pydantic-ai 0.1.1__tar.gz → 0.1.3__tar.gz

{pydantic_ai-0.1.1 → pydantic_ai-0.1.3}/Makefile

@@ -80,7 +80,7 @@ testcov: test ## Run tests and generate a coverage report
 
 .PHONY: test-mrp
 test-mrp: ## Build and  tests of mcp-run-python
-	cd mcp-run-python && npm run prepare
+	cd mcp-run-python && deno task build
 	uv run --package mcp-run-python pytest mcp-run-python -v
 
 .PHONY: update-examples

{pydantic_ai-0.1.1 → pydantic_ai-0.1.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pydantic-ai
-Version: 0.1.1
+Version: 0.1.3
 Summary: Agent Framework / shim to use Pydantic with LLMs
 Project-URL: Homepage, https://ai.pydantic.dev
 Project-URL: Source, https://github.com/pydantic/pydantic-ai
@@ -28,9 +28,9 @@ Classifier: Topic :: Internet
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
 Classifier: Topic :: Software Development :: Libraries :: Python Modules
 Requires-Python: >=3.9
-Requires-Dist: pydantic-ai-slim[anthropic,bedrock,cli,cohere,evals,groq,mcp,mistral,openai,vertexai]==0.1.1
+Requires-Dist: pydantic-ai-slim[anthropic,bedrock,cli,cohere,evals,groq,mcp,mistral,openai,vertexai]==0.1.3
 Provides-Extra: examples
-Requires-Dist: pydantic-ai-examples==0.1.1; extra == 'examples'
+Requires-Dist: pydantic-ai-examples==0.1.3; extra == 'examples'
 Provides-Extra: logfire
 Requires-Dist: logfire>=3.11.0; extra == 'logfire'
 Description-Content-Type: text/markdown

pydantic_ai-0.1.3/tests/models/cassettes/test_bedrock/test_bedrock_multiple_documents_in_history.yaml

@@ -0,0 +1,52 @@
+interactions:
+- request:
+    body: '{"messages": [{"role": "user", "content": [{"text": "Here is a PDF document: "}, {"document": {"name": "Document
+      1", "format": "pdf", "source": {"bytes": "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"}}}]},
+      {"role": "assistant", "content": [{"text": "foo bar"}]}, {"role": "user", "content": [{"text": "Here is another PDF
+      document: "}, {"document": {"name": "Document 2", "format": "pdf", "source": {"bytes": "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"}}}]},
+      {"role": "assistant", "content": [{"text": "foo bar 2"}]}, {"role": "user", "content": [{"text": "What is in the documents?"}]}],
+      "system": [], "inferenceConfig": {}}'
+    headers:
+      amz-sdk-invocation-id:
+      - !!binary |
+        NjNjMWFiZTUtMjVhMS00M2Y0LWE3ZWQtZjE3MmYwODllNWFl
+      amz-sdk-request:
+      - !!binary |
+        YXR0ZW1wdD0x
+      content-length:
+      - '35917'
+      content-type:
+      - !!binary |
+        YXBwbGljYXRpb24vanNvbg==
+    method: POST
+    uri: https://bedrock-runtime.us-east-1.amazonaws.com/model/us.anthropic.claude-3-7-sonnet-20250219-v1%3A0/converse
+  response:
+    headers:
+      connection:
+      - keep-alive
+      content-length:
+      - '506'
+      content-type:
+      - application/json
+    parsed_body:
+      metrics:
+        latencyMs: 1820
+      output:
+        message:
+          content:
+          - text: Based on the documents you've shared, both Document 1.pdf and Document 2.pdf contain the text "Dummy PDF
+              file". These appear to be placeholder or sample PDF documents rather than files with substantial content.
+          role: assistant
+      stopReason: end_turn
+      usage:
+        cacheReadInputTokenCount: 0
+        cacheReadInputTokens: 0
+        cacheWriteInputTokenCount: 0
+        cacheWriteInputTokens: 0
+        inputTokens: 145
+        outputTokens: 48
+        totalTokens: 193
+    status:
+      code: 200
+      message: OK
+version: 1

