fix: settings env var regex and split

private_gpt/settings/yaml.py

@@ -5,12 +5,12 @@
 
 from yaml import SafeLoader
 
-_env_replace_matcher = re.compile(r"\$\{(\w|_)+(:(\w|_)*)?}")
+_env_replace_matcher = re.compile(r"\$\{(\w|_)+:?.*}")
 
 
 @typing.no_type_check  # pyaml does not have good hints, everything is Any
 def load_yaml_with_envvars(
-    stream: TextIO, environ: dict[str, Any] = os.environ
+        stream: TextIO, environ: dict[str, Any] = os.environ
 ) -> dict[str, Any]:
     """Load yaml file with environment variable expansion.
 
@@ -22,7 +22,7 @@ def load_yaml_with_envvars(
     def load_env_var(_, node) -> str:
         """Extract the matched value, expand env variable, and replace the match."""
         value = str(node.value).removeprefix("${").removesuffix("}")
-        split = value.split(":")
+        split = value.split(":", 1)
         env_var = split[0]
         value = environ.get(env_var)
         default = None if len(split) == 1 else split[1]

tests/settings/test_settings_loader.py

@@ -18,10 +18,19 @@ def test_environment_variables_are_loaded() -> None:
 
 def test_environment_defaults_variables_are_loaded() -> None:
     sample_yaml = """
-    replaced: ${TEST_REPLACE_ME:default}
+    replaced: ${PGPT_EMBEDDING_HF_MODEL_NAME:BAAI/bge-small-en-v1.5}
     """
     loaded = load_yaml_with_envvars(io.StringIO(sample_yaml), {})
-    assert loaded["replaced"] == "default"
+    assert loaded["replaced"] == "BAAI/bge-small-en-v1.5"
+
+
+def test_environment_defaults_variables_are_loaded_with_duplicated_delimiters() -> None:
+    sample_yaml = """
+    replaced: ${PGPT_EMBEDDING_HF_MODEL_NAME::duped::}
+    """
+    loaded = load_yaml_with_envvars(io.StringIO(sample_yaml), {})
+    assert loaded["replaced"] == ":duped::"
+
 
 
 def test_environment_without_defaults_fails() -> None: