Add fiboa improve command #79 #21

CHANGELOG.md

@@ -9,10 +9,29 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 
 ### Added
 
+- Command `fiboa improve` with helpers to
+  - change the CRS
+  - change the GeoParquet version and compression
+  - fill missing perimeter/area values
+  - fix invalid geometries
+  - rename columns
 - Converter for Lithuania (EuroCrops)
-- Converter for Switzerland
 - Converter for Slovenia
 - Converter for Slovakia
+- Converter for Switzerland
+- `fiboa convert`: New parameter `--original-geometries` / `-og` to keep the original geometries
+
+### Changed
+
+- `fiboa convert`:
+  - Writes custom schemas to collection metadata
+  - Geometries are made valid using GeoPanda's `make_valid` method by default
+  - MultiPolygons are converted to Polygons by default
+- `fiboa validate` uses custom schemas for validation
+- `fiboa merge` keeps custom schemas when needed
+
+### Removed
+- `fiboa convert`: Removed the explicit parameter `explode_multipolygon` from the converter
 
 ### Fixed
 

README.md

@@ -8,19 +8,20 @@ A command-line interface (CLI) for working with fiboa.
 
 ## Getting Started
 
-In order to make working with fiboa easier we have developed command-line interface (CLI) tools such as 
+In order to make working with fiboa easier we have developed command-line interface (CLI) tools such as
 inspection, validation and file format conversions.
 
 ### Installation
 
-You will need to have **Python 3.9** or any later version installed. 
+You will need to have **Python 3.9** or any later version installed.
 
 Run `pip install fiboa-cli` in the CLI to install the validator.
 
 **Optional:** To install additional dependencies for specific [converters](#converter-for-existing-datasets),
 you can for example run: `pip install fiboa-cli[xyz]` with xyz being the converter name.
 
 **Note on versions:**
+
 - fiboa CLI >= 0.3.0 works with fiboa version > 0.2.0
 - fiboa CLI < 0.3.0 works with fiboa version = 0.1.0
 
@@ -44,6 +45,7 @@ fiboa CLI supports various commands to work with the files:
     - [Merge fiboa GeoParquet files](#merge-fiboa-geoparquet-files)
     - [Create JSON Schema from fiboa Schema](#create-json-schema-from-fiboa-schema)
     - [Validate a fiboa Schema](#validate-a-fiboa-schema)
+    - [Improve a fiboa Parquet file](#improve-a-fiboa-parquet-file)
     - [Update an extension template with new names](#update-an-extension-template-with-new-names)
     - [Converter for existing datasets](#converter-for-existing-datasets)
   - [Development](#development)
@@ -121,19 +123,38 @@ To validate a fiboa Schema YAML file, you can for example run:
 
 Check `fiboa validate-schema --help` for more details.
 
+### Improve a fiboa Parquet file
+
+Various "improvements" can be applied to a fiboa GeoParquet file.
+The commands allows to
+
+- change the CRS (`--crs`)
+- change the GeoParquet version (`-gp1`) and compression (`-pc`)
+- add/fill missing perimeter/area values (`-sz`)
+- fix invalid geometries (`-g`)
+- rename columns (`-r`)
+
+Example:
+
+- `fiboa improve file.parquet -o file2.parquet -g -sz -r old=new -pc zstd`
+
+Check `fiboa improve --help` for more details.
+
 ### Update an extension template with new names
 
 Once you've created and git cloned a new extension, you can use the CLI
 to update all template placeholders with proper names.
 
 For example, if your extension is meant to have
-- the title "Timestamps Extension", 
+
+- the title "Timestamps Extension",
 - the prefix `ts` (e.g. field `ts:created` or `ts:updated`),
 - is hosted at `https://github.io/fiboa/timestamps-extension`
   (organization: `fiboa`, repository `timestamps-extension`),
 - and you run fiboa in the folder of the extension.
 
 Then the following command could be used:
+
 - `fiboa rename-extension . -t Timestamps -p ts -s timestamps-extension -o fiboa`
 
 Check `fiboa rename-extension --help` for more details.
@@ -143,13 +164,14 @@ Check `fiboa rename-extension --help` for more details.
 The CLI ships various converters for existing datasets.
 
