Merge commit '72b090da2c50e540143fd312a2f9aa5f151e6136' into concedo_experimental

# Conflicts:
#	docs/backend/CANN.md
#	docs/function-calling.md
#	examples/embedding/embedding.cpp
#	examples/retrieval/retrieval.cpp
#	ggml/src/ggml-cann/CMakeLists.txt
#	ggml/src/ggml-cann/Doxyfile
#	ggml/src/ggml-cann/acl_tensor.cpp
#	ggml/src/ggml-cann/acl_tensor.h
#	ggml/src/ggml-cann/aclnn_ops.cpp
#	ggml/src/ggml-cann/aclnn_ops.h
#	ggml/src/ggml-cann/common.h
#	ggml/src/ggml-cann/ggml-cann.cpp
#	ggml/src/ggml-cpu/CMakeLists.txt
#	ggml/src/ggml-sycl/binbcast.cpp
#	ggml/src/ggml-sycl/common.hpp
#	ggml/src/ggml-sycl/concat.cpp
#	ggml/src/ggml-sycl/conv.cpp
#	ggml/src/ggml-sycl/cpy.cpp
#	ggml/src/ggml-sycl/dmmv.cpp
#	ggml/src/ggml-sycl/element_wise.cpp
#	ggml/src/ggml-sycl/getrows.cpp
#	ggml/src/ggml-sycl/ggml-sycl.cpp
#	ggml/src/ggml-sycl/gla.cpp
#	ggml/src/ggml-sycl/mmvq.cpp
#	ggml/src/ggml-sycl/norm.cpp
#	ggml/src/ggml-sycl/outprod.cpp
#	ggml/src/ggml-sycl/rope.cpp
#	ggml/src/ggml-sycl/softmax.cpp
#	ggml/src/ggml-sycl/tsembd.cpp
#	ggml/src/ggml-sycl/wkv.cpp
#	scripts/compare-commits.sh
#	tests/test-chat.cpp
#	tests/test-sampling.cpp

tools/server/tests/unit/test_completion.py

@@ -121,6 +121,30 @@ def test_completion_stream_with_openai_library():
     assert match_regex("(going|bed)+", output_text)
 
 
+# Test case from https://github.com/ggml-org/llama.cpp/issues/13780
+@pytest.mark.slow
+def test_completion_stream_with_openai_library_stops():
+    global server
+    server.model_hf_repo = "bartowski/Phi-3.5-mini-instruct-GGUF:Q4_K_M"
+    server.model_hf_file = None
+    server.start()
+    client = OpenAI(api_key="dummy", base_url=f"http://{server.server_host}:{server.server_port}/v1")
+    res = client.completions.create(
+        model="davinci-002",
+        prompt="System: You are helpfull assistant.\nAssistant:\nHey! How could I help?\nUser:\nTell me a joke.\nAssistant:\n",
+        stop=["User:\n", "Assistant:\n"],
+        max_tokens=200,
+        stream=True,
+    )
+    output_text = ''
+    for data in res:
+        choice = data.choices[0]
+        if choice.finish_reason is None:
+            assert choice.text is not None
+            output_text += choice.text
+    assert match_regex("Sure, here's one for[\\s\\S]*", output_text), f'Unexpected output: {output_text}'
+
+
 @pytest.mark.parametrize("n_slots", [1, 2])
 def test_consistent_result_same_seed(n_slots: int):
     global server

tools/server/tests/utils.py

@@ -328,6 +328,10 @@ class ServerProcess:
                     if 'function' not in tc:
                         raise ValueError(f"Expected function type, got {tc['type']}")
                     if tc['index'] >= len(tool_calls):
+                        assert 'id' in tc
+                        assert tc.get('type') == 'function'
+                        assert 'function' in tc and 'name' in tc['function'] and len(tc['function']['name']) > 0, \
+                            f"Expected function call with name, got {tc.get('function')}"
                         tool_calls.append(dict(
                             id="",
                             type="function",
@@ -340,10 +344,10 @@ class ServerProcess:
                     if tc.get('id') is not None:
                         tool_call['id'] = tc['id']
                     fct = tc['function']
+                    assert 'id' not in fct, f"Function call should not have id: {fct}"
                     if fct.get('name') is not None:
-                        tool_call['function']['name'] = fct['name']
+                        tool_call['function']['name'] = tool_call['function'].get('name', '') + fct['name']
                     if fct.get('arguments') is not None:
-                        assert len(fct['arguments']) > 0, f'Expected non empty arguments delta!'
                         tool_call['function']['arguments'] += fct['arguments']
 
             print(f'Streamed response had {content_parts} content parts, {tool_call_parts} tool call parts incl. {arguments_parts} arguments parts')