Changelog

This page documents the release history of py3plex. For the most up-to-date information, see GitHub Releases.

Release History

Version 1.0.0 (Current)

Release Date: 2024

Major Features:

  • DSL Query Language: SQL-like domain-specific language for querying multilayer networks

    • String syntax (execute_query) for simple queries

    • Builder API (Q.nodes(), Q.edges()) for complex, chainable queries

    • Layer algebra (L[layer], L.intersection(), L.union())

    • Comprehensive filter, compute, and aggregation operations

  • First-Class Uncertainty Quantification:

    • StatSeries and StatMatrix types for statistics with uncertainty

    • Resampling methods (SEED, PERTURBATION) for error propagation

    • Backward-compatible with existing code expecting scalars

    • Integrated across centrality, community detection, and dynamics modules

  • Temporal Network Support:

    • Time-stamped edges and temporal queries

    • Snapshot-based analysis

    • Temporal dynamics simulation

  • Enhanced Dynamics Framework:

    • Refactored dynamics API with consistent interface

    • SIR, SIS, and custom dynamics models

    • Layer-specific and time-varying parameters

    • Comprehensive result objects with measure extraction

  • Community Detection Improvements:

    • Multilayer Louvain with inter-layer coupling

    • Infomap integration (with external binary)

    • Layer-specific community detection workflows

    • Community comparison and stability analysis

  • Web GUI:

    • Interactive network exploration (FastAPI + SvelteKit)

    • Real-time visualization and query interface

    • REST API for programmatic access

  • CLI Tools:

    • py3plex command-line interface

    • Community detection, statistics, DSL linting

    • Network conversion and export utilities

  • Documentation Overhaul:

    • Comprehensive how-to guides

    • Conceptual documentation

    • API reference with examples

    • Tutorial notebooks

  • Testing & Quality:

    • 500+ unit and integration tests

    • Continuous integration (CI/CD)

    • Type hints throughout codebase

    • Code coverage tracking

API Changes:

  • Standardized function signatures across modules

  • Consistent parameter naming (gamma for resolution, omega for coupling)

  • Deprecated legacy APIs (warnings emitted, will be removed in 2.0)

Performance Improvements:

  • Optimized tensor operations

  • Sparse matrix support for large networks

  • Caching for expensive computations

Bug Fixes:

  • Numerous edge case fixes in community detection

  • Corrected centrality calculations for directed multilayer networks

  • Fixed memory leaks in dynamics simulations

  • Resolved NetworkX compatibility issues

Previous Versions (0.x Series)

Version 0.30 - 0.35 (2020-2023)

  • Incremental feature additions

  • Experimental DSL prototype

  • Initial dynamics framework

  • Community detection algorithms

Version 0.20 - 0.29 (2018-2020)

  • Multilayer centrality measures

  • Visualization improvements

  • NetworkX integration

Version 0.10 - 0.19 (2016-2018)

  • Initial public release

  • Core multilayer network data structures

  • Basic algorithms (degree, clustering)

  • Example notebooks

Version 0.1 - 0.9 (2014-2016)

  • Research prototype

  • Proof-of-concept implementations

  • Academic collaborations

How to Check Your Version

import py3plex
print(py3plex.__version__)

Expected output:

1.0.0

Upgrade Instructions

Upgrade to latest stable release:

pip install --upgrade py3plex

Install from specific version:

pip install py3plex==1.0.0

Install development version:

pip install git+https://github.com/SkBlaz/py3plex.git@main

Migration Guides

Migrating from 0.x to 1.0

Breaking Changes:

  1. Network construction API:

    • Old: network.add_nodes(node_list)

    • New: network.add_nodes([('node', 'layer'), ...]) (explicit tuples)

  2. Community detection returns:

    • Old: louvain_communities() returned dict with node IDs as keys

    • New: Returns dict with (node, layer) tuples as keys

  3. Centrality functions:

    • Old: Scalar returns

    • New: Can return StatSeries when uncertainty=True

    • Backward compatible: Use result.mean or np.array(result)

Deprecated APIs (still work but emit warnings):

  • network.get_nodes_by_layer(layer) → Use DSL: Q.nodes().from_layers(L[layer])

  • network.get_layer_edges(layer) → Use DSL: Q.edges().from_layers(L[layer])

  • compute_centrality(network, measure) → Use DSL: Q.nodes().compute(measure)

New Recommended Patterns:

# Old way (still works)
nodes = network.get_nodes_by_layer('layer1')
for node in nodes:
    degree = network.core_network.degree(node)

# New DSL way (recommended)
from py3plex.dsl import Q, L
result = (
    Q.nodes()
     .from_layers(L['layer1'])
     .compute("degree")
     .execute(network)
)

Migration Steps:

  1. Update py3plex: pip install --upgrade py3plex

  2. Run your existing code - warnings will indicate deprecated usage

  3. Gradually migrate to DSL for queries

  4. Update community detection code to expect tuple keys

  5. Handle uncertainty in centrality results if using uncertainty=True

API Changes and Deprecations

Removed in 1.0:

  • old_louvain_function() - replaced by louvain_communities()

  • legacy_centrality_wrapper() - use centrality functions directly

Deprecated (warnings only, will be removed in 2.0):

  • Direct core_network manipulation - use DSL or network methods

  • Legacy node/edge addition formats - use dict or tuple formats

Renamed:

  • multinet.multi_layer_network() - unchanged (primary API)

  • execute_dsl_query() → execute_query() (alias still works)

Contributing to Changelog

When submitting PRs, please update the “Unreleased” section in CHANGELOG.md in the repository root with:

  • Brief description of changes

  • Link to issue (if applicable)

  • Breaking changes clearly marked

  • Migration notes for API changes

See Contributing to py3plex for details.

Staying Updated

Next Steps