opensandbox-code-interpreter 1.1.1__tar.gz → 1.2.1__tar.gz

{opensandbox_code_interpreter-1.1.1 → opensandbox_code_interpreter-1.2.1}/.gitignore

@@ -294,3 +294,7 @@ obj/
 # fast-sandbox pinned source checkout (materialized by
 # manifests/release/build-fast-sandbox.sh from manifests/third-party/fast-sandbox.commit)
 .fast-sandbox/
+
+# Test-only localhost certificate pair for the egress OSEP-0025 resolver stub
+# (fixtures/README.md documents regeneration; these certify nothing).
+!components/egress/tests/resolver/fixtures/*.pem

{opensandbox_code_interpreter-1.1.1 → opensandbox_code_interpreter-1.2.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: opensandbox-code-interpreter
-Version: 1.1.1
+Version: 1.2.1
 Summary: OpenSandbox Code Interpreter Python SDK - Advanced code execution with persistent contexts
 Project-URL: Homepage, https://open-sandbox.ai
 Project-URL: Repository, https://github.com/opensandbox-group/OpenSandbox
@@ -223,24 +223,20 @@ Classifier: Programming Language :: Python :: 3.13
 Classifier: Topic :: Software Development :: Libraries
 Classifier: Typing :: Typed
 Requires-Python: >=3.10
-Requires-Dist: opensandbox<2.0.0,>=1.1.1
+Requires-Dist: opensandbox<2.0.0,>=1.2.1
 Requires-Dist: pydantic<3.0,>=2.4.2
 Description-Content-Type: text/markdown
 
 # OpenSandbox Code Interpreter SDK for Python
 
 
-A Python SDK for executing code in secure, isolated sandboxes. It provides a high-level API for running Python, Java,
-Go, TypeScript, and other languages safely, with support for code execution contexts.
+A Python SDK for executing code in secure, isolated sandboxes. It provides a high-level API for running Python, Java, Go, TypeScript, and other languages safely, with support for code execution contexts.
 
 ## Prerequisites
 
-This SDK requires a Docker image containing the Code Interpreter runtime environment. You must use the
-`opensandbox/code-interpreter` image (or a derivative) which includes pre-installed runtimes for Python, Java, Go,
-Node.js, etc.
+This SDK requires a Docker image containing the Code Interpreter runtime environment. You must use the `opensandbox/code-interpreter` image (or a derivative) which includes pre-installed runtimes for Python, Java, Go, Node.js, etc.
 
-For detailed information about supported languages and versions, refer to the
-[Environment Documentation](https://github.com/opensandbox-group/sandbox-images).
+For detailed information about supported languages and versions, refer to the [Environment Documentation](https://github.com/opensandbox-group/sandbox-images).
 
 ## Installation
 
@@ -258,8 +254,7 @@ uv add opensandbox-code-interpreter
 
 ## Quick Start
 
-The following example demonstrates how to create a sandbox with a specific runtime configuration and execute a simple
-script.
+The following example demonstrates how to create a sandbox with a specific runtime configuration and execute a simple script.
 
 ```python
 import asyncio
@@ -366,13 +361,11 @@ execution = await sandbox.commands.run("pip install pandas numpy")
 
 ### Docker Image
 
-The Code Interpreter SDK relies on a specialized environment. Ensure your sandbox provider has the
-`opensandbox/code-interpreter` image available.
+The Code Interpreter SDK relies on a specialized environment. Ensure your sandbox provider has the `opensandbox/code-interpreter` image available.
 
 ### Language Version Selection
 
-You can specify the desired version of a programming language by setting the corresponding environment variable when
-creating the `Sandbox`.
+You can specify the desired version of a programming language by setting the corresponding environment variable when creating the `Sandbox`.
 
 | Language | Environment Variable | Example Value | Default (if unset) |
 | -------- | -------------------- | ------------- | ------------------ |
@@ -385,9 +378,7 @@ creating the `Sandbox`.
 
 ### 0. Run with `language` (default language context)
 
-You can pass `language` directly (recommended: `SupportedLanguage.*`) and skip `create_context`.
-When `context.id` is omitted, **execd will create/reuse a default session for that language**, so
-state can persist across runs:
+You can pass `language` directly (recommended: `SupportedLanguage.*`) and skip `create_context`. When `context.id` is omitted, **execd will create/reuse a default session for that language**, so state can persist across runs:
 
 ```python
 from code_interpreter import SupportedLanguage

{opensandbox_code_interpreter-1.1.1 → opensandbox_code_interpreter-1.2.1}/README.md

@@ -1,17 +1,13 @@
 # OpenSandbox Code Interpreter SDK for Python
 
 
-A Python SDK for executing code in secure, isolated sandboxes. It provides a high-level API for running Python, Java,
-Go, TypeScript, and other languages safely, with support for code execution contexts.
+A Python SDK for executing code in secure, isolated sandboxes. It provides a high-level API for running Python, Java, Go, TypeScript, and other languages safely, with support for code execution contexts.
 
