v2.12

[v2.12] - 2025-2-3

Changed

• Enhanced Line Visualization and Efficiency in model_evaluation.py : Improved plotting by extending the 1:1 line, OLS line, and Theil-Sen line to x = xmax. Previously, these lines stopped at x = max(clipped_gdf['ActualDef']), limiting visualization. (#42)

• Updated Documentations and Label Wording in mct_pre_screen.ui:

  • Revised label wording from "Adjusted Predicted Density in the CHF" to "Adjusted Predicted Density in the CNF". (#43)
  • Added the Theil-Sen regression line explanation to the Model Fit Assessment documentation (#44)
  • Updated README.md to include a link to VT0007 Unplanned Deforestation Allocation Report and a figure illustrating Fitting and Prediction Phases and Chronology of the Testing and Application Stages (#50).

• Enabled Manual Input for Variables and Folder Directory in UDef-ARP.py:

  • Users can now manually enter partial or full paths for variables. (#47)
  • After selecting variables and running the initial results, users can update input values by typing new names into the line edits and re-running the processing. (#45)
  • Added functionality to manually input a folder directory.(#53)

• Improved Error Debugging with Detailed Stack Traces in UDef-ARP.py:

  • Enhanced error messages with detailed stack traces for easier debugging.
  • A custom error dialog featuring:
    • A Show Details button to expand and display the full error message
    • A Hide Details button to close the detailed view
    • A Copy Error Message button to quickly copy the error text to the clipboard for easy sharing.(#52)

• Preserve Original CSV Name and Rename Adjusted CSV File for Prediction in allocation_tool.py: In cases where bins are missing during the prediction phase, the original CSV file name remains unchanged. The adjusted file is now renamed as csv_adjusted_for_prediction to clearly differentiate it from the original file. (#54)

Fixed

• Resolved Spatial Resolution Issue in vulnerability_map.py : Removed int() from line 67 in the nrt_calculation function to prevent issues when the spatial resolution is smaller than one.(#46)

v2.11

[v2.11] - 2024-10-30

Added

• Added Warning Message for Exceeding Maximum Iterations in UDef-ARP.py and allocation_tool.py: Revised GUI functionality to display a warning message when the iteration count exceeds the user's maximum iteration setting in Allocated Risk Mapping during CNF and VP period. The warning prompts the user to increase the Maximum Iterations and rerun the tool. (#35)

• Added the Folder Checker in UDef-ARP.py: Added folder checker and QMessageBox reminder for missing working directory selection. (#39)

• Added Mask Input to geometric_classification Function in UDef-ARP.py and allocation_tool.py: Addressed an user error where Distance from Forest Edge images are mistakenly created without applying the mask, leading to incorrect vulnerability mapping results. Adding the mask as an input ensures accurate processing by requiring masked distance images for classification. (#41)

Changed

• Optimized the Reclassification Process in vulnerability_map.py : Optimized the reclassification process by implementing a for loop and improve readability. in geometric_classification function(#32) and geometric_classification_alternative function. (#37)

• Modified the replace_ref_system function in vulnerability_map.py, allocation_tool.py, and model_evaluation.py: To handle scenarios where the output data format is rst and the input format is tif. The revised version copies the projection from the tif WKT and corrects the projection system name in the rdc file. (#38)

• Relocated the Input File Checker in Udef-ARP.py: Relocated input file checker and QMessageBox to the start of the process function to ensure error dialogs prompting the user to select all input files displayed before encountering downstream errors. (#39)

• Revised GUI Parameter Terminology: Revised GUI parameter labels to align with TerrSet UDef-A tool terminology, enhancing user experience by improving consistency. (#41)

• Updated HELP Documentations: Revised HELP documentation to match latest content updates. (#41)

Fixed

• Fix precision issues in vulnerability_map.py: Explicitly setting llmin and llmax in the in geometric_classification function (#31, #32) and geometric_classification_alternative function (#37) to prevent precision issues.

• Fix AR iteration in prediction density array generation in allocation_tool.py: Fixed a bug where the AR iteration count was 2 or more, causing an issue when creating the prediction density array. This update automatically uses the prediction density array from the previous round instead of the original density map. (#34)

• Refine alternative geometric classification algorithm to correct zone width assignments in vulnerability_map.py: The original algorithm assigned narrow zones in areas of lowest vulnerability and wider zones in areas of highest vulnerability. This update reverses that logic, ensuring narrow zones are applied to areas of highest vulnerability, with progressively wider zones for lower-vulnerability areas. The fixed formula LLc = LLmin + (LLmax - (LLmin * r^c)). (#37)

v2.10

[v2.10] - 2024-08-15

Added

• Theil-Sen Regression Line in model_evaluation.py: Add Theil-Sen regression line using SciPy, with a default setting of a 95% confidence level and the ‘separate’ method for intercept calculation (np.median(y) - slope * np.median(x)). Updated the terminology from "Best Fit Line" to "OLS Line" in the plot. (#30)

Changed

• Optimized Alternative Geometric Classification in vulnerability_map.py : Enhanced the processing of raster images by conditionally calculating max_value only for images lacking metadata. For images with metadata, the entire array is skipped, relying on the GetMaximum() method. (#22, #24)

• UI Wording Update in Model Evaluation UI: Revised the wording from "Mask of the Non-Excluded Jurisdiction" to "Mask of the Jurisdiction". (#25)

• Enhanced Performance of Thiessen Polygon Creation in model_evaluation.py: Refined the mask polygon function to boost performance by removing the maximum polygon selection step and utilizing GeoPandas’ MultiPolygon for intersection. (#26)

• Plotting Style Adjustments in model_evaluation.py: Modified the plotting style to include an alpha effect, improving point density visibility. (#30)

Fixed

• Bug Fix in UDef-ARP.py: Corrected the get_image_resolution function by moving it from the MCT_FIT_CAL_Screen class to the MapChecker class. (#18)

v2.09

[v2.09] - 2024-05-1

Added

• Binary Map Checker Functionality in UDef-ARP.py : Introducing a binary map checker to automatically verify the binary input data. If the data is not a binary map, the function will display an error message. (#16)

• UDef-ARP Plugin for QGIS : The UDef-ARP QGIS Plugin zipped folder is available in the Assets for download. For installation instructions and source code, please visit UDef-ARP-Plugin Repository

Changed

• Enhanced Thiessen Polygon Method with 99.9% of Maximum Size Threshold in Model Evaluation : Improving the Thiessen Polygon Method by ensuring polygons maintain up to 99.9% of their maximum size. This modification facilitates the inclusion of more samples while accepting minor irregularities in some edge polygon shapes. (#14)

• Implemented New Plotting Style and Added MedAE Percentage Calculation in Model Evaluation: Introduces a new plotting style that incorporates a legend and a white boundary around each dot to improve visibility of overlaps. MedAE percentage is calculated as (MedAE/Grid_Size) * 100, where the Grid_Size represents the area of the assessment grids in hectares (ha), with a default value of 100,000 ha. The resulting MedAE% is expected to range between 0.1% and 0.9%. (#15)

• Reorganized Script Structure in UDef-ARP.py : The reorganization improves code maintainability and reusability. (#17)

v2.08

[v2.08] - 2024-03-18

Added

• Enhanced Plot Customization in Model Evaluation: Users now have the flexibility to customize the maximum values for both the x-axis and y-axis, enhancing their ability to visually compare different plots.

• Enhanced Allocated Risk Mapping: Implementing a weighted average of the relative frequency calculation for these missing bins in the prediction phase utilizing jurisdiction-wide data.