fix: forward stream_metadata to the chat API in legacy client

chat_completions(stream=True, stream_metadata=True) used stream_metadata
only to pick the raw dict-chunk parser locally, never adding it to the
request payload. The wire request didn't match the caller's intent and
relied on the server sending metadata chunks unconditionally. Forward the
flag (mirroring the modern CloudBackend, which always sends it) so the
request is correct and robust if the server ever gates metadata behind it.

Verified against the real API that the server currently emits block_metadata
regardless, so this is a latent-correctness fix, not a behavior change today.

pageindex/cloud_api.py

@@ -147,6 +147,12 @@ class LegacyCloudAPI:
             payload["temperature"] = temperature
         if enable_citations:
             payload["enable_citations"] = enable_citations
+        # Forward stream_metadata so the wire request matches the caller's intent
+        # (and stays correct if the server ever gates metadata chunks behind it),
+        # mirroring the modern CloudBackend which always sends it. It only affects
+        # streaming responses, where it selects the raw dict-chunk parser below.
+        if stream_metadata:
+            payload["stream_metadata"] = stream_metadata
 
         response = self._request(
             "POST",

tests/test_legacy_sdk_contract.py

@@ -238,7 +238,10 @@ def test_chat_completions_stream_metadata_returns_raw_chunks(monkeypatch):
     ))
 
     assert chunks == [{"object": "chat.completion.chunk"}]
-    assert "stream_metadata" not in calls[0]["json"]
+    # stream_metadata must be forwarded to the server so the wire request matches
+    # the caller's intent (and mirrors the modern CloudBackend), not kept as a
+    # client-only parser switch.
+    assert calls[0]["json"]["stream_metadata"] is True
 
 
 def test_chat_completions_stream_errors_are_pageindex_api_error(monkeypatch):