fix: collection title escaped unicode TDE-594

scripts/collection_from_items.py

@@ -5,7 +5,6 @@
 from boto3 import client
 from linz_logger import get_log
 
-from scripts.files.fs import write
 from scripts.files.fs_s3 import bucket_name_from_path, get_object_parallel_multithreading, list_json_in_uri
 from scripts.logging.time_helper import time_in_ms
 from scripts.stac.imagery.collection import ImageryCollection
@@ -62,7 +61,7 @@ def main() -> None:
     )
 
     destination = os.path.join(uri, "collection.json")
-    write(destination, json.dumps(collection.stac).encode("utf-8"))
+    collection.write_to(destination)
     get_log().info("collection written", destination=destination)
 
 

scripts/stac/imagery/collection.py

@@ -1,8 +1,10 @@
+import json
 from datetime import datetime
 from typing import Any, Dict, List, Optional
 
 import ulid
 
+from scripts.files.fs import write
 from scripts.stac.util.STAC_VERSION import STAC_VERSION
 
 
@@ -110,3 +112,6 @@ def update_extent(self, bbox: Optional[List[float]] = None, interval: Optional[L
             self.stac["extent"]["spatial"]["bbox"] = [bbox]
         if interval:
             self.stac["extent"]["temporal"]["interval"] = [interval]
+
+    def write_to(self, destination: str) -> None:
+        write(destination, json.dumps(self.stac, ensure_ascii=False).encode("utf-8"))

scripts/stac/tests/collection_test.py

@@ -1,7 +1,12 @@
+import json
+import os
+from shutil import rmtree
+from tempfile import mkdtemp
 from typing import Generator
 
 import pytest
 
+from scripts.files.fs import read
 from scripts.stac.imagery.collection import ImageryCollection
 from scripts.stac.imagery.item import ImageryItem
 
@@ -90,3 +95,25 @@ def test_add_item(mocker, setup_collection: ImageryCollection) -> None:  # type:
     assert {"rel": "item", "href": "./BR34_5000_0304.json", "type": "application/json"} in collection.stac["links"]
     assert collection.stac["extent"]["temporal"]["interval"] == [[start_datetime, end_datetime]]
     assert collection.stac["extent"]["spatial"]["bbox"] == [bbox]
+
+
+def test_write_collection(setup_collection: ImageryCollection) -> None:
+    target = mkdtemp()
+    collection_target = os.path.join(target, "collection.json")
+    setup_collection.write_to(collection_target)
+    collection = json.loads(read(collection_target))
+    rmtree(target)
+
+    assert collection["title"] == setup_collection.stac["title"]
+
+
+def test_write_collection_special_chars(setup_collection: ImageryCollection) -> None:
+    target = mkdtemp()
+    title = "Manawatū-Whanganui"
+    setup_collection.stac["title"] = title
+    collection_target = os.path.join(target, "collection.json")
+    setup_collection.write_to(collection_target)
+    collection = json.loads(read(collection_target))
+    rmtree(target)
+
+    assert collection["title"] == title