{pydantic_ai-0.1.1 → pydantic_ai-0.1.3}/tests/models/test_bedrock.py

@@ -30,13 +30,12 @@ from pydantic_ai.messages import (
     UserPromptPart,
     VideoUrl,
 )
-from pydantic_ai.models.bedrock import BedrockModelSettings
 from pydantic_ai.usage import Usage
 
 from ..conftest import IsDatetime, try_import
 
 with try_import() as imports_successful:
-    from pydantic_ai.models.bedrock import BedrockConverseModel
+    from pydantic_ai.models.bedrock import BedrockConverseModel, BedrockModelSettings
     from pydantic_ai.providers.bedrock import BedrockProvider
 
 pytestmark = [
@@ -560,3 +559,25 @@ async def test_bedrock_empty_system_prompt(allow_model_requests: None, bedrock_p
     assert result.output == snapshot(
         'The capital of France is Paris. Paris, officially known as "Ville de Paris," is not only the capital city but also the most populous city in France. It is located in the northern central part of the country along the Seine River. Paris is a major global city, renowned for its cultural, political, economic, and social influence. It is famous for its landmarks such as the Eiffel Tower, the Louvre Museum, Notre-Dame Cathedral, and the Champs-Élysées, among many other historic and modern attractions. The city has played a significant role in the history of art, fashion, gastronomy, and science.'
     )
+
+
+@pytest.mark.vcr()
+async def test_bedrock_multiple_documents_in_history(
+    allow_model_requests: None, bedrock_provider: BedrockProvider, document_content: BinaryContent
+):
+    m = BedrockConverseModel(model_name='us.anthropic.claude-3-7-sonnet-20250219-v1:0', provider=bedrock_provider)
+    agent = Agent(model=m)
+
+    result = await agent.run(
+        'What is in the documents?',
+        message_history=[
+            ModelRequest(parts=[UserPromptPart(content=['Here is a PDF document: ', document_content])]),
+            ModelResponse(parts=[TextPart(content='foo bar')]),
+            ModelRequest(parts=[UserPromptPart(content=['Here is another PDF document: ', document_content])]),
+            ModelResponse(parts=[TextPart(content='foo bar 2')]),
+        ],
+    )
+
+    assert result.output == snapshot(
+        'Based on the documents you\'ve shared, both Document 1.pdf and Document 2.pdf contain the text "Dummy PDF file". These appear to be placeholder or sample PDF documents rather than files with substantial content.'
+    )

{pydantic_ai-0.1.1 → pydantic_ai-0.1.3}/tests/models/test_instrumented.py

@@ -692,3 +692,47 @@ def test_messages_to_otel_events_serialization_errors():
             'event.name': 'gen_ai.tool.message',
         },
     ]
+
+
+def test_messages_to_otel_events_instructions():
+    messages = [
+        ModelRequest(instructions='instructions', parts=[UserPromptPart('user_prompt')]),
+        ModelResponse(parts=[TextPart('text1')]),
+    ]
+    assert [
+        InstrumentedModel.event_to_dict(e) for e in InstrumentedModel.messages_to_otel_events(messages)
+    ] == snapshot(
+        [
+            {'content': 'instructions', 'role': 'system', 'event.name': 'gen_ai.system.message'},
+            {'content': 'user_prompt', 'role': 'user', 'gen_ai.message.index': 0, 'event.name': 'gen_ai.user.message'},
+            {
+                'role': 'assistant',
+                'content': 'text1',
+                'gen_ai.message.index': 1,
+                'event.name': 'gen_ai.assistant.message',
+            },
+        ]
+    )
+
+
+def test_messages_to_otel_events_instructions_multiple_messages():
+    messages = [
+        ModelRequest(instructions='instructions', parts=[UserPromptPart('user_prompt')]),
+        ModelResponse(parts=[TextPart('text1')]),
+        ModelRequest(instructions='instructions2', parts=[UserPromptPart('user_prompt2')]),
+    ]
+    assert [
+        InstrumentedModel.event_to_dict(e) for e in InstrumentedModel.messages_to_otel_events(messages)
+    ] == snapshot(
+        [
+            {'content': 'instructions2', 'role': 'system', 'event.name': 'gen_ai.system.message'},
+            {'content': 'user_prompt', 'role': 'user', 'gen_ai.message.index': 0, 'event.name': 'gen_ai.user.message'},
+            {
+                'role': 'assistant',
+                'content': 'text1',
+                'gen_ai.message.index': 1,
+                'event.name': 'gen_ai.assistant.message',
+            },
+            {'content': 'user_prompt2', 'role': 'user', 'gen_ai.message.index': 2, 'event.name': 'gen_ai.user.message'},
+        ]
+    )

