Prepare v.0.2.0 Release #125
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This commit updates the requirements and removes support for py27. Signed-off-by: Lester James V. Miranda <[email protected]>
In this commit, the test_singleobj was now reimplemented using the pytest module. Signed-off by: Lester James V. Miranda <[email protected]>
All tests for GlobalBestPSO, LocalBestPSO, and BinaryPSO are now updated into pytest. In addition, we also have conftest.py to provide test fixtures. Signed-off-by: Lester James V. Miranda <[email protected]>
This commit adds some descriptive information for most pytest fixtures that were defined in the docstring. Signed-off-by: Lester James V. Miranda <[email protected]>
Reference #113 Signed-off-by: ljvmiranda921 <[email protected]>
Reference: #111 This commit adds a backend module consisting of various abstractions to common PSO operations. There are two files for this: - generators: for generating swarm positions, velocities, etc. - operators: for various swarm operations (update velocity, etc.) As of now, one gripe I have with my current implementation is that it consists of methods with a lot of parameters (it doesn't look pleasing or concise). My idea is to add a Swarm class that contains the position, velocity, c1, c2, w, etc. And just put them inside the optimizer. We'll see how it goes Signed-off-by: ljvmiranda921 <[email protected]>
Using default as None is not really helpful and adds another boilerplate of checking if init_pos is None. Instead, we just multiply by 1.00, which gives us the indentity of the matrix. Signed-off-by: Lester James V. Miranda <[email protected]>
Reference: #111 This commit replaces SwarmBase with backend operations for generating the swarm.
There's really no reason to restrict v_clamp to a certain type. As long as it is an iterable, you are fine. Signed-off-by: Lester James V. Miranda <[email protected]>
This commit updates global best with backend operations from the new module Signed-off-by: Lester James V. Miranda <[email protected]>
Until they're not implemented yet, let's remove the __init__ files for these functionalities. It may confuse the user of what we can actually do. Signed-off-by: ljvmiranda921 <[email protected]>
This commit adds new methods to the update that interfaces well with the Swarm class: create_swarm() as an easy abstraction for generating new swarms and update_gbest_neighborhood() for LocaBestPSO Signed-off-by: Lester James V. Miranda <[email protected]>
This class creates a DataClass-like interface for building swarms. It holds all special attributes in swarm optimization that you can reuse in case the user wants to build their own evolutionary loop. Signed-off-by: Lester James V. Miranda <[email protected]>
In this commit, the SwarmBase, LocalBestPSO, and GlobalBestPSO are now using the Swarm class interface as its backend. The user-facing API is still the same though. If you will inspect the optimize() method and compare it with previous commits, the structure is much cleaner and concise. Signed-off-by: Lester James V. Miranda <[email protected]>
In this commit, we are now comparing the set() of the reset method rather than checking if the value is None. New tests were also added to accommodate the swarm class. Signed-off-by: ljvmiranda921 <[email protected]>
This commit adds the Topology module's __init__ file. The concept behind this module is that if we can add topologies in a modular fashion, we can easily explore different PSO behaviour. Signed-off-by: Lester James V. Miranda <[email protected]>
added 24 commits
June 7, 2018 13:29
In this commit, we revert the name behavior back into the word options. This is so a more general term and will not semantically lock us when implementing new swarm algorithms. Signed-off-by: Lester James V. Miranda <[email protected]>
In this commit, we add a set of binary generators to generate discrete/binary versions of the swarm. Signed-off-by: Lester James V. Miranda <[email protected]>
Reference: #119 This commit exposes the init_pos option for user-defined initial positions in the SwarmOptimizer base class and gbest and lbest implementations. Signed-off-by: Lester James V. Miranda <[email protected]>
This commit ports base_discrete to use the backend module. We also renamed the DiscreteSwarmBase into DiscreteSwarmOptimizer for better semantic meaning. Signed-off-by: Lester James V. Miranda <[email protected]>
This commit ports the BinaryPSO to use the backend module. Python 2.7 dependencies were also removed. Signed-off-by: Lester James V. Miranda <[email protected]>
This commit fixes the reset() test in test_binary. Now, we call on the attributes from the Swarm class, not in the DiscreteSwarmOptimizer class. Signed-off-by: Lester James V. Miranda <[email protected]>
Port discrete PSO (and other Swarm fixes)
