pydantic-ai 1.2.1__tar.gz → 1.3.0__tar.gz

{pydantic_ai-1.2.1 → pydantic_ai-1.3.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pydantic-ai
-Version: 1.2.1
+Version: 1.3.0
 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,google,groq,huggingface,logfire,mcp,mistral,openai,retries,temporal,vertexai]==1.2.1
+Requires-Dist: pydantic-ai-slim[ag-ui,anthropic,bedrock,cli,cohere,evals,google,groq,huggingface,logfire,mcp,mistral,openai,retries,temporal,vertexai]==1.3.0
 Provides-Extra: a2a
 Requires-Dist: fasta2a>=0.4.1; extra == 'a2a'
 Provides-Extra: dbos
-Requires-Dist: pydantic-ai-slim[dbos]==1.2.1; extra == 'dbos'
+Requires-Dist: pydantic-ai-slim[dbos]==1.3.0; extra == 'dbos'
 Provides-Extra: examples
-Requires-Dist: pydantic-ai-examples==1.2.1; extra == 'examples'
+Requires-Dist: pydantic-ai-examples==1.3.0; extra == 'examples'
 Provides-Extra: prefect
-Requires-Dist: pydantic-ai-slim[prefect]==1.2.1; extra == 'prefect'
+Requires-Dist: pydantic-ai-slim[prefect]==1.3.0; 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. 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. 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-1.2.1 → pydantic_ai-1.3.0}/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. 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. 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-1.3.0/tests/cassettes/test_temporal/test_web_search_agent_run_in_workflow.yaml

