eric-canas/circulardict

CircularDict is a high-performance Python data structure that blends the functionality of dictionaries and circular buffers. Inheriting the usage of traditional dictionaries, it allows you to define constraints on size and memory usage. Ideal for caching large data structures while maintaining control over memory footprint.
https://github.com/eric-canas/circulardict

buffer circular-buffer circular-queue dict dictionary memory-management utility

Last synced: about 2 hours ago
JSON representation

CircularDict is a high-performance Python data structure that blends the functionality of dictionaries and circular buffers. Inheriting the usage of traditional dictionaries, it allows you to define constraints on size and memory usage. Ideal for caching large data structures while maintaining control over memory footprint.

Awesome Lists containing this project

README 

        
# CircularDict

CircularDict **CircularDict** is a Python `dict` that acts as a **Circular Buffer**. This dictionary maintains a **controlled size**, limited either by a specified **number of items** (_maxlen_) or **total size in bytes** (_maxsize_bytes_). Upon reaching the defined limit, **CircularDict** automatically removes the oldest entries, maintaining **memory usage** under control.



Built upon Python's `OrderedDict`, **CircularDict** inherits all **standard dictionary usage** and operations, augmented by the capability of **memory management**. It's particularly useful in scenarios such as **caching**, where limiting **memory consumption** is crucial. The class combines _dictionary_ and _circular-queue_ behaviors, providing an efficient, scalable and thread safe solution for various use cases.



## Installation



To install **CircularDict** simply run:



```bash

pip install circular-dict

```



## Usage



Working with **CircularDict** is as simple as using a standard Python `dict`, with additional parameters `maxlen` or `maxsize_bytes` on the initialization to control the buffer size. You can use one of them or both.

```python

from circular_dict import CircularDict

# Initialize a CircularDict with a maximum length of 3 items and a storage limit of 4Mb

my_dict = CircularDict(maxlen=3, maxsize_bytes=4*1024*1024)

```



### Example with `maxlen`

You can use `maxlen` to define the maximum amount of items that the dictionary can store. It is useful for defining fixed size buffers.



```python

from circular_dict import CircularDict



# Initialize a CircularDict with a maximum length of 3

my_buffer = CircularDict(maxlen=3)



# Fill it with 3 items

my_buffer['item1'] = 'value1'

my_buffer['item2'] = 'value2'

my_buffer['item3'] = 'value3'



print(f"When filling it: {circ_dict}")



# Add another item

my_buffer['item4'] = 'value4'



print(f"After adding an element beyond maxlen: {circ_dict}")

```



Output:

```bash

When filling it: {'item1': 'value1', 'item2': 'value2', 'item3': 'value3'}

After adding an element beyond maxlen: {'item2': 'value2', 'item3': 'value3', 'item4': 'value4'}

```



### Example with `maxsize_bytes`

You can use `maxsize_bytes` to define the maximum amount of memory that the `dict` can store. It is particularly beneficial when defining **caches**, to prevent **memory overflows**.



```python

from circular_dict import CircularDict

import numpy as np

import sys



# Initialize a CircularDict with a maximum length of 100KB

my_buffer = CircularDict(maxsize_bytes=100*1024)



# Add two arrays of ~40Kb (10*1024*4 bytes (int32) + 5 bytes (chars) + 100 bytes (numpy structure) + 50 bytes (str structure))

my_buffer['item1'] = np.zeros((10, 1024), dtype=np.int32)

my_buffer['item2'] = np.ones((10, 1024), dtype=np.int32)



print(f"{len(my_buffer)} Elements {tuple(my_buffer.keys())}. Dict size: {my_buffer.current_size/1024} Kb")



# Add a new element of ~32Kb will delete oldest elements ('item1') until fitting in the `dict`.

my_buffer['item3'] = np.ones((8, 1024), dtype=np.int32)



print(f"{len(my_buffer)} Elements {tuple(my_buffer.keys())}. Dict size: {my_buffer.current_size/1024} Kb")



# Create an element of ~160Kb (bigger than the defined maximum storage) to trigger a MemoryError

too_big_array = np.ones((40, 1024), dtype=np.int32)

try:

  # Try to add it to the dict

  my_buffer['item4'] = too_big_array

except MemoryError:

  print(f"Cannot add an element with {sys.getsizeof(too_big_array)/1024}Kb in a dict with"\

        f"maxsize_bytes of {my_buffer.maxsize_bytes/1024} Kb. Current elements {tuple(my_buffer.keys())}")

```



Output



```bash

2 Elements ('item1', 'item2'). Dict size: 80.35 Kb

2 Elements ('item2', 'item3'). Dict size: 72.35 Kb

Cannot add an element with 160.12Kb in a dict with maxsize_bytes of 100.0 Kb. Current elements ('item2', 'item3')

```



Please remember that the `maxsize_bytes` parameter considers the **total** memory footprint, including the sizes of _keys_ and _values_. If you try to add an item that exceeds the `maxsize_bytes`, a `MemoryError` will be raised. 



In environments like [PyPy](https://www.pypy.org/) where [`sys.getsizeof()` is not available](https://doc.pypy.org/en/latest/cpython_differences.html#miscellaneous) `maxsize_bytes` cannot be used and must be setted to `None`.