feat: upgrade ollama client to 0.4.5 (#345)

Co-authored-by: danran <[email protected]>

projects/extension/requirements.txt

@@ -1,6 +1,6 @@
 openai==1.44.0
 tiktoken==0.7.0
-ollama==0.2.1
+ollama==0.4.5
 anthropic==0.29.0
 cohere==5.5.8
 backoff==2.2.1

projects/extension/sql/idempotent/002-ollama.sql

@@ -162,7 +162,7 @@ as $python$
         args["images"] = images_1
 
     resp = client.generate(model, prompt, stream=False, **args)
-    return json.dumps(resp)
+    return resp.model_dump_json()
 $python$
 language plpython3u volatile parallel safe security invoker
 set search_path to pg_catalog, pg_temp
@@ -206,7 +206,7 @@ as $python$
 
     resp = client.chat(model, messages_1, stream=False, **args)
 
-    return json.dumps(resp)
+    return resp.model_dump_json()
 $python$
 language plpython3u volatile parallel safe security invoker
 set search_path to pg_catalog, pg_temp

projects/pgai/pgai/vectorizer/embedders/ollama.py

@@ -1,12 +1,10 @@
 import os
-from collections.abc import Mapping, Sequence
+from collections.abc import Sequence
 from functools import cached_property
-from typing import (
-    Any,
-    Literal,
-)
+from typing import Literal
 
 import ollama
+from ollama import ShowResponse
 from pydantic import BaseModel
 from typing_extensions import TypedDict, override
 
@@ -130,7 +128,7 @@ async def call_embed_api(self, documents: str | list[str]) -> EmbeddingResponse:
         )
         return EmbeddingResponse(embeddings=response["embeddings"], usage=usage)
 
-    async def _model(self) -> Mapping[str, Any]:
+    async def _model(self) -> ShowResponse:
         """
         Gets the model details from the Ollama API
         :return:

projects/pgai/pyproject.toml

@@ -15,7 +15,7 @@ dependencies = [
     "langchain-openai>=0.1,<1.0",
     "langchain-text-splitters>=0.2,<1.0",
     "pydantic>=2.0,<3.0",
-    "ollama>=0.3.3,<0.4.0",
+    "ollama>=0.4.5,<0.5.0",
     "openai>=1.44,<2.0",
     "python-dotenv>=1.0,<2.0",
     "structlog>=24.0,<25.0",