@@ -0,0 +1,151 @@
+interactions:
+- request:
+    headers:
+      accept:
+      - application/json
+      accept-encoding:
+      - gzip, deflate
+      connection:
+      - keep-alive
+      content-length:
+      - '336'
+      content-type:
+      - application/json
+      host:
+      - api.openai.com
+    method: POST
+    parsed_body:
+      include:
+      - reasoning.encrypted_content
+      input:
+      - content: In one sentence, what is the top news story in my country today?
+        role: user
+      model: gpt-5
+      stream: false
+      tool_choice: auto
+      tools:
+      - search_context_size: medium
+        type: web_search
+        user_location:
+          city: Mexico City
+          country: MX
+          type: approximate
+    uri: https://api.openai.com/v1/responses
+  response:
+    headers:
+      alt-svc:
+      - h3=":443"; ma=86400
+      connection:
+      - keep-alive
+      content-length:
+      - '12089'
+      content-type:
+      - application/json
+      openai-organization:
+      - pydantic-28gund
+      openai-processing-ms:
+      - '45591'
+      openai-project:
+      - proj_dKobscVY9YJxeEaDJen54e3d
+      openai-version:
+      - '2020-10-01'
+      strict-transport-security:
+      - max-age=31536000; includeSubDomains; preload
+      transfer-encoding:
+      - chunked
+    parsed_body:
+      background: false
+      billing:
+        payer: developer
+      created_at: 1760537022
+      error: null
+      id: resp_0cc772278fa4f4140068efa9be8878819ca691ebcbe9f1f6be
+      incomplete_details: null
+      instructions: null
+      max_output_tokens: null
+      max_tool_calls: null
+      metadata: {}
+      model: gpt-5-2025-08-07
+      object: response
+      output:
+      - encrypted_content: gAAAAABo76nsGZ3D61NVA0XbnShNHmz5T4RonYmRGx_pRGkEbVDw7m4Ld84NA0js4svXie5FHgTxnkjohBfv597Y9YkQPDKDKkT80tQ41fo4SYkJ4r3BUtgD-DNrpssBBlK_uT-FfWeJlwTTvz6hfl1jP8NK54ZThlLzeSyhoDkQoiJf7d7rmjro49oy8owOCGlod1oiBBcZoF4zXL04lZXeOmrN40e6zIrEXGIb-iC2NL7O2R3GbipQcbq6zy2RpIbmjgmXRTWjtaMmWcJwry4oqYsud2UHojpkhXeN-0Yh4gPfzWbgvYDO47Lc48t5UkNDuBLXezNyRX_Pmlocn-h9RKNEdkq1iZau4UyY_1Fo8LKm0hQuE4x4ZKzRed2KJM4YjwSomtcGshDGKmZw94kabHY9tDI6y5-mWFhHqvjZt50xJukx2iOC6KKKk0iKiqcCeNb4r2K0G_xhLIS-sjbG9rv_762G3AiZ041xEwbStrqovRuTu52YV6b70gnuzvEG0QF9hToUpbZZQx1prAIE1WILXAZrDxvHB2czq2PF3WP0QGWdbJpO3ZSm9osjX1qqNgoiOArBUYtOQUbYM7JtLz92Es53GVeIXlHSlE_VT92ISO6n8UyAGLYuWLd-nTELBbsjNWmOo1LFpUsi7h0rVR0pbs2jTUDRLCn74ucVxYX45dTx0P-OKfNrKDQNaRCyvi-oCFdrJUc-ecm-EaVoqz0b63HHzhKNXFsmPeFZl5gtGnXARqXrbcBgel-wviXlMLt8KaWWMrEU_yq8KjACNeQy8POZsimzWT95oMAjD_J35JgeXajkYIFHuh1ZRb0awKdXPCUp6kR-nL_cBZco6nZARImw-9VR7527KT30aGvPhSYCrtqP9QXmI94IB9mYvGsCCj1ooQ3JW4nzvThJ3KN9xNZaMkvE_pORulsxPLfKfELoy7i74HWBZ7X5_7zLB-Rv-Zj58YJiFtpVYGSzKd2UfbI8GvYaakGFfN2LqzG4k2UA2rIh97nbMfbdVtUL0PuINsHc2Qr1aHbeaiHavya4QCGt1w0cds37gy06dqdL5Hny2NzhSekkUK2vDGV-evUpIkDeK3hY93NQYgwppnmdvMDYOMUl_VnLH5m-3FIW0bTJkxC2zYR4A1lkD_xhawH2KD5lrZS4-qttoDdfWceV6PIW3J7EnP0j9vq5Twkx9phCXWOL0-_7fEKwwmC2VwVvf43rbqbF5ZQpLl_nNiicJX4nmV_umniCS_hUjmHTHHyagnJjUzD7ZHFjF7CenXS4uag4TlNEozrxEv4vcZtWd_A-zSpqYLdGty2Zp16nkeq2DhF137AZTohRpDttSis-oh18GQzgvyDb3kUyb23CHcH9DZwScI4QDiDjoE94mzOe4_MjSyrY4d4stLYr5DYlUc_zE3jWp10QiNz8Yz3K4sm8aM_cPteHRuwGBPmjYWKoNwhsLEiHP3aNE-MYpkcV4oAJk8MR3NU3asRRZUD5CNj66gkRlpcsd6dTUPSJISIZM415mE7uG8HF0inxGn_VXSIFxV7PfjdIs6WE58dwEEe7QXTpxWjkuCCykg43znRi0G3Tscuq8oU7VUe5dTAfGDA-nwkroCFWyfpHT_VX92l4BFM6Pm9GF1XJ3_cpbp_TMkrLNwkklGeZSbghTU0RqKdGfffyrdZNqRW16BESjHc7Z8SsC8QhLjGvavyFHJOBB3vgwRAFDskmFlhTKMAsJqaq_tqJ550UrrRTmjJ1sH3gOozKgryOU9i7w2rlnbn24YkA9yCvIVSaB7bW8_Z0TbhoWY_0D25OQqc8pgOGECEHDzGOX_k5wSevf6WEdxoL9wY0PpNxGz_TwA0DMOd4vModAeCGDujQSfPS7PDUm9qtZMZtwyjZ4PKQ_K09-sKg7X4PUQ9DAdygNmHQCjK0iqY-fhcJwkApX1ichrxy6M9kA1EM8rO5fD4pTaRLoqNVutc1sUBq93DXMEjE73UTCfrygIOK2brw4P1LBNQe_CP3J65obJQ8tMTXyemrnLPYIzw=
+        id: rs_0cc772278fa4f4140068efa9bf0dcc819ca630d7697c77a632
+        summary: []
+        type: reasoning
+      - action:
+          query: Mexico top news October 15, 2025
+          type: search
+        id: ws_0cc772278fa4f4140068efa9c5aa14819c871c50351e98af40
+        status: completed
+        type: web_search_call
