From 4ce6bb37dd11ad69c455214fbd0b28f32f50de44 Mon Sep 17 00:00:00 2001
From: "Stephan T. Lavavej" <stl@microsoft.com>
Date: Fri, 15 Mar 2024 22:45:55 -0700
Subject: [PATCH] `download_unicode_data_files.py`: Select latest or a specific
 version (#4469)

---
 .../download_unicode_data_files.py            | 46 ++++++++++++++-----
 1 file changed, 35 insertions(+), 11 deletions(-)

diff --git a/tools/unicode_properties_parse/download_unicode_data_files.py b/tools/unicode_properties_parse/download_unicode_data_files.py
index bf5c587fd1..7fada90868 100644
--- a/tools/unicode_properties_parse/download_unicode_data_files.py
+++ b/tools/unicode_properties_parse/download_unicode_data_files.py
@@ -1,22 +1,46 @@
 # Copyright (c) Microsoft Corporation.
 # SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 
+from pathlib import PurePosixPath
+import sys
+from urllib.error import HTTPError
 from urllib.request import urlretrieve
 
 
-Unicode_data_files = {
-    "DerivedCoreProperties.txt": "https://www.unicode.org/Public/UCD/latest/ucd/DerivedCoreProperties.txt",
-    "DerivedGeneralCategory.txt": "https://www.unicode.org/Public/UCD/latest/ucd/extracted/DerivedGeneralCategory.txt",
-    "EastAsianWidth.txt": "https://www.unicode.org/Public/UCD/latest/ucd/EastAsianWidth.txt",
-    "GraphemeBreakProperty.txt": "https://www.unicode.org/Public/UCD/latest/ucd/auxiliary/GraphemeBreakProperty.txt",
-    "GraphemeBreakTest.txt": "https://www.unicode.org/Public/UCD/latest/ucd/auxiliary/GraphemeBreakTest.txt",
-    "emoji-data.txt": "https://www.unicode.org/Public/UCD/latest/ucd/emoji/emoji-data.txt",
-}
+def get_base_url():
+    if len(sys.argv) != 2:
+        sys.exit(f"Usage: python {sys.argv[0]} [latest|<VERSION LIKE 15.0.0>]")
+
+    version = sys.argv[1]
+
+    if version == "latest":
+        return "https://unicode.org/Public/UCD/latest/"
+
+    return f"https://unicode.org/Public/{version}/"
+
+
+Unicode_data_files = [
+    "ucd/DerivedCoreProperties.txt",
+    "ucd/extracted/DerivedGeneralCategory.txt",
+    "ucd/EastAsianWidth.txt",
+    "ucd/auxiliary/GraphemeBreakProperty.txt",
+    "ucd/auxiliary/GraphemeBreakTest.txt",
+    "ucd/emoji/emoji-data.txt",
+]
+
 
 def download_unicode_data_files():
-    for filename, url in Unicode_data_files.items():
-        print(f"downloading {filename} from {url}")
-        urlretrieve(url, filename)
+    base_url = get_base_url()
+    print(f"   Base URL: {base_url}")
+
+    for data_file in Unicode_data_files:
+        url = base_url + data_file
+        filename = PurePosixPath(data_file).name
+        print(f"Downloading: {url}")
+        try:
+            urlretrieve(url, filename)
+        except HTTPError as http_error:
+            sys.exit(f"{http_error}")
 
 
 if __name__ == "__main__":