pydantic-ai 1.6.0__tar.gz → 1.8.0__tar.gz

{pydantic_ai-1.6.0 → pydantic_ai-1.8.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pydantic-ai
-Version: 1.6.0
+Version: 1.8.0
 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
@@ -26,15 +26,15 @@ Classifier: Topic :: Internet
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
 Classifier: Topic :: Software Development :: Libraries :: Python Modules
 Requires-Python: >=3.10
-Requires-Dist: pydantic-ai-slim[ag-ui,anthropic,bedrock,cli,cohere,evals,fastmcp,google,groq,huggingface,logfire,mcp,mistral,openai,retries,temporal,vertexai]==1.6.0
+Requires-Dist: pydantic-ai-slim[ag-ui,anthropic,bedrock,cli,cohere,evals,fastmcp,google,groq,huggingface,logfire,mcp,mistral,openai,retries,temporal,vertexai]==1.8.0
 Provides-Extra: a2a
 Requires-Dist: fasta2a>=0.4.1; extra == 'a2a'
 Provides-Extra: dbos
-Requires-Dist: pydantic-ai-slim[dbos]==1.6.0; extra == 'dbos'
+Requires-Dist: pydantic-ai-slim[dbos]==1.8.0; extra == 'dbos'
 Provides-Extra: examples
-Requires-Dist: pydantic-ai-examples==1.6.0; extra == 'examples'
+Requires-Dist: pydantic-ai-examples==1.8.0; extra == 'examples'
 Provides-Extra: prefect
-Requires-Dist: pydantic-ai-slim[prefect]==1.6.0; extra == 'prefect'
+Requires-Dist: pydantic-ai-slim[prefect]==1.8.0; extra == 'prefect'
 Description-Content-Type: text/markdown
 
 <div align="center">
@@ -78,7 +78,7 @@ We built Pydantic AI with one simple aim: to bring that FastAPI feeling to GenAI
 [Pydantic Validation](https://docs.pydantic.dev/latest/) is the validation layer of the OpenAI SDK, the Google ADK, the Anthropic SDK, LangChain, LlamaIndex, AutoGPT, Transformers, CrewAI, Instructor and many more. _Why use the derivative when you can go straight to the source?_ :smiley:
 
 2. **Model-agnostic**:
-Supports virtually every [model](https://ai.pydantic.dev/models/overview) and provider: OpenAI, Anthropic, Gemini, DeepSeek, Grok, Cohere, Mistral, and Perplexity; Azure AI Foundry, Amazon Bedrock, Google Vertex AI, Ollama, LiteLLM, Groq, OpenRouter, Together AI, Fireworks AI, Cerebras, Hugging Face, GitHub, Heroku, Vercel, Nebius, OVHcloud. If your favorite model or provider is not listed, you can easily implement a [custom model](https://ai.pydantic.dev/models/overview#custom-models).
+Supports virtually every [model](https://ai.pydantic.dev/models/overview) and provider: OpenAI, Anthropic, Gemini, DeepSeek, Grok, Cohere, Mistral, and Perplexity; Azure AI Foundry, Amazon Bedrock, Google Vertex AI, Ollama, LiteLLM, Groq, OpenRouter, Together AI, Fireworks AI, Cerebras, Hugging Face, GitHub, Heroku, Vercel, Nebius, OVHcloud, and Outlines. If your favorite model or provider is not listed, you can easily implement a [custom model](https://ai.pydantic.dev/models/overview#custom-models).
 
 3. **Seamless Observability**:
 Tightly [integrates](https://ai.pydantic.dev/logfire) with [Pydantic Logfire](https://pydantic.dev/logfire), our general-purpose OpenTelemetry observability platform, for real-time debugging, evals-based performance monitoring, and behavior, tracing, and cost tracking. If you already have an observability platform that supports OTel, you can [use that too](https://ai.pydantic.dev/logfire#alternative-observability-backends).