+      - encrypted_content: gAAAAABo76nsliyRJLQ8_6svrujSylQVBRqPDxfWb2I2yr61HB3Rvh8mkk4ZPdzVyktR-gyzPkY-X5VtSV3rjjMCRvU6nGi8j9MmS-pULY8splSAdXCTcxm5LoH6laEAtugW0p6HlxkaehV4RLM2RNqiFbx3KMciX3JTkg9AJcchLtGiSvaV4KrI6gO7BoMY_L9lPsTAOJMCWmD6eK3S13Rc61XPEsUhLq7W4h_mz-A_y6K5Hehqu4sEzLlWgAI3bVogiULK0UYQqgmRkL5VhZCw3RP7lqV_3eDD-7hJEgH7tnW3aYkSt-6f0PmuDoDdF5UOdhgjYLyr2KGyXokoRVF310d4b7OEJsFycF4-mqtD8-r6RxjmSJVp95Oye-3X2HhXKwanOhJbtymEa77qy2hIC_8p222na1_wBo9awrBaq8T1cyGauiMEEbVcqVkhwE3mQzGYz2EzbvM-KZ7fguzOijBP3kCaJV3mF4qAVkYG_VDn1jEPd-s1CW34HwZevytQBNizpEJwysAqNHbW6rUgXQiq8gkL3N0MIGffabwI1ZckYikOpQrUw7C9kgVMgy_8jr3nlJ3c3nqLwM5dxB95orArSk-GWyk7BRa_N6htBB2xwrmcbwO8lSF5LCcW5zc6N3O24Vne3YeDVB9rfVMrOwk0pKukSuYmiZpws5U4Yuc3ZFHJfpU-Nswbl4SrjxZATG5Eyzo-ecUMIW-MsifLMwLadoxaq3nafpwcQhE4AH4U3Rgi_NOeHuP6iMeStRxUEi_Q7cRASg7TGZgRYAXVmRjaB0X_LGjA_oITzmfOKewUxG2SwtTi3X0-_dxMzQXVNvBp6nKv3THhnM-qAp9rXowmseH-WRqvsuZR0rwuYHig1nLA_gibGmWJ9VwPrhasxmA6kc-PlGkvoYE59lLL3HQYUzNuJsosSRtHix7gcmkpASBIJlPzzvo_10aE9peHNoam78vm8QcFxKA6Lrynva5qCndxXGN7Phh13mkFKmYYGwdufaHDEpeNH0qAWv2GsVHpHKENe0Ll2XqJNeK_lF-zvzFBh0DeRfVy3Lq3G6ZtNvV_sJslzcNQzqqU7gRJEsnzC-bfqJsTGc1NWbYG86OgMV9FmsUohGlDVNIMxAWwtEQXJitWR7UQpei528hz6pJj6uKzrTqoNXAUVjJcQz5DW9wR8OYwdl95FzFHGmqKn8cPTVEvnTV9ISCHPNJ4Hy6_qb2OaUh624h51eo8fvOye_qBFE4B0nnrhbmdtN9PYoxdcF_hvk0Dxt5rvOmiqeTlxueE
+        id: rs_0cc772278fa4f4140068efa9c91ae0819cb57bf002b036ffcb
+        summary: []
+        type: reasoning
+      - action:
+          query: inundaciones Mexico octubre 2025 muertos desaparecidos Veracruz Hidalgo Puebla 15 de octubre
+          type: search
+        id: ws_0cc772278fa4f4140068efa9cb9a94819c86a1054697faa9d6
+        status: completed
+        type: web_search_call
+      - encrypted_content: gAAAAABo76nsdJIaRBHwtMtyHUET6TXVN2QSC9B3268qye-aJa-3t_U3_I8F5URYG19B9IuVC0RDpe2iBHtKklE6KIE_Ugzg57cRDo0wiTQxRDXYwjFxKPXCNyRQHYHkbPh9Dohk8nRcNt9mARzOpwgB6qknrb8odOOhyqe3q8bZ7aAqTh-qimwxkSssyAb-AOjYGAwSXmr2GErt284wA9jughihwSsiRxBmUSwnMx2Tcg8h1Kr4i2lnGcBPgWX_0bCpXGZY4A_JItQv-6Z0erPq5R9xPiYYlMDApTbBBOvFVCUXWuDx8XeAGYm0V18fJxXiEHuktll_Tj7o9c8GDbLF6V_hSsUsPAIQ0RCoKJ6hP4ZgE5CA1ml3WQa7IbPHBU-wKgHcKhwD5eBsRkE4DFfBx7zKTnZtB8FghGYYKkx4Ww7VZVqifoO8JAMvBsmZT3CWp6z69hCg4BOaXnfTE2llfA2B4b0oBTrCl9c4qG5dfx_ELVfRxOJyr8zI_Rxw_xj-fYgCrFb8xfa8XCvbV7thZMAac0eZyq8M9PGRbjvmKn7jNPYK-c2a46Ia6t_s93xxxUtTsCr-rvn2gKTMWm9BPTWqpqkAc2ZTCYxgT7QSs77IAOQuk_9OHDNBM-h0ujCYfzISsieGSnwIwMS3hfg9qiezUcNcwHOHxH7n3T2JfNRa2k37sJmxrFJOPyMezEqar1-XQvAEtXwNcNbYIykNZ4s8G3lCJAnvEzoRzCAmiiAndtejPlVrSZUhsWazjPcDnZ458-1nnMfb-vGK0Cybz3EdNWMFnWN5bjLIP3lc_wLPa_dWSz1Dm_bTckJWKAujV1jQITft04Nlo4i9vLW5vWb7Jsml4fiSYtvuALQkNzhOlMbdKvrL29UFobkfCB1hX4Cgj-Hkn7P5WjeYtxKXkOb171JGIEav5qQcfbImOuFxhg_7EUbaefIyGa1J2eGb7DRCcmbqnur5hvzMzPGNqHxWvEunVxf9ZvUEXsPoNPyC2Jkot_ntLSlwRtuuMqu9teK6NfWi0qCJeqbhEN-OWDpFtrmSEYLr1FNS5_XM6_VZ10o1GRW8notAm5RVdn7zDvjs_s5cXv4efs0UanOOAS6e_OIJeX7VHdiKLnHLXxZ-adp5NR2kKIw0hgtvcrCR7KHvr4itOny8zy_7EfeYgX8B66favNVeu6JLGEVIOs1bN3_evQYUyRPJAn3FwM_DIUZqQ1xJVewm1RAxvaENoT_gh3pib4vdjXxeGjJg7VsAkPpiHf0TSYvGKiAtdUuUHNfNv8X5MGuh40ST8ji-UHeWkUe8fzcVoypPVNApyO9bET3RgirucglYbObm3WwucxyZDkJC9nYb1LptN6K3xvJ8a97njRCnzjE4Z0SfOtRBotblGo0IN7cCVI-SJrOlXXdvm8PUOOofKmzM9Y-Seo_Vapq3NBfLGZar7Sg1RtHXTXpJABZZ1HNgfvmw0qqKka-cUyS2qByP6d1ZRc2MTAofNcJXEiiBF62f1JXK0yMwKyCacPBQ26J7xHvmDtz-lAZGO6n8GvWJy2UkGthYZLIHTfyKfc46t74ZunD1daW5_QK7dCMMdXylK8lQaSBtK8UIzlbDcT7cIeoO-9iGSEMj623RjwMqpAJLYW8cuoccnTnH3cJ0032hhuaTyvPnW98_2QdzERSY6booWeLI5QI4yWcmJgs8gnhQC1-UEZrZ-Y3_344Xv8t2Dk-ihlHMN0aZ5MpBkT_FXqspXn5a6HIVLEhdtckz1LWK6ZZxcMDBVZAy5pqTmIs_Ro303y1_5Z_oiKSl-ISCkUhCGCCb5JpLozG5tJZTqDij8u91L4c3fgzxs2Myk-Dk8eiAtNyR1kJTPLzFTqHrFvlTXbU5UaHqBjGLhlVXY-9H7vRB4Ssk0NWyQY5jv1L2jnhy64l6KtNBqY827HVOnsaT-XWlM4McnFT1MJ-qO41NzR-gTZva-KzjDgVj50_MBmxeuBLsnDVq26Iw2d9OQpyzo_qiN_Fs2G9zpaUr84-W9mNsoV3TljiQk9-D1xdy4ZXJ7oicW1Ah-Ua4lNGI8wKHujDDX-sw0FaEQfKKPrN4Ou_lXAStakNhnvn2C3sbPlUk6Mie52ky1Pyq9L3SMwFtNa_Sx0imAANBAqekA6BMwc_Q_Y8jgmtZatBAibaWfWfcCyWNy832S6wJ5L0_fOhYVwT4PMvOw9AIwQbTSAJtuLN7zxZOv2Qw5_UnatJR6MgXbQc0BrmWI_FGuZdOuv2TMGKZkV9ynaZ5L-Dk0P90lpZwNmpbddh1y7TQNY6Os6YPwwcyPSMgTD1fY3eJPWDEv8OxRIrFJeth_tJMb-IX2fozD7ZAXUXuEFqtCFZ7yqoO-AbW8rgLWU5ILiJJk7neHSyWrsiDf0eyILU-tC1QoGuP-Md5hehvrobjVQgkARdzIUAy_8RIeG956NoG1nqIM-5tu0A8xjv2rBf5vTw5p6rbtCiQtTacMVrprXLz-p23XqrOAZuEvhc7mMJBwEJcGq8w12oil-oakW3gbHWzKWUb_6bLGYaI7uBjSH_HDtcU79uONBYGDKRG_eZa0EcupHQ8twVBNviyz2bSubTN4gWvjz4k1nHkmQBagFQzy6ogcCVldZVyLtib2TfMH64DVeEFBEM-aUKcs81R42r8aZjW_I3RFRGS8hodNs0Y1ZfkYms_kBNOZc3pw2iP4rtOLDWatSGfUWN8owiQETiOI2AD70HoXcVBmjyRpxheI_mBVz6YIxMzf0esd5U118NzOyxqfQ66EvmAzmN0iSrbR2Vucx4CojBzweQ0KoWn_HV4U3qjkdXcmdvncbTK18LPoOhS6z7BhEKXDhd1bhI1f5AyirZhqR8xvK0r6aeQdkL7OWyQoDWJM-CRtn0S4O3qucwNA1XFVPt5fF8rysehSjLN4GH0AAueDBc9_BGx0fQMsmpRjUO71