v2.6.5

What's changed

• Additional memory safety updates for large iteration counts
• Better handling of sampling edge cases

v2.6.4

What's changed?

• Use partitioned parquet files for improved memory scaling for large landscapes and high fire counts
• More granular progress bars for spatial summaries
• Better handling of missing vector outputs

v2.6.3

What's changed

• Bug fixes for calculating burn probability from merged scenarios
• Better support for sampling only extra ignitions
• Fix typos and bugs in input validation functions

v2.6.2

What's changed

• Better handling of missing, implied, and partial inputs
• Fix bug with export transformers
• Update default batch sizes for better memory management with Arrow

v2.6.1

What's Changed?

• Better memory scaling for large landscapes
• More detailed labels in the map UI
• Better input validation and handling of edge cases

v2.6.0

What's changed?

• Add new transformer to merge burn outputs from multiple fire growth scenarios
• Restructure all intermediate outputs around Apache Arrow to improve performance
• Refactor code to share definitions and functions across fire growth model packages and base

v2.5.3

What's changed

• Work around issue with generating FBP summaries under conda

v2.5.2

What's changed

• Minor updates for better handling of edge cases

v2.5.0

What's changed

• Use Apache Arrow to more efficiently store and summarize per-fire outputs

v2.4.1

What's Changed

• Fix memory safety when summarizing FBP outputs in conda
• Minor plotting and error checking bug fixes