{pydantic_ai-0.1.1 → pydantic_ai-0.1.3}/tests/models/test_openai.py

@@ -4,6 +4,7 @@ import json
 from collections.abc import Sequence
 from dataclasses import dataclass, field
 from datetime import datetime, timezone
+from enum import Enum
 from functools import cached_property
 from typing import Annotated, Any, Callable, Literal, Union, cast
 
@@ -175,6 +176,7 @@ async def test_request_simple_success(allow_model_requests: None):
             'model': 'gpt-4o',
             'n': 1,
             'extra_headers': {'User-Agent': IsStr(regex=r'pydantic-ai\/.*')},
+            'extra_body': None,
         },
         {
             'messages': [
@@ -185,6 +187,7 @@ async def test_request_simple_success(allow_model_requests: None):
             'model': 'gpt-4o',
             'n': 1,
             'extra_headers': {'User-Agent': IsStr(regex=r'pydantic-ai\/.*')},
+            'extra_body': None,
         },
     ]
 
@@ -558,6 +561,7 @@ async def test_system_prompt_role(
             'model': 'gpt-4o',
             'n': 1,
             'extra_headers': {'User-Agent': IsStr(regex=r'pydantic-ai\/.*')},
+            'extra_body': None,
         }
     ]
 
@@ -634,6 +638,7 @@ async def test_image_url_input(allow_model_requests: None):
                 ],
                 'n': 1,
                 'extra_headers': {'User-Agent': IsStr(regex=r'pydantic-ai\/.*')},
+                'extra_body': None,
             }
         ]
     )
@@ -730,9 +735,15 @@ class MyDefaultDc:
     x: int = 1
 
 
+class MyEnum(Enum):
+    a = 'a'
+    b = 'b'
+
+
 @dataclass
 class MyRecursiveDc:
     field: MyRecursiveDc | None
+    my_enum: MyEnum = Field(description='my enum')
 