W0P-C4pY-eYvUu60u6ng2YELowM7Poak8xWOsvlGqYuXj2s4Tuj6hJPeD6vPoUDaxfFPjMijSkpfASxJvNbE8DpftoPrfiXIG-LDfGIXiJISgDzBgTllf3cw1lhSYq51WfqcBaWsIvXWGGbvnfqE2bA7nRwNUBkmxd9NC38-DAAySadaHIBjsIgv7VEKKp841cvTQ4bB-8Fp1d7pZOvMsq4GjaKCQNfjNcWL87KyyJ-GonLExJStrdNGxwS_9AYNA-6KEdtz-V9YRFzbkHl4406ueOody8O7jvucmoyKFTU1S2yJ5Wbd1gNYA-6wV_II78hbI6dXjTEGvdbRWIKkTktnD6pgDhHjkdS6uzsTcUYTkj60SxMujKv2xlRevDANdgFDB0-zKkv0ntRINr5_CHmnLbW3NYAAyGQ9wOgvfNEMno0EVLQU9F596lscKSXAl4djL1F_GSSn80lb0ujdu77Xh130yORuio6SSF9DAXg1wYI8O0Il2nyJg-0DWTyy6mCG4zPmzyS6yRH64o4kOGmZBuOJSFvcF2Qy0h0h0xBvYhN9fBJh8AcjuYP9TWB8s8Z3KKYOXYlPTPlbN4BaowBVfmzG6_aJaDvUsT1bk9IruW4ENtwf6OxXs0nqEwQEGBYWDAmcdS5hESe7JuVQ1O2bLkg2fYl63VznOUIrYW5hbHMI6pZ3EMsuJD5ITzULCl-WlX_qmduuAKnnFbSzua4Rv6ZUOk_GC-BCIfcnro0RTJTna2hl8M-YP0g05v68zDFl4x1s1xsvii8VF4p_t6EGaPY00iUgB__4uWwdIkuywTAaLgvdtWonA7aCvEZr6a04ZyUD7jf6Acd1cwxkCe2Ed-GVd9f6vQcuXUwAwnuK_gsWsqLfPfA_9U3mxMABZVXjuLwxZlxGW1rxdRQZUdqeQvIBLVt3ozMyqu4mQfs0r4gaLUFff_qHkckNvbvJwUDquokNysr0e1matigksVRDUUQc5aXvR8KeQZKUvjQLh5Rv_gvUmkBO2pQQUJCG3IPQqf6lU1Y_86sxHVegaoZ52jbbEq32NO9gQ-3LlMPvmm-FTmdfaBxNZZ2GVHgjLdvmshTd3XrdrRlqXc2dwFsLbIi6Efg6Zy6Vr8HeZiHWtJjiNT6jurXE0aNkq3SSWHEW6tNcn1BXToFUI7R4vozxEzwQG9e-acnS6OZsRie5dDmQ02CXSlJYu8s8TwyXEhy3vmaGDrtVIXlBGdb-6lAjspALKC4I6k_6uwiv492gFe7usw9mvHgIe9b5y2h0z1VUZ1lr152G1Fw_2R-1C6RxwUPZuD3ZVSw5LvO18VApN4_jDwD6l6i9QcCAGgEtz5rSp1SfbYXpTCFd5ulbZAh0JgU_OUnK4QlXZL7vdfs0cKU43IQaDAwg2zJjKodk87JDoe6YmDfrFWZ_qkPWdfbVtvKOMg2bMqJVU0LbKGH3H_MKnQVnHH7UkcdTfu7wwbYIdddQHQPEzuzOwBTg9nz46nSKfZCIQ3FES6gP7bBDuw3IhwzZq30hAS0XPspU8OQuvAQcUcpjSYlKkNvWvw1ZaGE-wJsA2UgyUDeQ_gUAmvjEqXTkomk5EDiDWg4R7TMi3bzbVUylLcNOopOi_X6Quij0dWoV6tZJqKq8z1BJUeV7Dl9dcnN24MTQ4YFMV3r59zAD4_DBWkMorwfllcFB1-i3mq2RdtErHFay7BATssuoSvaVhZbfcc6OsqE6UcZTUr7ctBCKIpg5EJTAZRgZcUTUQMQYaJmCgBi5QY4UIf9oxxd8URQKz-nT2G9vKrCTRC7h2J2aG3RaM4TVYzJQUADkbj5ejVMTELy2JjG_LQLMLzd_yR4VeSOfn3RVbo_P7A5P0bOp4ftma8Qe5kDC2L2H252tNr2gU82ozygJdqi_2ZB_ladw5fojKVMe2w1dBg_g-c5HmJlIxWp0N_dj8hDTS_zC2J6X6zIFkhW9JTSRgVzdDJd8oZEGNtTbAo__R74yEHll9L52bRu68NTxrT4cQ4RIbA5dQUl0QgNwRr67roltQeWXj1eS47Pe070ntnScAgTCF8431lM2XiK1iPP6fbi23SWE6oD4JQaaAQkeHoBqjqPcTeoPSHLmpKkZS55e6XPlxbQp_exrt8yLJv9fs7wrfG0egRzJwIRd_H8vxCg7f65EUIIAyUp8TTaHpYWQJMmmWtAa035TcuwzRtswUYcWLZZWbqBB90m9EM9huFAmF6xJypAfjWFRm0Mtwh7AgqTTEuw6cnNw5lzGQKJqnP2kCiZ6gVb97A67qQZMOs_zDp8E68RgEzDIm14JEAKrIs0nI3VqrNyy0DGzyvn6TWZUAYqUdUGjchOYc_tj3_cr7_Icu1VWG48SNIlXMKHi0k09efcyTzfaaTXQS5UdlenHvFg9j1WO6oyORpaGbA3L5XUNwH1yX6d9wN8
+        id: rs_0cc772278fa4f4140068efa9d096c0819c99fea06cc2948069
+        summary: []
+        type: reasoning
+      - content:
+        - annotations:
+          - end_index: 281
+            start_index: 188
+            title: Landslides and flooding cut off 300 communities in Mexico with dozens dead and missing
+            type: url_citation
+            url: https://apnews.com/article/5d036e18057361281e984b44402d3b1b?utm_source=openai
+          logprobs: []
+          text: Severe floods and landslides across Veracruz, Hidalgo, and Puebla have cut off hundreds of communities and
+            left dozens dead and many missing, prompting a major federal emergency response. ([apnews.com](https://apnews.com/article/5d036e18057361281e984b44402d3b1b?utm_source=openai))
+          type: output_text
+        id: msg_0cc772278fa4f4140068efa9e8d5c0819caea9f8f22692335d
+        role: assistant
+        status: completed
+        type: message
+      parallel_tool_calls: true
+      previous_response_id: null
+      prompt_cache_key: null
+      reasoning:
+        effort: medium
+        summary: null
+      safety_identifier: null
+      service_tier: default
+      status: completed
+      store: true
+      temperature: 1.0
+      text:
+        format:
+          type: text
+        verbosity: medium
+      tool_choice: auto
+      tools:
+      - filters: null
+        search_context_size: medium
+        type: web_search
+        user_location:
+          city: Mexico City
+          country: MX
+          region: null
+          timezone: null
+          type: approximate
+      top_logprobs: 0
+      top_p: 1.0
+      truncation: disabled
+      usage:
+        input_tokens: 12594
+        input_tokens_details:
+          cached_tokens: 3200
+        output_tokens: 1150
+        output_tokens_details:
+          reasoning_tokens: 1088
+        total_tokens: 13744
+      user: null
+    status:
+      code: 200
+      message: OK
+version: 1