 To get a list of available converters/datasets with title, license, etc. run:
+
 - `fiboa converters`
 
 Use any of the IDs from the list to convert an existing dataset to fiboa:
 
 - `fiboa convert de_nrw`
 
-See [Implement a converter](#implement-a-converter) for details about how to 
+See [Implement a converter](#implement-a-converter) for details about how to
 
 ## Development
 

fiboa_cli/__init__.py

@@ -6,16 +6,18 @@
 import click
 import pandas as pd
 
+from .const import COMPRESSION_METHODS, CORE_COLUMNS
 from .convert import convert as convert_
 from .convert import list_all_converter_ids, list_all_converters
 from .create_geojson import create_geojson as create_geojson_
 from .create_geoparquet import create_geoparquet as create_geoparquet_
 from .describe import describe as describe_
-from .merge import merge as merge_, DEFAULT_COLUMNS, DEFAULT_CRS
+from .improve import improve as improve_
+from .merge import merge as merge_, DEFAULT_CRS
 from .jsonschema import jsonschema as jsonschema_
 from .rename_extension import rename_extension as rename_extension_
 from .util import (check_ext_schema_for_cli, log, parse_converter_input_files,
-                   valid_file_for_cli, valid_file_for_cli_with_ext,
+                   parse_map, valid_file_for_cli, valid_file_for_cli_with_ext,
                    valid_files_folders_for_cli, valid_folder_for_cli)
 from .validate import validate as validate_
 from .validate_schema import validate_schema as validate_schema_
@@ -376,7 +378,7 @@ def jsonschema(schema, out, fiboa_version, id_):
 )
 @click.option(
     '--compression', '-pc',
-    type=click.Choice(["brotli", "gzip", "lz4", "snappy", "zstd", "none"]),
+    type=click.Choice(COMPRESSION_METHODS),
     help='Compression method for the Parquet file.',
     show_default=True,
     default="brotli"
@@ -385,7 +387,7 @@ def jsonschema(schema, out, fiboa_version, id_):
     '--geoparquet1', '-gp1',
     is_flag=True,
     type=click.BOOL,
-    help='Enforces generating a GeoParquet 1.0 file bounding box. Defaults to GeoParquet 1.1 with bounding box.',
+    help='Enforces generating a GeoParquet 1.0 file. Defaults to GeoParquet 1.1 with bounding box.',
     default=False
 )
 @click.option(
@@ -394,13 +396,20 @@ def jsonschema(schema, out, fiboa_version, id_):
     help='Url of mapping file. Some converters use additional sources with mapping data.',
     default=None
 )
-def convert(dataset, out, input, cache, source_coop, collection, compression, geoparquet1, mapping_file):
+@click.option(
+    '--original-geometries', '-og',
+    is_flag=True,
+    type=click.BOOL,
+    help='Keep the source geometries as provided, i.e. this option disables that geomtries are made valid and converted to Polygons.',
+    default=False
+)
+def convert(dataset, out, input, cache, source_coop, collection, compression, geoparquet1, mapping_file, original_geometries):
     """
     Converts existing field boundary datasets to fiboa.
     """
     log(f"fiboa CLI {__version__} - Convert '{dataset}'\n", "success")
     try:
-        convert_(dataset, out, input, cache, source_coop, collection, compression, geoparquet1, mapping_file)
+        convert_(dataset, out, input, cache, source_coop, collection, compression, geoparquet1, mapping_file, original_geometries)
     except Exception as e:
         log(e, "error")
         sys.exit(1)
@@ -518,7 +527,7 @@ def rename_extension(folder, title, slug, org = "fiboa", prefix = None):
     multiple=True,
     help='Additional column names to include.',
     show_default=True,
-    default=DEFAULT_COLUMNS,
+    default=CORE_COLUMNS,
 )
 @click.option(
     '--exclude', '-e',
@@ -536,7 +545,7 @@ def rename_extension(folder, title, slug, org = "fiboa", prefix = None):
 )
 @click.option(
     '--compression', '-pc',
-    type=click.Choice(["brotli", "gzip", "lz4", "snappy", "zstd", "none"]),
+    type=click.Choice(COMPRESSION_METHODS),
     help='Compression method for the Parquet file.',
     show_default=True,
     default="brotli"
@@ -545,7 +554,7 @@ def rename_extension(folder, title, slug, org = "fiboa", prefix = None):
     '--geoparquet1', '-gp1',
     is_flag=True,
     type=click.BOOL,
-    help='Enforces generating a GeoParquet 1.0 file bounding box. Defaults to GeoParquet 1.1 with bounding box.',
+    help='Enforces generating a GeoParquet 1.0 file. Defaults to GeoParquet 1.1 with bounding box.',
     default=False
 )
 def merge(datasets, out, crs, include, exclude, extension, compression, geoparquet1):
@@ -564,6 +573,76 @@ def merge(datasets, out, crs, include, exclude, extension, compression, geoparqu
         sys.exit(1)
 
