ui.Chat() now correctly handles new ollama.chat() return value introduced in ollama 0.4

CHANGELOG.md

@@ -5,6 +5,12 @@ All notable changes to Shiny for Python will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [UNRELEASED]
+
+### Bug fixes
+
+* `ui.Chat()` now correctly handles new `ollama.chat()` return value introduced in `ollama` v0.4. (#1787)
+
 ## [1.2.1] - 2024-11-14
 
 ### Bug fixes

shiny/templates/chat/hello-providers/ollama/app.py

@@ -29,7 +29,7 @@ async def _():
     # Create a response message stream
     # Assumes you've run `ollama run llama3` to start the server
     response = ollama.chat(
-        model="llama3",
+        model="llama3.2",
         messages=messages,
         stream=True,
     )

shiny/ui/_chat.py

@@ -914,15 +914,6 @@ def _get_token_count(
         if self._tokenizer is None:
             self._tokenizer = get_default_tokenizer()
 
-        if self._tokenizer is None:
-            raise ValueError(
-                "A tokenizer is required to impose `token_limits` on messages. "
-                "To get a generic default tokenizer, install the `tokenizers` "
-                "package (`pip install tokenizers`). "
-                "To get a more precise token count, provide a specific tokenizer "
-                "to the `Chat` constructor."
-            )
-
         encoded = self._tokenizer.encode(content)
         if isinstance(encoded, TokenizersEncoding):
             return len(encoded.ids)

shiny/ui/_chat_normalize.py

@@ -231,11 +231,19 @@ def normalize_chunk(self, chunk: "dict[str, Any]") -> ChatMessage:
         return super().normalize_chunk(msg)
 
     def can_normalize(self, message: Any) -> bool:
-        if not isinstance(message, dict):
-            return False
-        if "message" not in message:
+        try:
+            from ollama import ChatResponse
+
+            # Ollama<0.4 used TypedDict (now it uses pydantic)
+            # https://github.com/ollama/ollama-python/pull/276
+            if isinstance(ChatResponse, dict):
+                return "message" in message and super().can_normalize(
+                    message["message"]
+                )
+            else:
+                return isinstance(message, ChatResponse)
+        except Exception:
             return False
-        return super().can_normalize(message["message"])
 
     def can_normalize_chunk(self, chunk: Any) -> bool:
         return self.can_normalize(chunk)

shiny/ui/_chat_tokenizer.py

@@ -45,12 +45,23 @@ def encode(
 TokenEncoding = Union[TiktokenEncoding, TokenizersTokenizer]
 
 
-def get_default_tokenizer() -> TokenizersTokenizer | None:
+def get_default_tokenizer() -> TokenizersTokenizer:
     try:
         from tokenizers import Tokenizer
 
         return Tokenizer.from_pretrained("bert-base-cased")  # type: ignore
-    except Exception:
-        pass
-
-    return None
+    except ImportError:
+        raise ValueError(
+            "Failed to download a default tokenizer. "
+            "A tokenizer is required to impose `token_limits` on `chat.messages()`. "
+            "To get a generic default tokenizer, install the `tokenizers` "
+            "package (`pip install tokenizers`). "
+        )
+    except Exception as e:
+        raise ValueError(
+            "Failed to download a default tokenizer. "
+            "A tokenizer is required to impose `token_limits` on `chat.messages()`. "
+            "Try downloading a different tokenizer using "
+            "`tokenizers.Tokenizer.from_pretrained()`. "
+            f"Error: {e}"
+        ) from e

tests/pytest/test_chat.py

@@ -333,6 +333,20 @@ def test_openai_normalization():
     assert msg == {"content": "Hello ", "role": "assistant"}
 
 
+def test_ollama_normalization():
+    from ollama import ChatResponse
+    from ollama import Message as OllamaMessage
+
+    # Mock return object from ollama.chat()
+    msg = ChatResponse(
+        message=OllamaMessage(content="Hello world!", role="assistant"),
+    )
+
+    msg_dict = {"content": "Hello world!", "role": "assistant"}
+    assert normalize_message(msg) == msg_dict
+    assert normalize_message_chunk(msg) == msg_dict
+
+
 # ------------------------------------------------------------------------------------
 # Unit tests for as_provider_message()
 #
@@ -462,9 +476,7 @@ def test_as_ollama_message():
     import ollama
     from ollama import Message as OllamaMessage
 
-    assert "typing.Sequence[ollama._types.Message]" in str(
-        ollama.chat.__annotations__["messages"]
-    )
+    assert "ollama._types.Message" in str(ollama.chat.__annotations__["messages"])
 
     from shiny.ui._chat_provider_types import as_ollama_message