{pydantic_ai-1.2.1 → pydantic_ai-1.3.0}/tests/conftest.py

@@ -14,7 +14,7 @@ from datetime import datetime
 from functools import cached_property
 from pathlib import Path
 from types import ModuleType
-from typing import TYPE_CHECKING, Any, TypeAlias
+from typing import TYPE_CHECKING, Any, TypeAlias, cast
 
 import httpx
 import pytest
@@ -463,21 +463,13 @@ async def vertex_provider(vertex_provider_auth: None):  # pragma: lax no cover
         pytest.skip('Requires properly configured local google vertex config to pass')
 
     try:
-        from google.genai import Client
-
-        from pydantic_ai.providers.google import GoogleProvider
+        from pydantic_ai.providers.google import GoogleProvider, VertexAILocation
     except ImportError:  # pragma: lax no cover
         pytest.skip('google is not installed')
 
     project = os.getenv('GOOGLE_PROJECT', 'pydantic-ai')
     location = os.getenv('GOOGLE_LOCATION', 'global')
-    client = Client(vertexai=True, project=project, location=location)
-
-    try:
-        yield GoogleProvider(client=client)
-    finally:
-        client.aio._api_client._httpx_client.close()  # type: ignore
-        await client.aio._api_client._async_httpx_client.aclose()  # type: ignore
+    yield GoogleProvider(project=project, location=cast(VertexAILocation, location))
 
 
 @pytest.fixture()

{pydantic_ai-1.2.1 → pydantic_ai-1.3.0}/tests/evals/test_dataset.py

@@ -1555,6 +1555,16 @@ async def test_evaluate_async_logfire(
                             'gen_ai.operation.name': {},
                             'n_cases': {},
                             'name': {},
+                            'logfire.experiment.metadata': {
+                                'type': 'object',
+                                'properties': {
+                                    'averages': {
+                                        'type': 'object',
+                                        'title': 'ReportCaseAggregate',
+                                        'x-python-datatype': 'PydanticModel',
+                                    }
+                                },
+                            },
                             'task_name': {},
                         },
                         'type': 'object',
@@ -1563,6 +1573,18 @@ async def test_evaluate_async_logfire(
                     'logfire.msg_template': 'evaluate {name}',
                     'logfire.span_type': 'span',
                     'n_cases': 2,
+                    'logfire.experiment.metadata': {
+                        'n_cases': 2,
+                        'averages': {
+                            'name': 'Averages',
+                            'scores': {'confidence': 1.0},
+                            'labels': {},
+                            'metrics': {},
+                            'assertions': 1.0,
+                            'task_duration': 1.0,
+                            'total_duration': 9.0,
+                        },
+                    },
                     'name': 'mock_async_task',
                     'task_name': 'mock_async_task',
                 },

pydantic_ai-1.3.0/tests/models/cassettes/test_bedrock/test_bedrock_model_stream_empty_text_delta.yaml