{pydantic_ai-1.6.0 → pydantic_ai-1.8.0}/README.md

@@ -39,7 +39,7 @@ We built Pydantic AI with one simple aim: to bring that FastAPI feeling to GenAI
 [Pydantic Validation](https://docs.pydantic.dev/latest/) is the validation layer of the OpenAI SDK, the Google ADK, the Anthropic SDK, LangChain, LlamaIndex, AutoGPT, Transformers, CrewAI, Instructor and many more. _Why use the derivative when you can go straight to the source?_ :smiley:
 
 2. **Model-agnostic**:
-Supports virtually every [model](https://ai.pydantic.dev/models/overview) and provider: OpenAI, Anthropic, Gemini, DeepSeek, Grok, Cohere, Mistral, and Perplexity; Azure AI Foundry, Amazon Bedrock, Google Vertex AI, Ollama, LiteLLM, Groq, OpenRouter, Together AI, Fireworks AI, Cerebras, Hugging Face, GitHub, Heroku, Vercel, Nebius, OVHcloud. If your favorite model or provider is not listed, you can easily implement a [custom model](https://ai.pydantic.dev/models/overview#custom-models).
+Supports virtually every [model](https://ai.pydantic.dev/models/overview) and provider: OpenAI, Anthropic, Gemini, DeepSeek, Grok, Cohere, Mistral, and Perplexity; Azure AI Foundry, Amazon Bedrock, Google Vertex AI, Ollama, LiteLLM, Groq, OpenRouter, Together AI, Fireworks AI, Cerebras, Hugging Face, GitHub, Heroku, Vercel, Nebius, OVHcloud, and Outlines. If your favorite model or provider is not listed, you can easily implement a [custom model](https://ai.pydantic.dev/models/overview#custom-models).
 
 3. **Seamless Observability**:
 Tightly [integrates](https://ai.pydantic.dev/logfire) with [Pydantic Logfire](https://pydantic.dev/logfire), our general-purpose OpenTelemetry observability platform, for real-time debugging, evals-based performance monitoring, and behavior, tracing, and cost tracking. If you already have an observability platform that supports OTel, you can [use that too](https://ai.pydantic.dev/logfire#alternative-observability-backends).

{pydantic_ai-1.6.0 → pydantic_ai-1.8.0}/pyproject.toml

@@ -105,6 +105,11 @@ dev = [
     "pip>=25.2",
     "genai-prices>=0.0.28",
     "mcp-run-python>=0.0.20",
+    # Needed to test Outlines (not included in the default installation)
+    "pydantic-ai-slim[outlines-transformers]",
+    "pydantic-ai-slim[outlines-llamacpp]",
+    "pydantic-ai-slim[outlines-mlxlm]",
+    "pydantic-ai-slim[outlines-sglang]",
 ]
 lint = ["mypy>=1.11.2", "pyright>=1.1.390", "ruff>=0.6.9"]
 docs = [

{pydantic_ai-1.6.0 → pydantic_ai-1.8.0}/tests/cassettes/test_settings/test_stop_settings[anthropic].yaml

@@ -21,7 +21,7 @@ interactions:
         - text: What is the capital of France? Give me an answer that contains the word "Paris", but is not the first word.
           type: text
         role: user
-      model: claude-3-5-sonnet-latest
+      model: claude-sonnet-4-5
       stop_sequences:
       - Paris
       stream: false
@@ -43,7 +43,7 @@ interactions:
       - text: 'The beautiful city of '
         type: text
       id: msg_01376yZQxHcw9pER2Ab2SvQb
-      model: claude-3-5-sonnet-20241022
+      model: claude-sonnet-4-5-20250929
       role: assistant
       stop_reason: stop_sequence
       stop_sequence: Paris

{pydantic_ai-1.6.0 → pydantic_ai-1.8.0}/tests/conftest.py

@@ -498,7 +498,7 @@ def model(
             from pydantic_ai.models.anthropic import AnthropicModel
             from pydantic_ai.providers.anthropic import AnthropicProvider
 