Add pyswarms.backend module This commit adds the `backend` module for PySwarms. It enables users to define their own optimization loop using primitives provided by this module (white-box approach). In addition, this commit also updates the existing implementations of GlobalBestPSO, LocalBestPSO, and BinaryPSO to use these backend primitives. Some major updates: 1. Topologies are introduced to segregate computation of best_cost/pos and velocity or position updates. 2. A Swarms class is implemented to serve as a DataClass for data storage. 3. SwarmBase is now called SwarmOptimizer (more semantically meaningful) 4. The user can now set his/her initial positions. 5. Remove py27 compatibility Additional tasks: 1. Documentation and new notebook examples 2. Check feature selection example (#110 ) Signed-off-by: Lester James V. Miranda <[email protected]>
…o-1.7.5 Update sphinx to 1.7.5
In this commit, we'll use the Markdown file going forward. Turns out the PyPI is now more accommodating of .md files so we'll follow them from now on. In the same vein, we also updated the references found in the setup.py and MANIFEST.in of our module. In py34, there are some problems when we're not referencing properly in MANIFEST.in, and I wonder why. Signed-off-by: Lester James V. Miranda <[email protected]>
In this commit, we do a documentation cleanup for the previous documentation. Specifically, we updated the following: - Better grammar on some sentences/paragraphs. - Clearer docstring in some old methods - Add PlotEnvironment docs (wonder why it's not been added before) - Overall cleaner code and organization Signed-off-by: Lester James V. Miranda <[email protected]>
This commit adds a documentation for the backend module. I also updated some docstrings for clearer explanations Signed-off-by: Lester James V. Miranda <[email protected]>
In this commit, we add an assets directory to contain all images for our documentation. The images were generated using the TiKz package in LaTeX. Signed-off-by: Lester James V. Miranda <[email protected]>
Add some explanation about the PySwarms API and on how to use the backend module. Signed-off-by: Lester James V. Miranda <[email protected]>
This commit updates the default parameters in create_swarm (options, by default, is now an empty dict), and in the update_velocity of the ring and star topologies (clamp, by default, is now None) Signed-off-by: Lester James V. Miranda <[email protected]>
This commit should update all Jupyter notebook examples to the dev version. Hopefully everything is okay after this. This also adds a new notebook for demonstrating the custom optimization loop. Signed-off-by: Lester James V. Miranda <[email protected]>
added 3 commits
June 10, 2018 21:48
This commit adds a branch matrix in both README and index.rst. This gives a better overview of the doc and build status of both master and development branches. Signed-off-by: Lester James V. Miranda <[email protected]>
|
Congrats on this @ljvmiranda921 ! This is a great idea. I spent the past few weeks running into a similar challenge on SwarmOpt, and decided it best to have modules of core functions for replicating algorithms and experimentation, and then classes for any implementation I decide to optimize towards efficient runtime and ability to handle large spaces. I'll definitely be reaching out in the coming weeks to compare notes as I get my first overhaul in this direction completed. Thanks for continuing to provide consistent inspiration on what a thoughtful, well-documented project can be! 👏 👏 👏 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This is PySwarms development branch merging into master branch for v.0.2.0 release.
Major updates
pyswarms.backendmodule for custom swarm algorithms. Users can now use some primitives provided in this module to write their own optimzation loop, providing a more "white-box" approach in swarm intelligence. (User-provided initial positions #119) (PR Add backend module #115) (PR Add a separate Swarm class #116) (PR Add Topologies #117)The PySwarms Backend module
The new backend module exposes some swarm optimization primitives so that users can create their custom swarm implementations without relying too much from our base classes. There are two main components for backend, the
Swarmclass and theTopologybase class. Using these classes, you can construct your own optimization loop like the one below:The Swarm class
This class acts as a data class that holds all necessary attributes in a given swarm. The idea is to continually update the attributes located there. You can easily initialize this class by providing the initial position and velocity matrices.
The Topology class
The topology class abstracts away common operations in swarm optimization: (1) determining the best particle in the swarm, (2) computing the next position, and (3) computing the velocity matrix. As of now, we only have the
RingandStartopologies implemented. Hopefully we can add more in the future.