@@ -0,0 +1,66 @@
+interactions:
+- request:
+    body: '{"messages": [{"role": "user", "content": [{"text": "Hi"}]}], "system": [], "inferenceConfig": {}}'
+    headers:
+      amz-sdk-invocation-id:
+      - !!binary |
+        Zjg3NDYxYzEtZDQzZS00YThkLTkwNDMtNDcwOTNjMWZlMGM2
+      amz-sdk-request:
+      - !!binary |
+        YXR0ZW1wdD0x
+      content-length:
+      - '98'
+      content-type:
+      - !!binary |
+        YXBwbGljYXRpb24vanNvbg==
+    method: POST
+    uri: https://bedrock-runtime.us-east-1.amazonaws.com/model/openai.gpt-oss-120b-1%3A0/converse-stream
+  response:
+    body:
+      string: !!binary |
+        AAAAsgAAAFKKQLTFCzpldmVudC10eXBlBwAMbWVzc2FnZVN0YXJ0DTpjb250ZW50LXR5cGUHABBh
+        cHBsaWNhdGlvbi9qc29uDTptZXNzYWdlLXR5cGUHAAVldmVudHsicCI6ImFiY2RlZmdoaWprbG1u
+        b3BxcnN0dXZ3eHl6QUJDREVGR0hJSktMTU5PUFFSU1RVVldYWVowIiwicm9sZSI6ImFzc2lzdGFu
+        dCJ9QP+CaAAAAKIAAABXmsrXyAs6ZXZlbnQtdHlwZQcAEWNvbnRlbnRCbG9ja0RlbHRhDTpjb250
+        ZW50LXR5cGUHABBhcHBsaWNhdGlvbi9qc29uDTptZXNzYWdlLXR5cGUHAAVldmVudHsiY29udGVu
+        dEJsb2NrSW5kZXgiOjAsImRlbHRhIjp7InRleHQiOiIifSwicCI6ImFiY2RlZmdoaSJ9NTLrNAAA
+        ALkAAABW+v1BzQs6ZXZlbnQtdHlwZQcAEGNvbnRlbnRCbG9ja1N0b3ANOmNvbnRlbnQtdHlwZQcA
+        EGFwcGxpY2F0aW9uL2pzb24NOm1lc3NhZ2UtdHlwZQcABWV2ZW50eyJjb250ZW50QmxvY2tJbmRl
+        eCI6MCwicCI6ImFiY2RlZmdoaWprbG1ub3BxcnN0dXZ3eHl6QUJDREVGR0hJSktMTU5PUFFSU1RV
+        VldYWVowIn2IAETUAAABWQAAAFducmUhCzpldmVudC10eXBlBwARY29udGVudEJsb2NrRGVsdGEN
+        OmNvbnRlbnQtdHlwZQcAEGFwcGxpY2F0aW9uL2pzb24NOm1lc3NhZ2UtdHlwZQcABWV2ZW50eyJj
+        b250ZW50QmxvY2tJbmRleCI6MSwiZGVsdGEiOnsicmVhc29uaW5nQ29udGVudCI6eyJ0ZXh0Ijoi
+        VGhlIHVzZXIganVzdCBzYXlzIFwiSGlcIi4gV2UgbmVlZCB0byByZXNwb25kIGFwcHJvcHJpYXRl
+        bHksIGZyaWVuZGx5IGdyZWV0aW5nLiBObyBzcGVjaWFsIGluc3RydWN0aW9ucy4gU2hvdWxkIGJl
+        IHNob3J0LiJ9fSwicCI6ImFiY2RlZmdoaWprbG1ub3BxcnN0dXZ3eHl6QUJDREVGR0hJSktMTU5P
+        UFFSU1RVVldYIn2DKrHfAAAAsQAAAFbKjQoMCzpldmVudC10eXBlBwAQY29udGVudEJsb2NrU3Rv
+        cA06Y29udGVudC10eXBlBwAQYXBwbGljYXRpb24vanNvbg06bWVzc2FnZS10eXBlBwAFZXZlbnR7
+        ImNvbnRlbnRCbG9ja0luZGV4IjoxLCJwIjoiYWJjZGVmZ2hpamtsbW5vcHFyc3R1dnd4eXpBQkNE
+        RUZHSElKS0xNTk9QUVJTIn1Cad1JAAAA5AAAAFdNeXohCzpldmVudC10eXBlBwARY29udGVudEJs
+        b2NrRGVsdGENOmNvbnRlbnQtdHlwZQcAEGFwcGxpY2F0aW9uL2pzb24NOm1lc3NhZ2UtdHlwZQcA
+        BWV2ZW50eyJjb250ZW50QmxvY2tJbmRleCI6MiwiZGVsdGEiOnsidGV4dCI6IkhlbGxvISBIb3cg
+        Y2FuIEkgaGVscCJ9LCJwIjoiYWJjZGVmZ2hpamtsbW5vcHFyc3R1dnd4eXpBQkNERUZHSElKS0xN
+        Tk9QUVJTVFVWV1hZWjAxIn2L4Y7bAAAAuwAAAFf3OiI7CzpldmVudC10eXBlBwARY29udGVudEJs
+        b2NrRGVsdGENOmNvbnRlbnQtdHlwZQcAEGFwcGxpY2F0aW9uL2pzb24NOm1lc3NhZ2UtdHlwZQcA
+        BWV2ZW50eyJjb250ZW50QmxvY2tJbmRleCI6MiwiZGVsdGEiOnsidGV4dCI6IiB5b3UgdG9kYXk/
+        In0sInAiOiJhYmNkZWZnaGlqa2xtbm9wcXJzdHV2dyJ9dKsF8AAAAI4AAABW6fwlWws6ZXZlbnQt
+        dHlwZQcAEGNvbnRlbnRCbG9ja1N0b3ANOmNvbnRlbnQtdHlwZQcAEGFwcGxpY2F0aW9uL2pzb24N
+        Om1lc3NhZ2UtdHlwZQcABWV2ZW50eyJjb250ZW50QmxvY2tJbmRleCI6MiwicCI6ImFiY2RlZmdo
+        aWoifbla37UAAACDAAAAUY8IdEkLOmV2ZW50LXR5cGUHAAttZXNzYWdlU3RvcA06Y29udGVudC10
+        eXBlBwAQYXBwbGljYXRpb24vanNvbg06bWVzc2FnZS10eXBlBwAFZXZlbnR7InAiOiJhYiIsInN0
+        b3BSZWFzb24iOiJlbmRfdHVybiJ9kFbppAAAAOQAAABOKRLS4Qs6ZXZlbnQtdHlwZQcACG1ldGFk
+        YXRhDTpjb250ZW50LXR5cGUHABBhcHBsaWNhdGlvbi9qc29uDTptZXNzYWdlLXR5cGUHAAVldmVu
+        dHsibWV0cmljcyI6eyJsYXRlbmN5TXMiOjc1M30sInAiOiJhYmNkZWZnaGlqa2xtbiIsInVzYWdl
+        Ijp7ImlucHV0VG9rZW5zIjo3MCwib3V0cHV0VG9rZW5zIjo0Mywic2VydmVyVG9vbFVzYWdlIjp7
+        fSwidG90YWxUb2tlbnMiOjExM319saJH4A==
+    headers:
+      connection:
+      - keep-alive
+      content-type:
+      - application/vnd.amazon.eventstream
+      transfer-encoding:
+      - chunked
+    status:
+      code: 200
+      message: OK
+version: 1