-            return AnthropicModel('claude-3-5-sonnet-latest', provider=AnthropicProvider(api_key=anthropic_api_key))
+            return AnthropicModel('claude-sonnet-4-5', provider=AnthropicProvider(api_key=anthropic_api_key))
         elif request.param == 'mistral':
             from pydantic_ai.models.mistral import MistralModel
             from pydantic_ai.providers.mistral import MistralProvider
@@ -536,6 +536,18 @@ def model(
                 'Qwen/Qwen2.5-72B-Instruct',
                 provider=HuggingFaceProvider(provider_name='nebius', api_key=huggingface_api_key),
             )
+        elif request.param == 'outlines':
+            from outlines.models.transformers import from_transformers
+            from transformers import AutoModelForCausalLM, AutoTokenizer
+
+            from pydantic_ai.models.outlines import OutlinesModel
+
+            return OutlinesModel(
+                from_transformers(
+                    AutoModelForCausalLM.from_pretrained('erwanf/gpt2-mini'),
+                    AutoTokenizer.from_pretrained('erwanf/gpt2-mini'),
+                )
+            )
         else:
             raise ValueError(f'Unknown model: {request.param}')
     except ImportError:

{pydantic_ai-1.6.0 → pydantic_ai-1.8.0}/tests/evals/test_dataset.py

@@ -1530,7 +1530,7 @@ async def test_evaluate_async_logfire(
             return TaskOutput(answer='Paris')
         return TaskOutput(answer='Unknown')  # pragma: no cover
 
-    await example_dataset.evaluate(mock_async_task)
+    await example_dataset.evaluate(mock_async_task, metadata={'key': 'value'})
 
     spans = capfire.exporter.exported_spans_as_dict(parse_json_attributes=True)
     spans.sort(key=lambda s: s['start_time'])
@@ -1556,6 +1556,7 @@ async def test_evaluate_async_logfire(
                             'gen_ai.operation.name': {},
                             'n_cases': {},
                             'name': {},
+                            'metadata': {'type': 'object'},
                             'logfire.experiment.metadata': {
                                 'type': 'object',
                                 'properties': {
@@ -1571,11 +1572,13 @@ async def test_evaluate_async_logfire(
                         'type': 'object',
                     },
                     'logfire.msg': 'evaluate mock_async_task',
+                    'metadata': {'key': 'value'},
                     'logfire.msg_template': 'evaluate {name}',
                     'logfire.span_type': 'span',
                     'n_cases': 2,
                     'logfire.experiment.metadata': {
                         'n_cases': 2,
+                        'metadata': {'key': 'value'},
                         'averages': {
                             'name': 'Averages',
                             'scores': {'confidence': 1.0},
@@ -1750,3 +1753,23 @@ async def test_evaluate_async_logfire(
             ),
         ]
     )
+
+
+async def test_evaluate_with_experiment_metadata(example_dataset: Dataset[TaskInput, TaskOutput, TaskMetadata]):
+    """Test that experiment metadata passed to evaluate() appears in the report."""
+
+    async def task(inputs: TaskInput) -> TaskOutput:
+        return TaskOutput(answer=inputs.query.upper())
+
+    # Pass experiment metadata to evaluate()
+    experiment_metadata = {
+        'model': 'gpt-4o',
+        'prompt_version': 'v2.1',
+        'temperature': 0.7,
+        'max_tokens': 1000,
+    }
+
+    report = await example_dataset.evaluate(task, metadata=experiment_metadata)
+
+    # Verify that the report contains the experiment metadata
+    assert report.experiment_metadata == experiment_metadata

