Updating write_tags file and adding basic CLI

.gitignore

@@ -46,6 +46,7 @@ pip-delete-this-directory.txt
 # Temporary files
 *scratch*.py
 scripts/test*.py
+scripts/out*.geojson
 .pypirc
 .coverage
 .vscode

requirements.txt

@@ -1,2 +1,2 @@
-pydantic>=2.8.2
+pydantic>=2.8.0
 pytest>=8.3.2

scripts/write_tags.py

@@ -1,17 +1,19 @@
 import json
+import re
 
 with open("scripts/tags.json", "r+", encoding="utf-8") as f:
     tags = json.load(f)
 
 tags_filled = {k: v for k, v in tags.items() if v}
 
-with open("src/overturetoosm/resources.py", "w+", encoding="utf-8") as f:
-    f.write(
-        f'''"""A mapping of Overture tags to OSM tags."""
+new_dict_str = json.dumps(tags_filled, indent=4).replace(":", ": ")
+pattern = r"places_tags: Dict\[str, Dict\[str, str\]\]\s*=\s*\{.*\}"
+replacement = f"places_tags: Dict[str, Dict[str, str]] = {new_dict_str}"
+
+with open("src/overturetoosm/resources.py", "r", encoding="utf-8") as f:
+    contents = f.read()
 
-tags: dict[str, dict[str, str]] = {json.dumps(tags_filled, indent=4)}
-"""dict[str, dict[str, str]]: A mapping of Overture to OSM tags, 
-excluding blank values. This is downstream from the `scripts/tag.json` 
-file."""
-'''
-    )
+replace = re.sub(pattern, replacement, contents, flags=re.DOTALL)
+
+with open("src/overturetoosm/resources.py", "w+", encoding="utf-8") as f:
+    f.write(replace)

src/overturetoosm/__main__.py

@@ -0,0 +1,6 @@
+"""Entrypoint for the package."""
+
+from .cli import main
+
+if __name__ == "__main__":
+    main()

src/overturetoosm/cli.py

@@ -0,0 +1,70 @@
+"""Command line interface for the overturetoosm package."""
+
+import argparse
+import json
+from typing import Dict
+
+from . import process_address, process_building, process_geojson, process_place
+
+
+def parse_kwargs(pairs) -> Dict[str, str]:
+    """Parse key-value pairs into a dictionary."""
+    kwargs = {}
+    for pair in pairs:
+        key, value = pair.split("=")
+        kwargs[key] = value
+    return kwargs
+
+
+def main():
+    """Configure the argument parser for the CLI."""
+    parser = argparse.ArgumentParser(
+        description="Convert Overture data to the OSM schema in the GeoJSON format."
+    )
+
+    parser.add_argument(
+        "-i", "--input", required=True, help="Path to the input GeoJSON file"
+    )
+    out = parser.add_argument_group("Output options")
+    output_group = out.add_mutually_exclusive_group(required=True)
+    output_group.add_argument("-o", "--output", help="Path to the output GeoJSON file")
+    output_group.add_argument(
+        "--in-place",
+        action="store_true",
+        help="Convert the input file in place (overwrites the original file)",
+    )
+
+    parser.add_argument(
+        "-t",
+        "--type",
+        choices=["place", "building", "address"],
+        type=str,
+        required=True,
+        help="Type of feature to convert (place, building, or address)",
+    )
+    parser.add_argument("kwargs", nargs="*", help="Additional arguments")
+
+    args = parser.parse_args()
+
+    kwargs = parse_kwargs(args.kwargs)
+
+    fx_dict = {
+        "place": process_place,
+        "building": process_building,
+        "address": process_address,
+    }
+
+    with open(args.input, "r", encoding="utf-8") as f:
+        contents: dict = json.load(f)
+        conf = float(kwargs.get("confidence", 0))
+        kwargs.pop("confidence", None)
+        geojson = process_geojson(
+            contents, fx=fx_dict.get(args.type, None), confidence=conf, options=kwargs
+        )
+
+    if args.in_place:
+        with open(args.input, "w+", encoding="utf-8") as f:
+            json.dump(geojson, f, indent=4)
+    else:
+        with open(args.output, "w+", encoding="utf-8") as f:
+            json.dump(geojson, f, indent=4)

src/overturetoosm/objects.py

@@ -380,4 +380,4 @@ def to_osm(self, style: str) -> Dict[str, str]:
 
 def source_statement(source: List[Sources]) -> str:
     """Return a source statement from a list of sources."""
-    return ", ".join(i.dataset.strip(", ") for i in source) + " via overturetoosm"
+    return ", ".join({i.dataset.strip(", ") for i in source}) + " via overturetoosm"