numpy_ringbuffer

numpy_ringbuffer copied to clipboard

Maintain copy of unwrapped array

1

Array unwrap is computationally expensive for large arrays. This PR makes sure the underlying array is not unnecessarily unwrapped if it has not changed since it was last unwrapped. All...

nup002

Fixes issues of unsafe memory access

1

This adds the tests `test_raises_error_indexing_empty_buffer` and `test_negative_indices_give_recent_data_with_unfull_buffer` to ensure the class doesn't access memory unsafely. Closes #13 Closes #14

fpdotmonkey

Add handling of negative indices to getitem

1

This will handle cases of `ringbuf[-]`, which were failing when `self._arr` is not full. This may address the needs of your Issue #5, but I suspect it will not handle...

pourhouse

In a partly-full buffer, if you take a negative index, it will refer that index to the end of the full length of the buffer as apposed to the end...

fpdotmonkey

If you index a `RingBuffer` before having appended any values to it, you will unsafely get whatever was in memory at the buffer's location ```python from numpy_ringbuffer import RingBuffer r...

fpdotmonkey

The Readme.md says that collections.deque operations are available, but RingBuffer inherits from collections ABC, Sequence? At any rate, I can't access the deque.clear() method. Am I missing something?

paulobear

`__getitem__` does not obey `_left_index` and `_right_index`

7

Hi @eric-wieser, simple indexing does not work as expected in combination with `pop()`. E.g., ```python r = RingBuffer(capacity=4, dtype=np.int32) r.extend((1,2,3,4)) r.pop() print(r[-1]) >>> 4 ``` One would expect to see...

b4be1

Maybe you should answer to this SO question http://stackoverflow.com/questions/4151320/efficient-circular-buffer

s-celles

Add wheel to workflow dependencies

1

Should resolve the workflow error: ``` DEPRECATION: numpy-ringbuffer is being installed using the legacy 'setup.py install' method, because it does not have a 'pyproject.toml' and the 'wheel' package is not...

charlesbmi

Fix extend-with-overflow

1

Previous behavior: ```python >>> import numpy as np >>> from numpy_ringbuffer import RingBuffer >>> r = RingBuffer(5) >>> r.extend(np.arange(8)) >>> r ``` Expected/fixed behavior: ```python >>> import numpy as np...

charlesbmi