fix(waterdata): raise RequestTooLarge for an unchunkable over-budget request (not a silent 414)

dataretrieval/waterdata/chunking.py

@@ -871,13 +871,30 @@ def __init__(
         self.canonical_url: str | None = None
 
         axes = _extract_axes(args)
-        # No chunkable axes → skip ``build_request`` entirely; the
-        # common Water Data call shape shouldn't pay for an unused
-        # request prep on the passthrough hot path. The fetcher
-        # will run with the user's args verbatim; if that produces
-        # an over-budget URL, the server (or httpx itself) rejects.
         if not axes:
-            return
+            # No chunkable axis: nothing to split. If the single request fits,
+            # run it verbatim (the common passthrough). ``_safe_request_bytes``
+            # treats an un-constructable URL (httpx.InvalidURL, > 64 KB) as over
+            # budget.
+            if _safe_request_bytes(build_request, args, url_limit) <= url_limit:
+                return
+            # Over budget. A filter the chunker doesn't manage — cql-json — is
+            # passed through unchanged (chunking applies only to cql-text); the
+            # server, not us, judges it. Otherwise this is an in-domain shape we
+            # would normally chunk but can't (a single large CQL ``IN`` clause
+            # with no top-level ``OR``, or one oversized value), so raise an
+            # actionable error instead of shipping it for an opaque HTTP 414.
+            filter_expr = args.get("filter")
+            if filter_expr is not None and not _is_chunkable(
+                filter_expr, args.get("filter_lang")
+            ):
+                return
+            raise RequestTooLarge(
+                f"Request exceeds {url_limit} bytes (URL + body) and has no "
+                f"chunkable multi-value argument to split (e.g. a single large "
+                f"CQL `IN` clause, or one oversized value). Narrow the query, "
+                f"simplify the filter, or split the call manually."
+            )
 
         # Constructing the initial request can itself trip
         # ``httpx.InvalidURL`` (URL > 64 KB) — that's the canonical

tests/waterdata_chunking_test.py

@@ -161,10 +161,30 @@ def test_extract_axes_skips_singletons_and_never_chunk_params():
 
 
 def test_chunk_plan_returns_passthrough_when_no_chunkable_axes():
-    """Scalar args with nothing to chunk → passthrough, even at a
-    URL limit the request technically exceeds (the server may 414,
-    but ``ChunkPlan`` has nothing to split)."""
+    """Scalar args with nothing to chunk and a request within the limit →
+    passthrough (no axes)."""
     args = {"monitoring_location_id": "scalar-only"}
+    plan = ChunkPlan(args, _fake_build, url_limit=1000)
+    assert plan.axes == []
+    assert plan.total == 1
+
+
+def test_chunk_plan_raises_when_unchunkable_request_exceeds_limit():
+    """A request with nothing to chunk that still exceeds the byte limit (e.g.
+    a single large CQL ``IN`` clause with no top-level ``OR``) raises
+    RequestTooLarge instead of being shipped for the server to reject with an
+    opaque HTTP 414."""
+    args = {"monitoring_location_id": "scalar-only"}
+    with pytest.raises(RequestTooLarge):
+        ChunkPlan(args, _fake_build, url_limit=10)
+
+
+def test_chunk_plan_passes_through_unchunkable_cql_json_over_limit():
+    """A cql-json filter is outside the chunker's domain (it splits only
+    cql-text), so an over-budget cql-json request is passed through unchanged
+    instead of raising — the server judges it, not us. Guards against the
+    chunker hijacking the deliberate cql-json passthrough."""
+    args = {"filter": "a OR b OR c", "filter_lang": "cql-json"}
     plan = ChunkPlan(args, _fake_build, url_limit=10)
     assert plan.axes == []
     assert plan.total == 1