{pydantic_ai-1.6.0 → pydantic_ai-1.8.0}/tests/evals/test_reporting.py

@@ -950,3 +950,415 @@ async def test_evaluation_renderer_no_evaluator_failures_column():
 │ test_case │ {'query': 'What is 2+2?'} │ {'answer': '4'} │ accuracy: 0.950 │    0.100 │
 └───────────┴───────────────────────────┴─────────────────┴─────────────────┴──────────┘
 """)
+
+
+async def test_evaluation_renderer_with_experiment_metadata(sample_report_case: ReportCase):
+    """Test EvaluationRenderer with experiment metadata."""
+    report = EvaluationReport(
+        cases=[sample_report_case],
+        name='test_report',
+        experiment_metadata={'model': 'gpt-4o', 'temperature': 0.7, 'prompt_version': 'v2'},
+    )
+
+    output = report.render(
+        include_input=True,
+        include_metadata=False,
+        include_expected_output=False,
+        include_output=False,
+        include_durations=True,
+        include_total_duration=False,
+        include_removed_cases=False,
+        include_averages=True,
+        include_errors=False,
+        include_error_stacktrace=False,
+        include_evaluator_failures=True,
+        input_config={},
+        metadata_config={},
+        output_config={},
+        score_configs={},
+        label_configs={},
+        metric_configs={},
+        duration_config={},
+        include_reasons=False,
+    )
+
+    assert output == snapshot("""\
+╭─ Evaluation Summary: test_report ─╮
+│ model: gpt-4o                     │
+│ temperature: 0.7                  │
+│ prompt_version: v2                │
+╰───────────────────────────────────╯
+┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━┓
+┃ Case ID   ┃ Inputs                    ┃ Scores       ┃ Labels                 ┃ Metrics         ┃ Assertions ┃ Duration ┃
+┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━┩
+│ test_case │ {'query': 'What is 2+2?'} │ score1: 2.50 │ label1: hello          │ accuracy: 0.950 │ ✔          │  100.0ms │
+├───────────┼───────────────────────────┼──────────────┼────────────────────────┼─────────────────┼────────────┼──────────┤
+│ Averages  │                           │ score1: 2.50 │ label1: {'hello': 1.0} │ accuracy: 0.950 │ 100.0% ✔   │  100.0ms │
+└───────────┴───────────────────────────┴──────────────┴────────────────────────┴─────────────────┴────────────┴──────────┘
+""")
+
+
+async def test_evaluation_renderer_with_long_experiment_metadata(sample_report_case: ReportCase):
+    """Test EvaluationRenderer with very long experiment metadata."""
+    report = EvaluationReport(
+        cases=[sample_report_case],
+        name='test_report',
+        experiment_metadata={
+            'model': 'gpt-4o-2024-08-06',
+            'temperature': 0.7,
+            'prompt_version': 'v2.1.5',
+            'system_prompt': 'You are a helpful assistant',
+            'max_tokens': 1000,
+            'top_p': 0.9,
+            'frequency_penalty': 0.1,
+            'presence_penalty': 0.1,
+        },
+    )
+
+    output = report.render(
+        include_input=False,
+        include_metadata=False,
+        include_expected_output=False,
+        include_output=False,
+        include_durations=True,
+        include_total_duration=False,
+        include_removed_cases=False,
+        include_averages=False,
+        include_errors=False,
+        include_error_stacktrace=False,
+        include_evaluator_failures=True,
+        input_config={},
+        metadata_config={},
+        output_config={},
+        score_configs={},
+        label_configs={},
+        metric_configs={},
+        duration_config={},
+        include_reasons=False,
+    )
+
+    assert output == snapshot("""\
+╭─ Evaluation Summary: test_report ──────────╮
+│ model: gpt-4o-2024-08-06                   │
+│ temperature: 0.7                           │
+│ prompt_version: v2.1.5                     │
+│ system_prompt: You are a helpful assistant │
+│ max_tokens: 1000                           │
+│ top_p: 0.9                                 │
+│ frequency_penalty: 0.1                     │
+│ presence_penalty: 0.1                      │
+╰────────────────────────────────────────────╯
+┏━━━━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━┓
+┃ Case ID   ┃ Scores       ┃ Labels        ┃ Metrics         ┃ Assertions ┃ Duration ┃
+┡━━━━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━┩
+│ test_case │ score1: 2.50 │ label1: hello │ accuracy: 0.950 │ ✔          │  100.0ms │
+└───────────┴──────────────┴───────────────┴─────────────────┴────────────┴──────────┘
+""")
+
+
+async def test_evaluation_renderer_diff_with_experiment_metadata(sample_report_case: ReportCase):
+    """Test EvaluationRenderer diff table with experiment metadata."""
+    baseline_report = EvaluationReport(
+        cases=[sample_report_case],
+        name='baseline_report',
+        experiment_metadata={'model': 'gpt-4', 'temperature': 0.5},
+    )
+
+    new_report = EvaluationReport(
+        cases=[sample_report_case],
+        name='new_report',
+        experiment_metadata={'model': 'gpt-4o', 'temperature': 0.7},
+    )
+
+    output = new_report.render(
+        baseline=baseline_report,
+        include_input=False,
+        include_metadata=False,
+        include_expected_output=False,
+        include_output=False,
+        include_durations=True,
+        include_total_duration=False,
+        include_removed_cases=False,
+        include_averages=True,
