docs: doc_id is per-call table-setting — keep it identical across a conversation

The targeting block doc_id adds is re-set on every call and sits in the
cached prompt prefix, so a round-trip that drops (or changes) doc_id
silently diverges the prefix and loses the cache continuation. State the
rule on all three chat surfaces' doc_id docs, and pin it with a prefix
test that passes the same doc_id on both calls.

pageindex/client.py

@@ -368,6 +368,9 @@ class PageIndexClient:
                 is appended to the managed system prompt.
             stream: Enable streaming responses.
             doc_id: Document ID or list of IDs to scope the conversation.
+                Keep it identical across a conversation's calls — the
+                targeting block it adds is re-set each call and is part
+                of the cached prompt prefix.
             temperature: Sampling temperature, passed through to the model.
             stream_metadata: With stream=True, yield chunk dicts instead of
                 text pieces.
@@ -438,6 +441,9 @@ class PageIndexClient:
                 tool outputs are emitted as ``response.output_item.done``
                 events and the single final event is ``response.completed``.
             doc_id: Document ID or list of IDs to scope the conversation.
+                Keep it identical across a conversation's calls — the
+                targeting block it adds is re-set each call and is part
+                of the cached prompt prefix.
             instructions: Appended to the managed system prompt.
             temperature / top_p: Passed through to the model.
             max_turns: Cap on agent turns per call.
@@ -489,6 +495,9 @@ class PageIndexClient:
                 (its native event objects, including SDK-synthesized
                 convenience events), one message sequence per turn.
             doc_id: Document ID or list of IDs to scope the conversation.
+                Keep it identical across a conversation's calls — the
+                targeting block it adds is re-set each call and is part
+                of the cached prompt prefix.
             system: Appended after the managed system blocks.
             temperature / top_p / top_k / stop_sequences: Passed through.
             max_turns: Cap on agent turns per call (default 10, like the

tests/test_local_chat.py

@@ -302,6 +302,25 @@ def test_responses_round_trip_extends_prefix(client, store_path, fake_model):
     assert second.inputs[0][:len(previous_final)] == previous_final
 
 
+def test_responses_round_trip_prefix_with_doc_id(client, store_path, fake_model):
+    """Same contract with doc targeting: re-passing the same doc_id re-sets
+    an identical leading block, so the prefix still extends item-for-item."""
+    seed_doc(store_path, "pi-a", "report.pdf")
+    first = fake_model([
+        [_call_item("get_document", {"doc_name": "report.pdf"})],
+        [_msg_item("The answer")],
+    ])
+    result = client.responses("What status?", doc_id="pi-a")
+
+    second = fake_model([[_msg_item("Done")]])
+    follow_up = ([{"role": "user", "content": "What status?"}]
+                 + result["output"]
+                 + [{"role": "user", "content": "and now?"}])
+    client.responses(follow_up, doc_id="pi-a")
+    previous_final = first.inputs[-1]
+    assert second.inputs[0][:len(previous_final)] == previous_final
+
+
 @needs_agents
 def test_responses_stream_passthrough(client, store_path, fake_model):
     seed_doc(store_path, "pi-a", "report.pdf")