{pydantic_ai-1.2.1 → pydantic_ai-1.3.0}/tests/models/test_bedrock.py

@@ -32,7 +32,9 @@ from pydantic_ai import (
 )
 from pydantic_ai.agent import Agent
 from pydantic_ai.exceptions import ModelRetry
+from pydantic_ai.messages import AgentStreamEvent
 from pydantic_ai.models import ModelRequestParameters
+from pydantic_ai.run import AgentRunResult, AgentRunResultEvent
 from pydantic_ai.tools import ToolDefinition
 from pydantic_ai.usage import RequestUsage, RunUsage
 
@@ -1272,3 +1274,33 @@ async def test_bedrock_no_tool_choice(bedrock_provider: BedrockProvider):
             ]
         }
     )
+
+
+async def test_bedrock_model_stream_empty_text_delta(allow_model_requests: None, bedrock_provider: BedrockProvider):
+    model = BedrockConverseModel(model_name='openai.gpt-oss-120b-1:0', provider=bedrock_provider)
+    agent = Agent(model)
+
+    result: AgentRunResult | None = None
+    events: list[AgentStreamEvent] = []
+    async for event in agent.run_stream_events('Hi'):
+        if isinstance(event, AgentRunResultEvent):
+            result = event.result
+        else:
+            events.append(event)
+
+    assert result is not None
+    # The response stream contains `{'contentBlockDelta': {'delta': {'text': ''}, 'contentBlockIndex': 0}}`, but our response should not have any empty text parts.
+    assert not any(part.content == '' for part in result.response.parts if isinstance(part, TextPart))
+    assert events == snapshot(
+        [
+            PartStartEvent(
+                index=0,
+                part=ThinkingPart(
+                    content='The user just says "Hi". We need to respond appropriately, friendly greeting. No special instructions. Should be short.'
+                ),
+            ),
+            PartStartEvent(index=1, part=TextPart(content='Hello! How can I help')),
+            FinalResultEvent(tool_name=None, tool_call_id=None),
+            PartDeltaEvent(index=1, delta=TextPartDelta(content_delta=' you today?')),
+        ]
+    )

{pydantic_ai-1.2.1 → pydantic_ai-1.3.0}/tests/models/test_deepseek.py

@@ -1,24 +1,17 @@
 from __future__ import annotations as _annotations
 
-from typing import Any
-
 import pytest
-from dirty_equals import IsListOrTuple
 from inline_snapshot import snapshot
 
 from pydantic_ai import (
     Agent,
-    FinalResultEvent,
     ModelRequest,
     ModelResponse,
-    PartDeltaEvent,
-    PartStartEvent,
     TextPart,
-    TextPartDelta,
     ThinkingPart,
-    ThinkingPartDelta,
     UserPromptPart,
 )
+from pydantic_ai.run import AgentRunResult, AgentRunResultEvent
 from pydantic_ai.usage import RequestUsage
 
 from ..conftest import IsDatetime, IsStr, try_import
@@ -71,27 +64,42 @@ async def test_deepseek_model_thinking_stream(allow_model_requests: None, deepse
     deepseek_model = OpenAIChatModel('deepseek-reasoner', provider=DeepSeekProvider(api_key=deepseek_api_key))
     agent = Agent(model=deepseek_model)
 
-    event_parts: list[Any] = []
-    async with agent.iter(user_prompt='Hello') as agent_run:
-        async for node in agent_run:
-            if Agent.is_model_request_node(node) or Agent.is_call_tools_node(node):
-                async with node.stream(agent_run.ctx) as request_stream:
-                    async for event in request_stream:
-                        event_parts.append(event)
+    result: AgentRunResult | None = None
+    async for event in agent.run_stream_events(user_prompt='How do I cross the street?'):
+        if isinstance(event, AgentRunResultEvent):
+            result = event.result
 
-    assert event_parts == IsListOrTuple(
-        positions={
-            0: snapshot(
-                PartStartEvent(
-                    index=0, part=ThinkingPart(content='H', id='reasoning_content', provider_name='deepseek')
-                )
+    assert result is not None
+    assert result.all_messages() == snapshot(
+        [
+            ModelRequest(
+                parts=[
+                    UserPromptPart(
+                        content='How do I cross the street?',
+                        timestamp=IsDatetime(),
+                    )
+                ]
+            ),
+            ModelResponse(
+                parts=[
+                    ThinkingPart(
+                        content=IsStr(),
+                        id='reasoning_content',
+                        provider_name='deepseek',
+                    ),
+                    TextPart(content='Hello there! 😊 How can I help you today?'),
+                ],
+                usage=RequestUsage(
+                    input_tokens=6,
+                    output_tokens=212,
+                    details={'prompt_cache_hit_tokens': 0, 'prompt_cache_miss_tokens': 6, 'reasoning_tokens': 198},
+                ),
+                model_name='deepseek-reasoner',
+                timestamp=IsDatetime(),
+                provider_name='deepseek',
+                provider_details={'finish_reason': 'stop'},
+                provider_response_id='33be18fc-3842-486c-8c29-dd8e578f7f20',
+                finish_reason='stop',
             ),
-            1: snapshot(PartDeltaEvent(index=0, delta=ThinkingPartDelta(content_delta='mm', provider_name='deepseek'))),
-            2: snapshot(PartDeltaEvent(index=0, delta=ThinkingPartDelta(content_delta=',', provider_name='deepseek'))),
-            198: snapshot(PartStartEvent(index=1, part=TextPart(content='Hello'))),
-            199: snapshot(FinalResultEvent(tool_name=None, tool_call_id=None)),
-            200: snapshot(PartDeltaEvent(index=1, delta=TextPartDelta(content_delta=' there'))),
-            201: snapshot(PartDeltaEvent(index=1, delta=TextPartDelta(content_delta='!'))),
-        },
-        length=211,
+        ]
     )