 
 @dataclass
@@ -826,9 +837,13 @@ def tool_with_tuples(x: tuple[int], y: tuple[str] = ('abc',)) -> str:
                             },
                             'type': 'object',
                         },
+                        'MyEnum': {'enum': ['a', 'b'], 'type': 'string'},
                         'MyRecursiveDc': {
-                            'properties': {'field': {'anyOf': [{'$ref': '#/$defs/MyRecursiveDc'}, {'type': 'null'}]}},
-                            'required': ['field'],
+                            'properties': {
+                                'field': {'anyOf': [{'$ref': '#/$defs/MyRecursiveDc'}, {'type': 'null'}]},
+                                'my_enum': {'description': 'my enum', 'anyOf': [{'$ref': '#/$defs/MyEnum'}]},
+                            },
+                            'required': ['field', 'my_enum'],
                             'type': 'object',
                         },
                     },
@@ -857,11 +872,15 @@ def tool_with_tuples(x: tuple[int], y: tuple[str] = ('abc',)) -> str:
                             'additionalProperties': False,
                             'required': ['field'],
                         },
+                        'MyEnum': {'enum': ['a', 'b'], 'type': 'string'},
                         'MyRecursiveDc': {
-                            'properties': {'field': {'anyOf': [{'$ref': '#/$defs/MyRecursiveDc'}, {'type': 'null'}]}},
+                            'properties': {
+                                'field': {'anyOf': [{'$ref': '#/$defs/MyRecursiveDc'}, {'type': 'null'}]},
+                                'my_enum': {'description': 'my enum', 'anyOf': [{'$ref': '#/$defs/MyEnum'}]},
+                            },
                             'type': 'object',
                             'additionalProperties': False,
-                            'required': ['field'],
+                            'required': ['field', 'my_enum'],
                         },
                     },
                     'additionalProperties': False,
@@ -998,7 +1017,7 @@ def tool_with_tuples(x: tuple[int], y: tuple[str] = ('abc',)) -> str:
                         }
                     },
                     'additionalProperties': False,
-                    'properties': {'x': {'oneOf': [{'type': 'integer'}, {'$ref': '#/$defs/MyDefaultDc'}]}},
+                    'properties': {'x': {'anyOf': [{'type': 'integer'}, {'$ref': '#/$defs/MyDefaultDc'}]}},
                     'required': ['x'],
                     'type': 'object',
                 }
@@ -1079,12 +1098,15 @@ def tool_with_tuples(x: tuple[int], y: tuple[str] = ('abc',)) -> str:
                 {
                     'additionalProperties': False,
                     'properties': {
-                        'x': {'maxItems': 1, 'minItems': 1, 'prefixItems': [{'type': 'integer'}], 'type': 'array'},
+                        'x': {
+                            'prefixItems': [{'type': 'integer'}],
+                            'type': 'array',
+                            'description': 'minItems=1, maxItems=1',
+                        },
                         'y': {
-                            'maxItems': 1,
-                            'minItems': 1,
                             'prefixItems': [{'type': 'string'}],
                             'type': 'array',
+                            'description': 'minItems=1, maxItems=1',
                         },
                     },
                     'required': ['x', 'y'],
@@ -1160,28 +1182,46 @@ def test_strict_schema():
                 'MyModel': {
                     'additionalProperties': False,
                     'properties': {
-                        'my_discriminated_union': {'oneOf': [{'$ref': '#/$defs/Apple'}, {'$ref': '#/$defs/Banana'}]},
+                        'my_discriminated_union': {'anyOf': [{'$ref': '#/$defs/Apple'}, {'$ref': '#/$defs/Banana'}]},
                         'my_list': {'items': {'type': 'number'}, 'type': 'array'},
                         'my_patterns': {
                             'additionalProperties': False,
-                            'patternProperties': {'^my-pattern$': {'type': 'string'}},
+                            'description': "patternProperties={'^my-pattern$': {'type': 'string'}}",
                             'type': 'object',
                             'properties': {},
                             'required': [],
                         },
-                        'my_recursive': {'anyOf': [{'$ref': '#/$defs/MyModel'}, {'type': 'null'}]},
+                        'my_recursive': {'anyOf': [{'$ref': '#'}, {'type': 'null'}]},
                         'my_tuple': {
-                            'maxItems': 1,
-                            'minItems': 1,
                             'prefixItems': [{'type': 'integer'}],
                             'type': 'array',
+                            'description': 'minItems=1, maxItems=1',
                         },
                     },
                     'required': ['my_recursive', 'my_patterns', 'my_tuple', 'my_list', 'my_discriminated_union'],
                     'type': 'object',
                 },
             },
-            '$ref': '#/$defs/MyModel',
+            'properties': {
+                'my_recursive': {'anyOf': [{'$ref': '#'}, {'type': 'null'}]},
+                'my_patterns': {
+                    'type': 'object',
+                    'description': "patternProperties={'^my-pattern$': {'type': 'string'}}",
+                    'additionalProperties': False,
+                    'properties': {},
+                    'required': [],
+                },
+                'my_tuple': {
+                    'prefixItems': [{'type': 'integer'}],
+                    'type': 'array',
+                    'description': 'minItems=1, maxItems=1',
+                },
+                'my_list': {'items': {'type': 'number'}, 'type': 'array'},
+                'my_discriminated_union': {'anyOf': [{'$ref': '#/$defs/Apple'}, {'$ref': '#/$defs/Banana'}]},
+            },
+            'required': ['my_recursive', 'my_patterns', 'my_tuple', 'my_list', 'my_discriminated_union'],
+            'type': 'object',
+            'additionalProperties': False,
         }
     )
 

{pydantic_ai-0.1.1 → pydantic_ai-0.1.3}/tests/test_agent.py

@@ -15,6 +15,7 @@ from pydantic_ai import Agent, ModelRetry, RunContext, UnexpectedModelBehavior,
 from pydantic_ai.messages import (
     BinaryContent,
     ModelMessage,
+    ModelMessagesTypeAdapter,
     ModelRequest,
     ModelResponse,
     ModelResponsePart,
@@ -1675,8 +1676,11 @@ def test_custom_output_type_invalid() -> None:
 def test_binary_content_all_messages_json():
     agent = Agent('test')
 
-    result = agent.run_sync(['Hello', BinaryContent(data=b'Hello', media_type='text/plain')])
-    assert json.loads(result.all_messages_json()) == snapshot(
+    content = BinaryContent(data=b'Hello', media_type='text/plain')
+    result = agent.run_sync(['Hello', content])
+
+    serialized = result.all_messages_json()
+    assert json.loads(serialized) == snapshot(
         [
             {
                 'parts': [
@@ -1698,6 +1702,10 @@ def test_binary_content_all_messages_json():
         ]
     )
 