+        include_errors=False,
+        include_error_stacktrace=False,
+        include_evaluator_failures=True,
+        input_config={},
+        metadata_config={},
+        output_config={},
+        score_configs={},
+        label_configs={},
+        metric_configs={},
+        duration_config={},
+        include_reasons=False,
+    )
+
+    assert output == snapshot("""\
+╭─ Evaluation Diff: baseline_report → new_report ─╮
+│ model: gpt-4 → gpt-4o                           │
+│ temperature: 0.5 → 0.7                          │
+╰─────────────────────────────────────────────────╯
+┏━━━━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━┓
+┃ Case ID   ┃ Scores       ┃ Labels                 ┃ Metrics         ┃ Assertions ┃ Duration ┃
+┡━━━━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━┩
+│ test_case │ score1: 2.50 │ label1: hello          │ accuracy: 0.950 │ ✔          │  100.0ms │
+├───────────┼──────────────┼────────────────────────┼─────────────────┼────────────┼──────────┤
+│ Averages  │ score1: 2.50 │ label1: {'hello': 1.0} │ accuracy: 0.950 │ 100.0% ✔   │  100.0ms │
+└───────────┴──────────────┴────────────────────────┴─────────────────┴────────────┴──────────┘
+""")
+
+
+async def test_evaluation_renderer_diff_with_only_new_metadata(sample_report_case: ReportCase):
+    """Test EvaluationRenderer diff table where only new report has metadata."""
+    baseline_report = EvaluationReport(
+        cases=[sample_report_case],
+        name='baseline_report',
+        experiment_metadata=None,  # No metadata
+    )
+
+    new_report = EvaluationReport(
+        cases=[sample_report_case],
+        name='new_report',
+        experiment_metadata={'model': 'gpt-4o', 'temperature': 0.7},
+    )
+
+    output = new_report.render(
+        baseline=baseline_report,
+        include_input=False,
+        include_metadata=False,
+        include_expected_output=False,
+        include_output=False,
+        include_durations=True,
+        include_total_duration=False,
+        include_removed_cases=False,
+        include_averages=False,
+        include_errors=False,
+        include_error_stacktrace=False,
+        include_evaluator_failures=True,
+        input_config={},
+        metadata_config={},
+        output_config={},
+        score_configs={},
+        label_configs={},
+        metric_configs={},
+        duration_config={},
+        include_reasons=False,
+    )
+
+    assert output == snapshot("""\
+╭─ Evaluation Diff: baseline_report → new_report ─╮
+│ + model: gpt-4o                                 │
+│ + temperature: 0.7                              │
+╰─────────────────────────────────────────────────╯
+┏━━━━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━┓
+┃ Case ID   ┃ Scores       ┃ Labels        ┃ Metrics         ┃ Assertions ┃ Duration ┃
+┡━━━━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━┩
+│ test_case │ score1: 2.50 │ label1: hello │ accuracy: 0.950 │ ✔          │  100.0ms │
+└───────────┴──────────────┴───────────────┴─────────────────┴────────────┴──────────┘
+""")
+
+
+async def test_evaluation_renderer_diff_with_only_baseline_metadata(sample_report_case: ReportCase):
+    """Test EvaluationRenderer diff table where only baseline report has metadata."""
+    baseline_report = EvaluationReport(
+        cases=[sample_report_case],
+        name='baseline_report',
+        experiment_metadata={'model': 'gpt-4', 'temperature': 0.5},
+    )
+
+    new_report = EvaluationReport(
+        cases=[sample_report_case],
+        name='new_report',
+        experiment_metadata=None,  # No metadata
+    )
+
+    output = new_report.render(
+        baseline=baseline_report,
+        include_input=False,
+        include_metadata=False,
+        include_expected_output=False,
+        include_output=False,
+        include_durations=True,
+        include_total_duration=False,
+        include_removed_cases=False,
+        include_averages=False,
+        include_errors=False,
+        include_error_stacktrace=False,
+        include_evaluator_failures=True,
+        input_config={},
+        metadata_config={},
+        output_config={},
+        score_configs={},
+        label_configs={},
+        metric_configs={},
+        duration_config={},
+        include_reasons=False,
+    )
+
+    assert output == snapshot("""\
+╭─ Evaluation Diff: baseline_report → new_report ─╮
+│ - model: gpt-4                                  │
+│ - temperature: 0.5                              │
+╰─────────────────────────────────────────────────╯
+┏━━━━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━┓
+┃ Case ID   ┃ Scores       ┃ Labels        ┃ Metrics         ┃ Assertions ┃ Duration ┃
+┡━━━━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━┩
+│ test_case │ score1: 2.50 │ label1: hello │ accuracy: 0.950 │ ✔          │  100.0ms │
+└───────────┴──────────────┴───────────────┴─────────────────┴────────────┴──────────┘
+""")
+
+
+async def test_evaluation_renderer_diff_with_same_metadata(sample_report_case: ReportCase):
+    """Test EvaluationRenderer diff table where both reports have the same metadata."""
+    metadata = {'model': 'gpt-4o', 'temperature': 0.7}
+
+    baseline_report = EvaluationReport(
+        cases=[sample_report_case],
+        name='baseline_report',
+        experiment_metadata=metadata,
+    )
+
+    new_report = EvaluationReport(
