Encode unchecked multi-value filters as repeated query parameters

CHANGELOG.md

@@ -8,6 +8,12 @@ adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
 ## [Unreleased]
 
+### Fixed
+
+- Expand unchecked iterable query values and tuple sequence parameters into
+  repeated array keys instead of sending their Python tuple representation.
+  Existing array suffixes are retained without duplication. (#123)
+
 ### Added
 
 - Add method `Query.results` for returning results as an iterator instead of sequence ([#37](https://github.com/nasa/python_cmr/issues/37))

cmr/queries.py

@@ -188,6 +188,9 @@ def parameters(self, **kwargs: Any) -> Self:
         Provide query parameters as keyword arguments. The keyword needs to match the name
         of the method, and the value should either be the value or a tuple of values.
 
+        Iterable values for parameters without a dedicated method are sent as
+        repeated array query parameters, one value per key.
+
         Example: parameters(short_name="AST_L1T", point=(42.5, -101.25))
 
         :returns: self
@@ -252,9 +255,10 @@ def _build_url(self) -> str:
         for key, val in self.params.items():
 
             # list params require slightly different formatting
-            if isinstance(val, list):
+            if isinstance(val, (list, tuple)):
+                array_key = key if key.endswith("[]") else f"{key}[]"
                 for list_val in val:
-                    formatted_params.append(f"{key}[]={list_val}")
+                    formatted_params.append(f"{array_key}={list_val}")
 
             elif isinstance(val, bool):
                 formatted_params.append(f"{key}={str(val).lower()}")

tests/test_multi_value_parameters.py

@@ -0,0 +1,105 @@
+"""Exercise repeated query parameters through URL construction and requests."""
+
+import json
+from urllib.parse import parse_qs, urlsplit
+
+import pytest
+import requests
+
+from cmr import (
+    CollectionQuery, GranuleQuery, ServiceQuery, ToolQuery, VariableQuery,
+)
+
+
+@pytest.mark.parametrize("query_type", [
+    CollectionQuery, GranuleQuery, ServiceQuery, ToolQuery, VariableQuery,
+])
+def test_unchecked_values_are_individual_query_parameters(query_type):
+    query = query_type().parameters(data_center=["FIRST", "SECOND"])
+    original = dict(query.params)
+    assert parse_qs(urlsplit(query._build_url()).query) == {
+        "data_center[]": ["FIRST", "SECOND"],
+    }
+    assert query.params == original
+    assert query.params["data_center[]"] == ("FIRST", "SECOND")
+
+
+@pytest.mark.parametrize("factory", [list, tuple, iter, set])
+def test_unchecked_iterables_retain_values_on_repeated_builds(factory):
+    query = CollectionQuery().parameters(data_center=factory(["A", "B"]))
+    first = query._build_url()
+    assert query._build_url() == first
+    assert sorted(parse_qs(urlsplit(first).query)["data_center[]"]) == [
+        "A", "B",
+    ]
+
+
+@pytest.mark.parametrize("values,expected", [
+    ([], {}),
+    (["A"], {"data_center[]": ["A"]}),
+    (["A", "A"], {"data_center[]": ["A", "A"]}),
+    (range(2), {"data_center[]": ["0", "1"]}),
+])
+def test_empty_singleton_duplicate_and_numeric_values(values, expected):
+    query = CollectionQuery().parameters(data_center=values)
+    assert parse_qs(urlsplit(query._build_url()).query) == expected
+
+
+@pytest.mark.parametrize("values", [["one", "two"], ("one", "two")])
+def test_declared_sequence_parameter_keeps_one_array_suffix(values):
+    query = VariableQuery().instance_format(values)
+    assert parse_qs(urlsplit(query._build_url()).query) == {
+        "instance_format[]": ["one", "two"],
+    }
+
+
+def test_scalar_parameters_and_tuple_method_arguments_are_unchanged():
+    query = GranuleQuery().parameters(
+        short_name="EXAMPLE", point=(-100, 42),
+        custom_enabled=True, custom_limit=0,
+    )
+    assert parse_qs(urlsplit(query._build_url()).query) == {
+        "short_name": ["EXAMPLE"], "point[]": ["-100.0,42.0"],
+        "custom_enabled": ["true"], "custom_limit": ["0"],
+    }
+
+
+@pytest.mark.parametrize("operation", ["results", "get", "hits"])
+def test_repeated_values_reach_the_http_request(monkeypatch, operation):
+    sent = []
+
+    def send(session, request, **kwargs):
+        sent.append(request)
+        response = requests.Response()
+        response.status_code = 200
+        response.request = request
+        response.headers["CMR-Hits"] = "2"
+        response._content = json.dumps({
+            "feed": {"entry": [{"id": len(sent)}]},
+        }).encode()
+        if len(sent) == 1:
+            response.headers["cmr-search-after"] = "next-page"
+        return response
+
+    monkeypatch.setattr(requests.Session, "send", send)
+    query = CollectionQuery().parameters(data_center=["FIRST", "SECOND"])
+    query.headers["Client-Id"] = "local-test"
+    original_headers = dict(query.headers)
+    if operation == "results":
+        assert list(query.results(page_size=1)) == [{"id": 1}, {"id": 2}]
+    elif operation == "get":
+        with pytest.warns(DeprecationWarning):
+            assert query.get(1) == [{"id": 1}]
+    else:
+        assert query.hits() == 2
+    for request in sent:
+        params = parse_qs(urlsplit(request.url).query)
+        assert params["data_center[]"] == ["FIRST", "SECOND"]
+        assert params["page_size"] == (["0"] if operation == "hits" else ["1"])
+        assert "data_center[][]" not in params
+        assert request.headers["Client-Id"] == "local-test"
+    if operation == "results":
+        assert len(sent) == 2
+        assert "cmr-search-after" not in sent[0].headers
+        assert sent[1].headers["cmr-search-after"] == "next-page"
+    assert query.headers == original_headers