 
+## IMPROVE (add area, perimeter, and fix geometries)
+@click.command()
+@click.argument('input', nargs=1, type=click.Path(exists=True))
+@click.option(
+    '--out', '-o',
+    type=click.Path(exists=False),
+    help='Path to write the GeoParquet file to. If not given, overwrites the input file.',
+    default=None
+)
+@click.option(
+    '--rename-column', '-r',
+    type=click.STRING,
+    callback=lambda ctx, param, value: parse_map(value),
+    multiple=True,
+    help='Renaming of columns. Provide the old name and the new name separated by an equal sign. Can be used multiple times.'
+)
+@click.option(
+    '--add-sizes', '-sz',
+    is_flag=True,
+    type=click.BOOL,
+    help='Computes missing sizes (area, perimeter)',
+    default=False
+)
+@click.option(
+    '--fix-geometries', '-g',
+    is_flag=True,
+    type=click.BOOL,
+    help='Tries to fix invalid geometries that are repored by the validator (uses GeoPanda\'s make_valid method internally)',
+    default=False
+)
+@click.option(
+    '--explode-geometries', '-e',
+    is_flag=True,
+    type=click.BOOL,
+    help='Converts MultiPolygons to Polygons',
+    default=False
+)
+@click.option(
+    '--crs',
+    type=click.STRING,
+    help='Coordinate Reference System (CRS) to use for the GeoParquet file.',
+    show_default=True,
+    default=None,
+)
+@click.option(
+    '--compression', '-pc',
+    type=click.Choice(COMPRESSION_METHODS),
+    help='Compression method for the Parquet file.',
+    show_default=True,
+    default="brotli"
+)
+@click.option(
+    '--geoparquet1', '-gp1',
+    is_flag=True,
+    type=click.BOOL,
+    help='Enforces generating a GeoParquet 1.0 file. Defaults to GeoParquet 1.1 with bounding box.',
+    default=False
+)
+def improve(input, out, rename_column, add_sizes, fix_geometries, explode_geometries, crs, compression, geoparquet1):
+    """
+    "Improves" a fiboa GeoParquet file according to the given parameters.
+    """
+    log(f"fiboa CLI {__version__} - Improve datasets\n", "success")
+    try:
+        improve_(input, out, rename_column, add_sizes, fix_geometries, explode_geometries, crs, compression, geoparquet1)
+    except Exception as e:
+        log(e, "error")
+        sys.exit(1)
+
+
 cli.add_command(describe)
 cli.add_command(validate)
 cli.add_command(validate_schema)
@@ -574,6 +653,7 @@ def merge(datasets, out, crs, include, exclude, extension, compression, geoparqu
 cli.add_command(converters)
 cli.add_command(rename_extension)
 cli.add_command(merge)
+cli.add_command(improve)
 
 if __name__ == '__main__':
     cli()