+        cases=[sample_report_case],
+        name='new_report',
+        experiment_metadata=metadata,
+    )
+
+    output = new_report.render(
+        include_input=False,
+        include_metadata=False,
+        include_expected_output=False,
+        include_output=False,
+        include_durations=True,
+        include_total_duration=False,
+        include_removed_cases=False,
+        include_averages=False,
+        include_error_stacktrace=False,
+        include_evaluator_failures=True,
+        input_config={},
+        metadata_config={},
+        output_config={},
+        score_configs={},
+        label_configs={},
+        metric_configs={},
+        duration_config={},
+        include_reasons=False,
+        baseline=baseline_report,
+        include_errors=False,  # Prevent failures table from being added
+    )
+    assert output == snapshot("""\
+╭─ Evaluation Diff: baseline_report → new_report ─╮
+│ model: gpt-4o                                   │
+│ temperature: 0.7                                │
+╰─────────────────────────────────────────────────╯
+┏━━━━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━┓
+┃ Case ID   ┃ Scores       ┃ Labels        ┃ Metrics         ┃ Assertions ┃ Duration ┃
+┡━━━━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━┩
+│ test_case │ score1: 2.50 │ label1: hello │ accuracy: 0.950 │ ✔          │  100.0ms │
+└───────────┴──────────────┴───────────────┴─────────────────┴────────────┴──────────┘
+""")
+
+
+async def test_evaluation_renderer_diff_with_changed_metadata(sample_report_case: ReportCase):
+    """Test EvaluationRenderer diff table where both reports have the same metadata."""
+
+    baseline_report = EvaluationReport(
+        cases=[sample_report_case],
+        name='baseline_report',
+        experiment_metadata={
+            'updated-key': 'original value',
+            'preserved-key': 'preserved value',
+            'old-key': 'old value',
+        },
+    )
+
+    new_report = EvaluationReport(
+        cases=[sample_report_case],
+        name='new_report',
+        experiment_metadata={
+            'updated-key': 'updated value',
+            'preserved-key': 'preserved value',
+            'new-key': 'new value',
+        },
+    )
+
+    output = new_report.render(
+        include_input=False,
+        include_metadata=False,
+        include_expected_output=False,
+        include_output=False,
+        include_durations=True,
+        include_total_duration=False,
+        include_removed_cases=False,
+        include_averages=False,
+        include_error_stacktrace=False,
+        include_evaluator_failures=True,
+        input_config={},
+        metadata_config={},
+        output_config={},
+        score_configs={},
+        label_configs={},
+        metric_configs={},
+        duration_config={},
+        include_reasons=False,
+        baseline=baseline_report,
+        include_errors=False,  # Prevent failures table from being added
+    )
+    assert output == snapshot("""\
+╭─ Evaluation Diff: baseline_report → new_report ─╮
+│ + new-key: new value                            │
+│ - old-key: old value                            │
+│ preserved-key: preserved value                  │
+│ updated-key: original value → updated value     │
+╰─────────────────────────────────────────────────╯
+┏━━━━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━┓
+┃ Case ID   ┃ Scores       ┃ Labels        ┃ Metrics         ┃ Assertions ┃ Duration ┃
+┡━━━━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━┩
+│ test_case │ score1: 2.50 │ label1: hello │ accuracy: 0.950 │ ✔          │  100.0ms │
+└───────────┴──────────────┴───────────────┴─────────────────┴────────────┴──────────┘
+""")
+
+
+async def test_evaluation_renderer_diff_with_no_metadata(sample_report_case: ReportCase):
+    """Test EvaluationRenderer diff table where both reports have the same metadata."""
+
+    baseline_report = EvaluationReport(
+        cases=[sample_report_case],
+        name='baseline_report',
+    )
+
+    new_report = EvaluationReport(
+        cases=[sample_report_case],
+        name='new_report',
+    )
+
+    output = new_report.render(
+        include_input=False,
+        include_metadata=False,
+        include_expected_output=False,
+        include_output=False,
+        include_durations=True,
+        include_total_duration=False,
+        include_removed_cases=False,
+        include_averages=False,
+        include_error_stacktrace=False,
+        include_evaluator_failures=True,
+        input_config={},
+        metadata_config={},
+        output_config={},
+        score_configs={},
+        label_configs={},
+        metric_configs={},
+        duration_config={},
+        include_reasons=False,
+        baseline=baseline_report,
+        include_errors=False,  # Prevent failures table from being added
+    )
+    assert output == snapshot("""\
+                    Evaluation Diff: baseline_report → new_report                     \n\
+┏━━━━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━┓
+┃ Case ID   ┃ Scores       ┃ Labels        ┃ Metrics         ┃ Assertions ┃ Duration ┃
+┡━━━━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━┩
+│ test_case │ score1: 2.50 │ label1: hello │ accuracy: 0.950 │ ✔          │  100.0ms │
+└───────────┴──────────────┴───────────────┴─────────────────┴────────────┴──────────┘
+""")

