feat: add client.exists() to check if a model is available locally

ollama/__init__.py

@@ -52,6 +52,7 @@
 create = _client.create
 delete = _client.delete
 list = _client.list
+exists = _client.exists
 copy = _client.copy
 show = _client.show
 ps = _client.ps

ollama/_client.py

@@ -629,6 +629,18 @@ def list(self) -> ListResponse:
       '/api/tags',
     )
 
+  def exists(self, model: str) -> bool:
+    """Check if a model is available locally.
+
+    Args:
+      model: The model name to check (e.g. 'llama3.1:8b').
+
+    Returns:
+      True if the model exists locally, False otherwise.
+    """
+    models = self.list().models or []
+    return any(m.model == model for m in models)
+
   def delete(self, model: str) -> StatusResponse:
     r = self._request_raw(
       'DELETE',
@@ -1270,6 +1282,19 @@ async def list(self) -> ListResponse:
       '/api/tags',
     )
 
+  async def exists(self, model: str) -> bool:
+    """Check if a model is available locally.
+
+    Args:
+      model: The model name to check (e.g. 'llama3.1:8b').
+
+    Returns:
+      True if the model exists locally, False otherwise.
+    """
+    resp = await self.list()
+    models = resp.models or []
+    return any(m.model == model for m in models)
+
   async def delete(self, model: str) -> StatusResponse:
     r = await self._request_raw(
       'DELETE',

tests/test_exists.py

@@ -0,0 +1,82 @@
+import json
+
+import pytest
+from pytest_httpserver import HTTPServer
+
+from ollama._client import AsyncClient, Client
+
+
+pytestmark = pytest.mark.anyio
+
+
+@pytest.fixture
+def anyio_backend():
+  return 'asyncio'
+
+
+def test_client_exists_true(httpserver: HTTPServer):
+  """exists() returns True when model is present."""
+  httpserver.expect_ordered_request(
+    '/api/tags',
+    method='GET',
+  ).respond_with_json({
+    'models': [
+      {'name': 'llama3.1:8b', 'model': 'llama3.1:8b', 'size': 4661224676},
+      {'name': 'qwen2.5:latest', 'model': 'qwen2.5:latest', 'size': 4430121000},
+    ]
+  })
+
+  client = Client(host=f'http://{httpserver.host}:{httpserver.port}')
+  assert client.exists('llama3.1:8b') is True
+
+
+def test_client_exists_false(httpserver: HTTPServer):
+  """exists() returns False when model is not present."""
+  httpserver.expect_ordered_request(
+    '/api/tags',
+    method='GET',
+  ).respond_with_json({
+    'models': [
+      {'name': 'llama3.1:8b', 'model': 'llama3.1:8b', 'size': 4661224676},
+    ]
+  })
+
+  client = Client(host=f'http://{httpserver.host}:{httpserver.port}')
+  assert client.exists('gemma2:2b') is False
+
+
+def test_client_exists_empty_list(httpserver: HTTPServer):
+  """exists() returns False when no models are available."""
+  httpserver.expect_ordered_request(
+    '/api/tags',
+    method='GET',
+  ).respond_with_json({'models': []})
+
+  client = Client(host=f'http://{httpserver.host}:{httpserver.port}')
+  assert client.exists('llama3.1:8b') is False
+
+
+async def test_async_client_exists_true(httpserver: HTTPServer):
+  """Async exists() returns True when model is present."""
+  httpserver.expect_ordered_request(
+    '/api/tags',
+    method='GET',
+  ).respond_with_json({
+    'models': [
+      {'name': 'llama3.1:8b', 'model': 'llama3.1:8b', 'size': 4661224676},
+    ]
+  })
+
+  client = AsyncClient(host=f'http://{httpserver.host}:{httpserver.port}')
+  assert await client.exists('llama3.1:8b') is True
+
+
+async def test_async_client_exists_false(httpserver: HTTPServer):
+  """Async exists() returns False when model is not present."""
+  httpserver.expect_ordered_request(
+    '/api/tags',
+    method='GET',
+  ).respond_with_json({'models': []})
+
+  client = AsyncClient(host=f'http://{httpserver.host}:{httpserver.port}')
+  assert await client.exists('nonexistent') is False