fiboa_cli/const.py

@@ -10,3 +10,14 @@
 STAC_COLLECTION_SCHEMA = "http://schemas.stacspec.org/v{version}/collection-spec/json-schema/collection.json"
 GEOPARQUET_SCHEMA = "https://geoparquet.org/releases/v{version}/schema.json"
 STAC_TABLE_EXTENSION = "https://stac-extensions.github.io/table/v1.2.0/schema.json"
+
+COMPRESSION_METHODS = ["brotli", "gzip", "lz4", "snappy", "zstd", "none"]
+
+CORE_COLUMNS = [
+    "id",
+    "geometry",
+    "area",
+    "perimeter",
+    "determination_datetime",
+    "determination_method",
+]

fiboa_cli/convert.py

@@ -13,7 +13,8 @@ def convert(
         collection = False,
         compression = None,
         geoparquet1 = False,
-        mapping_file=None,
+        mapping_file = None,
+        original_geometries = False,
     ):
     if dataset in IGNORED_DATASET_FILES:
         raise Exception(f"'{dataset}' is not a converter")
@@ -37,6 +38,7 @@ def convert(
         compression = compression,
         geoparquet1 = geoparquet1,
         mapping_file = mapping_file,
+        original_geometries = original_geometries,
     )
 
 def list_all_converter_ids():

fiboa_cli/convert_utils.py

@@ -40,7 +40,7 @@ def convert(
         license = None,
         compression = None,
         geoparquet1 = False,
-        explode_multipolygon = False,
+        original_geometries = False,
         index_as_id = False,
         **kwargs):
     """
@@ -160,11 +160,12 @@ def convert(
             else:
                 log(f"Column '{key}' not found in dataset, skipping migration", "warning")
 
-    # 4b. For geometry column, convert multipolygon type to polygon
-    if explode_multipolygon:
+    # 4b. For geometry column, fix geometries
+    if not original_geometries:
+        gdf.geometry = gdf.geometry.make_valid()
         gdf = gdf.explode()
 
-    if has_migration or has_col_migrations or has_col_filters or has_col_additions or explode_multipolygon:
+    if has_migration or has_col_migrations or has_col_filters or has_col_additions:
         log("GeoDataFrame after migrations and filters:")
         print(gdf.head())
 

fiboa_cli/datasets/be_wa.py

@@ -91,7 +91,6 @@ def file_migration(data, path, uri, layer):
         license=LICENSE,
         layer_filter=lambda layer, uri: layer == LAYER,
         file_migration=file_migration,
-        explode_multipolygon=True,
         index_as_id=True,
         **kwargs
     )

fiboa_cli/datasets/ch.py

@@ -69,7 +69,6 @@ def convert(output_file, cache = None, **kwargs):
         column_migrations=COLUMN_MIGRATIONS,
         column_filters=COLUMN_FILTERS,
         providers=PROVIDERS,
-        explode_multipolygon=True,
         index_as_id=True,
         fid_as_index=True,
         **kwargs

fiboa_cli/datasets/ec_fr.py

@@ -25,6 +25,5 @@ def convert(output_file, cache = None, **kwargs):
         column_filters=base.COLUMN_FILTERS,
         attribution=base.ATTRIBUTION,
         license=LICENSE,
-        explode_multipolygon=True,
         **kwargs
     )

fiboa_cli/datasets/es_cat.py

@@ -35,7 +35,6 @@
 
 COLUMN_MIGRATIONS = {
     "campanya": lambda col: pd.to_datetime(col, format='%Y'),
-    "geometry": lambda col: col.make_valid(),
 }
 
 MISSING_SCHEMAS = {
@@ -62,6 +61,5 @@ def convert(output_file, cache = None, **kwargs):
         license=LICENSE,
         layer="CULTIUS_DUN2023",
         index_as_id=True,
-        explode_multipolygon=True,
         **kwargs
     )

fiboa_cli/datasets/fi.py

@@ -33,8 +33,6 @@
 COLUMN_MIGRATIONS = {
     # Make year (1st january) from column "VUOSI"
     "VUOSI": lambda col: pd.to_datetime(col, format='%Y'),
-    # Todo: generate a generic solution for making geometries valid
-    "geometry": lambda col: col.make_valid()
 }
 
 def migrate(gdf):