{pydantic_ai-1.6.0 → pydantic_ai-1.8.0}/tests/models/cassettes/test_anthropic/test_anthropic_model_empty_message_on_history.yaml

@@ -25,7 +25,7 @@ interactions:
         - text: I need a potato!
           type: text
         role: user
-      model: claude-3-5-sonnet-latest
+      model: claude-sonnet-4-5
       stream: false
       system: |+
         You are a helpful assistant.
@@ -56,7 +56,7 @@ interactions:
           What specific information about potatoes would be most helpful to you?
         type: text
       id: msg_01PAZFa5ciacA9ptgEDMbkZM
-      model: claude-3-5-sonnet-20241022
+      model: claude-sonnet-4-5-20250929
       role: assistant
       stop_reason: end_turn
       stop_sequence: null

{pydantic_ai-1.6.0 → pydantic_ai-1.8.0}/tests/models/cassettes/test_anthropic/test_anthropic_model_thinking_part.yaml

@@ -21,7 +21,7 @@ interactions:
         - text: How do I cross the street?
           type: text
         role: user
-      model: claude-3-7-sonnet-latest
+      model: claude-sonnet-4-5
       stream: false
       thinking:
         budget_tokens: 1024
@@ -82,7 +82,7 @@ interactions:
           Would you like me to explain any of these steps in more detail?
         type: text
       id: msg_01BnZvs3naGorn93wjjCDwbd
-      model: claude-3-7-sonnet-20250219
+      model: claude-sonnet-4-5-20250929
       role: assistant
       stop_reason: end_turn
       stop_sequence: null
@@ -167,7 +167,7 @@ interactions:
         - text: Considering the way to cross the street, analogously, how do I cross the river?
           type: text
         role: user
-      model: claude-3-7-sonnet-latest
+      model: claude-sonnet-4-5
       stream: false
       thinking:
         budget_tokens: 1024
@@ -235,7 +235,7 @@ interactions:
           Is there a specific river crossing scenario you're curious about?
         type: text
       id: msg_019Z9a1qnqUCxd7Fj6PuuetE
-      model: claude-3-7-sonnet-20250219
+      model: claude-sonnet-4-5-20250929
       role: assistant
       stop_reason: end_turn
       stop_sequence: null