Persistent append-only data structures.
Project description
appendonly README
This package provides a set of data structures for use in ZODB applications where standard BTrees are poor fits for an application’s requirements.
In particular, these data structures are designed to minimize conflict errors when doing frequent “append” operations to queues and stacks.
appendonly.AppendStack
This class provides a LIFO stack of separately-persisted objects:
The stack manages a set of “layer” objects, with a configurable limit on the number of layers. Each layer has a configurable maximum length, and a sequential generation number.
The stack appends items to most recent layer until the layer is filled; it then adds a new layer.
If the number of layers then exceeds the configured maximum, the stack pruneds the oldest layer(s) to conform to that limit.
When pruning, the stack calls an application-supplied callback for archiving / cleanup of the pruned layer.
Iteration over the stack yields (generation, index, object) tuples. in reverse order to that in which the objects were appended.
The stack is implemented as a single persistent record, with custom ZODB conflict resolution code.
appendonly.Archive
This class provides a linked list of separately-persisted copies of layer data pruned from an AppendStack. The intended use would be something like:
from appendonly import AppendStack from appendonly import Archive class RecentItems(object): def __init__(self): self._recent = AppendStack() self._archive = Archive() def pushItem(object): self._stack.push(object, self._archive.addLayer) def __iter__(self): for generation, index, item in self._stack: yield item for generation, index, item in self._archive: yield items
appendonly.Accumulator
This class provides a list-like data structure, where the only mutations allowed are to append to or clear the list. Intended uses are for a set of pending operations / changes / notifications, which get processed as a unit (at which point the accumulator is cleared).
appendonly Changelog
1.2 (2014-12-28)
Add support for Python 3.4.
Make ZODB dependency unconditional, now that it fully supports Py3k.
Add support for testing on Travis.
1.1 (2013-11-28)
Add Accumulator.extend method to enhance list-ness.
Add support for PyPy.
Add support for Python 3.2 / 3.3.
1.0.1 (2013-02-25)
Fix brown-bag in 1.0 release (Accumulator.append changes were not persisted).
1.0 (2013-02-25)
Add a conflict-free ‘Accumulator’ class: manages a queue of items which can be iterated, appended to, or conumed entirely (no partial / pop).
Automate tests for supported Python versions via ‘tox’.
Drop support for Python 2.4 / 2.5.
0.10 (2012-02-21)
Add an ‘Archive’ class, intended to support long-term storage of the layer data pruned from an ‘AppendStack’.
0.9 (2010-08-09)
Initial public release.
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
File details
Details for the file appendonly-1.2.tar.gz
.
File metadata
- Download URL: appendonly-1.2.tar.gz
- Upload date:
- Size: 11.8 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 30e057e2ec8917491112fd32146695f82597de2f171ff4c1977d5ca604196a60 |
|
MD5 | 55ccc4b4eb0387688eafde15de7b450a |
|
BLAKE2b-256 | 3a43542e9ef723a09c1f11fe4f18bd4057bbdc2b117579125b936ebe271d23ad |