{pydantic_ai-1.2.1 → pydantic_ai-1.3.0}/tests/models/test_groq.py

@@ -5160,7 +5160,6 @@ async def test_tool_use_failed_error_streaming(allow_model_requests: None, groq_
             ),
             ModelResponse(
                 parts=[
-                    TextPart(content=''),
                     ThinkingPart(
                         content="""\
 The user requests to call the tool with non-existent parameters to test error handling. We need to call the function "get_something_by_name" with wrong parameters. The function expects a single argument object with "name". Non-existent parameters means we could provide a wrong key, or missing name. Let's provide an object with wrong key "nonexistent": "value". That should cause error. So we call the function with {"nonexistent": "test"}.
@@ -5205,7 +5204,6 @@ We need to output the call.\
             ),
             ModelResponse(
                 parts=[
-                    TextPart(content=''),
                     ThinkingPart(content='We need to call with correct param: name. Use a placeholder name.'),
                     ToolCallPart(
                         tool_name='get_something_by_name',

{pydantic_ai-1.2.1 → pydantic_ai-1.3.0}/tests/models/test_huggingface.py

@@ -9,7 +9,6 @@ from typing import Any, Literal, cast
 from unittest.mock import Mock
 
 import pytest
-from dirty_equals import IsListOrTuple
 from inline_snapshot import snapshot
 from typing_extensions import TypedDict
 
@@ -18,19 +17,14 @@ from pydantic_ai import (
     AudioUrl,
     BinaryContent,
     DocumentUrl,
-    FinalResultEvent,
     ImageUrl,
     ModelRequest,
     ModelResponse,
     ModelRetry,
-    PartDeltaEvent,
-    PartStartEvent,
     RetryPromptPart,
     SystemPromptPart,
     TextPart,
-    TextPartDelta,
     ThinkingPart,
-    ThinkingPartDelta,
     ToolCallPart,
     ToolReturnPart,
     UserPromptPart,
@@ -38,6 +32,7 @@ from pydantic_ai import (
 )
 from pydantic_ai.exceptions import ModelHTTPError
 from pydantic_ai.result import RunUsage
+from pydantic_ai.run import AgentRunResult, AgentRunResultEvent
 from pydantic_ai.settings import ModelSettings
 from pydantic_ai.tools import RunContext
 from pydantic_ai.usage import RequestUsage
@@ -978,35 +973,32 @@ async def test_hf_model_thinking_part_iter(allow_model_requests: None, huggingfa
     )
     agent = Agent(m)
 
-    event_parts: list[Any] = []
-    async with agent.iter(user_prompt='How do I cross the street?') as agent_run:
-        async for node in agent_run:
-            if Agent.is_model_request_node(node) or Agent.is_call_tools_node(node):
-                async with node.stream(agent_run.ctx) as request_stream:
-                    async for event in request_stream:
-                        event_parts.append(event)
-
-    assert event_parts == snapshot(
-        IsListOrTuple(
-            positions={
-                0: PartStartEvent(index=0, part=ThinkingPart(content='')),
-                1: PartDeltaEvent(index=0, delta=ThinkingPartDelta(content_delta='\n')),
-                2: PartDeltaEvent(index=0, delta=ThinkingPartDelta(content_delta='Okay')),
-                3: PartDeltaEvent(index=0, delta=ThinkingPartDelta(content_delta=',')),
-                4: PartDeltaEvent(index=0, delta=ThinkingPartDelta(content_delta=' the')),
-                5: PartDeltaEvent(index=0, delta=ThinkingPartDelta(content_delta=' user')),
-                6: PartDeltaEvent(index=0, delta=ThinkingPartDelta(content_delta=' is')),
-                7: PartDeltaEvent(index=0, delta=ThinkingPartDelta(content_delta=' asking')),
-                413: PartStartEvent(index=1, part=TextPart(content='Cross')),
-                414: FinalResultEvent(tool_name=None, tool_call_id=None),
-                415: PartDeltaEvent(index=1, delta=TextPartDelta(content_delta='ing')),
-                416: PartDeltaEvent(index=1, delta=TextPartDelta(content_delta=' the')),
-                417: PartDeltaEvent(index=1, delta=TextPartDelta(content_delta=' street')),
-                418: PartDeltaEvent(index=1, delta=TextPartDelta(content_delta=' safely')),
-                419: PartDeltaEvent(index=1, delta=TextPartDelta(content_delta=' requires')),
-                420: PartDeltaEvent(index=1, delta=TextPartDelta(content_delta=' attent')),
-                421: PartDeltaEvent(index=1, delta=TextPartDelta(content_delta='iveness')),
-            },
-            length=1062,
-        )
+    result: AgentRunResult | None = None
+    async for event in agent.run_stream_events(user_prompt='How do I cross the street?'):
+        if isinstance(event, AgentRunResultEvent):
+            result = event.result
+
+    assert result is not None
+    assert result.all_messages() == snapshot(
+        [
+            ModelRequest(
+                parts=[
+                    UserPromptPart(
+                        content='How do I cross the street?',
+                        timestamp=IsDatetime(),
+                    )
+                ]
+            ),
+            ModelResponse(
+                parts=[
+                    ThinkingPart(content=IsStr()),
+                    TextPart(content=IsStr()),
+                ],
+                model_name='Qwen/Qwen3-235B-A22B',
+                timestamp=IsDatetime(),
+                provider_name='huggingface',
+                provider_details={'finish_reason': 'stop'},
+                provider_response_id='chatcmpl-357f347a3f5d4897b36a128fb4e4cf7b',
+            ),
+        ]
     )