+    # We also need to be able to round trip the serialized messages.
+    messages = ModelMessagesTypeAdapter.validate_json(serialized)
+    assert messages == result.all_messages()
+
 
 def test_instructions_raise_error_when_system_prompt_is_set():
     agent = Agent('test', instructions='An instructions!')
@@ -1805,3 +1813,59 @@ def test_instructions_with_message_history():
             ),
         ]
     )
+
+
+def test_empty_final_response():
+    def llm(messages: list[ModelMessage], info: AgentInfo) -> ModelResponse:
+        if len(messages) == 1:
+            return ModelResponse(parts=[TextPart('foo'), ToolCallPart('my_tool', {'x': 1})])
+        elif len(messages) == 3:
+            return ModelResponse(parts=[TextPart('bar'), ToolCallPart('my_tool', {'x': 2})])
+        else:
+            return ModelResponse(parts=[])
+
+    agent = Agent(FunctionModel(llm))
+
+    @agent.tool_plain
+    def my_tool(x: int) -> int:
+        return x * 2
+
+    result = agent.run_sync('Hello')
+    assert result.output == 'bar'
+
+    assert result.new_messages() == snapshot(
+        [
+            ModelRequest(parts=[UserPromptPart(content='Hello', timestamp=IsNow(tz=timezone.utc))]),
+            ModelResponse(
+                parts=[
+                    TextPart(content='foo'),
+                    ToolCallPart(tool_name='my_tool', args={'x': 1}, tool_call_id=IsStr()),
+                ],
+                model_name='function:llm:',
+                timestamp=IsNow(tz=timezone.utc),
+            ),
+            ModelRequest(
+                parts=[
+                    ToolReturnPart(
+                        tool_name='my_tool', content=2, tool_call_id=IsStr(), timestamp=IsNow(tz=timezone.utc)
+                    )
+                ]
+            ),
+            ModelResponse(
+                parts=[
+                    TextPart(content='bar'),
+                    ToolCallPart(tool_name='my_tool', args={'x': 2}, tool_call_id=IsStr()),
+                ],
+                model_name='function:llm:',
+                timestamp=IsNow(tz=timezone.utc),
+            ),
+            ModelRequest(
+                parts=[
+                    ToolReturnPart(
+                        tool_name='my_tool', content=4, tool_call_id=IsStr(), timestamp=IsNow(tz=timezone.utc)
+                    )
+                ]
+            ),
+            ModelResponse(parts=[], model_name='function:llm:', timestamp=IsNow(tz=timezone.utc)),
+        ]
+    )

{pydantic_ai-0.1.1 → pydantic_ai-0.1.3}/tests/test_mcp.py

@@ -1,5 +1,7 @@
 """Tests for the MCP (Model Context Protocol) server implementation."""
 
+from pathlib import Path
+
 import pytest
 from dirty_equals import IsInstance
 from inline_snapshot import snapshot
@@ -38,6 +40,14 @@ async def test_stdio_server():
         assert result.content == snapshot([TextContent(type='text', text='32.0')])
 
 
+async def test_stdio_server_with_cwd():
+    test_dir = Path(__file__).parent
+    server = MCPServerStdio('python', ['mcp_server.py'], cwd=test_dir)
+    async with server:
+        tools = await server.list_tools()
+        assert len(tools) == 1
+
+
 def test_sse_server():
     sse_server = MCPServerHTTP(url='http://localhost:8000/sse')
     assert sse_server.url == 'http://localhost:8000/sse'

{pydantic_ai-0.1.1 → pydantic_ai-0.1.3}/tests/test_tools.py

@@ -549,6 +549,20 @@ def test_init_plain_tool_invalid():
         Tool(ctx_tool, takes_ctx=False)
 
 
+@pytest.mark.parametrize(
+    'args, expected',
+    [
+        ('', {}),
+        ({'x': 42, 'y': 'value'}, {'x': 42, 'y': 'value'}),
+        ('{"a": 1, "b": "c"}', {'a': 1, 'b': 'c'}),
+    ],
+)
+def test_tool_call_part_args_as_dict(args: Union[str, dict[str, Any]], expected: dict[str, Any]):
+    part = ToolCallPart(tool_name='foo', args=args)
+    result = part.args_as_dict()
+    assert result == expected
+
+
 def test_return_pydantic_model():
     agent = Agent('test')