Version 0.3.0

New documentation and a lot of performance improvements.

To get the best performance, it is recommended to use geopandas>=0.8 and have pygeos installed.

API changes:

• Convexeity is now Convexity (#171)
• local_ centrality (betweenness, closeness, straightness) has been included in respective global versions (#178)

New features:

• CheckTessellationInput to check building footprint data for potential issues during Tessellation (#163)
• On demand DistanceBand spatial weights (neighbors) for larger weight which would not fit in memory (#165)

Enhancements:

• New documentation (#167)
• Support network analysis for full network (#176)
• Options for preprocess (#180)
• Expose underlying simpson and shannon functions (#183)
• MeanInterbuildingDistance performance (#187)
• StreetProfile performance refactor (#186)
• Retain attributes in network_false_nodes (#189)
• Performance improvements in elements module (#190)
• Performance refactor of SharedWallsRatio (#191)
• Performance refactor of NeighboringStreetOrientationDeviation (#192)
• Minor performance improvements (#193, #196)
• Allow specification of verbosity (#195)
• Perfomance enhancements in sw_high (#198)

Bug fixes:

• Density TypeError for islands (#164)
• Preserve CRS in network_false_nodes (#181)
• Fixed Squareness for non-Polygon geom types (#182)
• CRS lost with older geopandas (#188)

Acknowledgements

Thanks to everyone who contributed to this release either directly or indirectly!
People with a "+" by their names contributed a patch for the first time.

Alessandra Feliciotti +
Martin Fleischmann

Version 0.2.1

Bug fix release.

• fixed regression causing MeanInterbuildingDistance failure (#161)

Version 0.2.0

Version 0.2.0 with several updates, one minor API change and significant performance improvements.

API changes:

• AverageCharacter allows calculation of multiple modes (mean, median, mode) at the same time. Each can be accessed via its own attribute. Apart from mean, none is accessible using .series anymore. (#147)

Enhancements:

• Shannon index (#158)
• Simpson allows Gini-Simpson and Inverse Simpson index modes (#157)
• Diversity classes support categorical values (#159)
• SegmentsLength allows sum and mean at the same time (#146)
• AverageCharacter allows calculation of multiple modes (mean, median, mode) at the same time (#147)
• Better compatibility with OSMnx Graphs (#149)
• Orientation support LineString geometry (#156)
• AreaRatio for uneven number of features (#135)
• Performance improvements (#144, #145, #152, #155)

Bugfixes:

• float precision errors in network_false_nodes (#133)
• network_false_nodes for multiindex (#136)
• BlocksCount no neighbors error (#139, #140)
• LineString Z support in nx_to_gdf (#148)
• accidental 'rtd' print (#150)
• CentroidCorner may fail for Polygon Z (#151)

Version 0.1.1

Small bug-fix release bringing following changes:

• fix for AreaRatio resetting index of resulting Series (#127)
• fix for network_false_nodes to work with GeoSeries (#128)
• fix for incomplete spatial_weights and missing neighbors. Instead of raising KeyError
  momepy now returns np.nan for affected row. np.nan is also returned if there are no
  neighbors (instead of 0). (#131)

Version 0.1

🎉 The first official release of momepy is finally here. 🎉

After a long period of beta and RC development, I am happy to announce that since today, momepy is officially released in the first stable version 0.1.0.

During the final weeks before the release, momepy was reviewed for a publication in the Journal of Open Source Software (openjournals/joss-reviews#1807). As that is now concluded, momepy is being released together with the software paper introducing its abilities, thorough documentation (docs.momepy.org) and user guide (guide.momepy.org).

I have to thank the Axel and Margaret Ax:son Johnson Foundation for the support of the development of the package (as a part of The Urban Form Resilience Project), the University of Strathclyde, JOSS and namely @geraintpalmer and @martibosch for their thorough review of the software.

For the quick start, head to Getting started!

v0.1 release candidate 4

This should be the last release candidate ahead of the release of v0.1.

It is intended to be submitted to JOSS for a review. v0.1 should then include any changes based on the review process.

v0.1 release candidate 2

The second release candidate of the first stable release of momepy. We are heading towards a release in following weeks. Preferably v0.1 should be the one included in JOSS paper.