0.2.6

Fixed

• Function detection in elapsed_time_decorator for Google Colab environment

0.2.5

Added

• Option to pass list of hierarchy_depth values for multiple theme / type pairs
• Info about current theme / type pair to the HierarchyDepthOutOfBoundsWarning

0.2.4

Added

• Places hierarchy based on the official taxonomy #63
• Option to change minimal confidence score for places and select only primary category for the wide form transformation #63

Changed

• Added option to use any non-negative integer as a hierarchy_depth value for wide form processing #64
• Shortened hash parts for generated file names to 8 characters per part

Fixed

• Bug where a constant value has been overwritten instead of being copied before modifying

0.2.3

Fixed

• Changed wide format places definition for older release versions
• Changed get all columns function for wide format places definition
• Bug where code crashed when release index hit zero matches #11

0.2.2

Fixed

• Changed wide format definitions for different release versions

0.2.1

Added

• Wide format release index to precalculate all possible columns #43
• Flag include_all_possible_columns to keep or prune empty columns #43
• overturemaestro.advanced_functions.wide_form.get_all_possible_column_names for getting a list of all possible column names #46
• overturemaestro.cache.clear_cache function for clearing local release index cache from the API

0.2.0

Added

• Automatic total time wrapper decorator to aggregate nested function calls
• Parameter columns_to_download for selecting columns to download from the dataset #23
• Option to pass a list of pyarrow filters and columns for download for each theme type pair when downloading multiple datasets at once
• Automatic columns detection in pyarrow filters when passing columns_to_download
• New advanced_functions module with a wide format for machine learning purposes #38

Changed

• Refactored available release versions caching #24
• Removed hive partitioned parquet schema columns from GeoDataFrame loading

Deprecated

• Nested fields in PyArrow filter in CLI is now expected to be separated by a dot, not a comma #22

0.1.2

Added

• Option to pass max number of workers for downloading the data #30

0.1.1

Changed

• Modified release index consolidation script

0.1.0

Added

• CLI #3
• Option to filter data with bounding box #4
• Tests for the library #6
• Automatic newest release version loading #7
• Library docs #2
• README content
• Verbosity modes
• Total operation time
• Overloads for the functions typing
• Function for displaying all available release versions
• GitHub Action workflows for docs deployment

Changed

• Moved location of the pregenerated release indexes to the global cache #19
• Moved scikit-learn and polars to the dedicated dependency group #9
• Sped up intersection algorithm
• Reduced number of max concurrent connections for parquet files download

Fixed

• Memory leak during concurrent parquet files download
• Added automatic retry for downloads with 10 retries