 ## Prerequisites
 
-This SDK requires a Docker image containing the Code Interpreter runtime environment. You must use the
-`opensandbox/code-interpreter` image (or a derivative) which includes pre-installed runtimes for Python, Java, Go,
-Node.js, etc.
+This SDK requires a Docker image containing the Code Interpreter runtime environment. You must use the `opensandbox/code-interpreter` image (or a derivative) which includes pre-installed runtimes for Python, Java, Go, Node.js, etc.
 
-For detailed information about supported languages and versions, refer to the
-[Environment Documentation](https://github.com/opensandbox-group/sandbox-images).
+For detailed information about supported languages and versions, refer to the [Environment Documentation](https://github.com/opensandbox-group/sandbox-images).
 
 ## Installation
 
@@ -29,8 +25,7 @@ uv add opensandbox-code-interpreter
 
 ## Quick Start
 
-The following example demonstrates how to create a sandbox with a specific runtime configuration and execute a simple
-script.
+The following example demonstrates how to create a sandbox with a specific runtime configuration and execute a simple script.
 
 ```python
 import asyncio
@@ -137,13 +132,11 @@ execution = await sandbox.commands.run("pip install pandas numpy")
 
 ### Docker Image
 
-The Code Interpreter SDK relies on a specialized environment. Ensure your sandbox provider has the
-`opensandbox/code-interpreter` image available.
+The Code Interpreter SDK relies on a specialized environment. Ensure your sandbox provider has the `opensandbox/code-interpreter` image available.
 
 ### Language Version Selection
 
-You can specify the desired version of a programming language by setting the corresponding environment variable when
-creating the `Sandbox`.
+You can specify the desired version of a programming language by setting the corresponding environment variable when creating the `Sandbox`.
 
 | Language | Environment Variable | Example Value | Default (if unset) |
 | -------- | -------------------- | ------------- | ------------------ |
@@ -156,9 +149,7 @@ creating the `Sandbox`.
 
 ### 0. Run with `language` (default language context)
 
-You can pass `language` directly (recommended: `SupportedLanguage.*`) and skip `create_context`.
-When `context.id` is omitted, **execd will create/reuse a default session for that language**, so
-state can persist across runs:
+You can pass `language` directly (recommended: `SupportedLanguage.*`) and skip `create_context`. When `context.id` is omitted, **execd will create/reuse a default session for that language**, so state can persist across runs:
 
 ```python
 from code_interpreter import SupportedLanguage

{opensandbox_code_interpreter-1.1.1 → opensandbox_code_interpreter-1.2.1}/pyproject.toml

@@ -43,7 +43,7 @@ classifiers = [
 ]
 dependencies = [
     "pydantic>=2.4.2,<3.0",
-    "opensandbox>=1.1.1,<2.0.0",
+    "opensandbox>=1.2.1,<2.0.0",
 ]
 
 [project.urls]

{opensandbox_code_interpreter-1.1.1 → opensandbox_code_interpreter-1.2.1}/src/code_interpreter/adapters/code_adapter.py

@@ -33,15 +33,15 @@ from opensandbox.adapters.converter.execution_event_dispatcher import (
     ExecutionEventDispatcher,
 )
 from opensandbox.adapters.converter.response_handler import (
-    extract_request_id,
+    build_api_exception_from_httpx,
     handle_api_error,
     require_parsed,
 )
 from opensandbox.config import ConnectionConfig
-from opensandbox.exceptions import InvalidArgumentException, SandboxApiException
+from opensandbox.exceptions import InvalidArgumentException
 from opensandbox.models.execd import Execution, ExecutionHandlers
 from opensandbox.models.sandboxes import SandboxEndpoint
-from opensandbox.transport import unwrap_retry_transport
+from opensandbox.transport import streaming_transport
 
 from code_interpreter.adapters.converter.code_execution_converter import (
     CodeExecutionConverter,
@@ -77,7 +77,6 @@ class CodesAdapter(Codes):
     """
 
     RUN_CODE_PATH = "/code"
-    CREATE_CONTEXT_PATH = "/code/context"
 
     def __init__(
         self, execd_endpoint: SandboxEndpoint, connection_config: ConnectionConfig
@@ -128,9 +127,8 @@ class CodesAdapter(Codes):
             "Accept": "text/event-stream",
             "Cache-Control": "no-cache",
         }
-        # SSE bootstraps bypass the retry wrapper: request bodies are
-        # not replayable and a non-idempotent status opt-in would cause
-        # duplicate execution on a resent SSE POST.
+        # SSE bootstraps keep only the wake-timeout retry: a non-idempotent
+        # status opt-in would cause duplicate execution on a resent SSE POST.
         sse_transport = (
             self.connection_config.new_owned_transport()
             if self._owns_clients
@@ -144,7 +142,7 @@ class CodesAdapter(Codes):
                 write=timeout_seconds,
                 pool=None,
             ),
-            transport=unwrap_retry_transport(sse_transport),
+            transport=streaming_transport(sse_transport),
         )
 
     async def _get_client(self):
@@ -316,6 +314,7 @@ class CodesAdapter(Codes):
                 result=[],
                 error=None,
             )
+            execution._is_code_execution = True
 
             # Use SSE client for streaming responses (read timeout disabled)
             client = await self._get_sse_client()
@@ -324,15 +323,10 @@ class CodesAdapter(Codes):
             async with client.stream("POST", url, json=api_request) as response:
                 if response.status_code != 200:
                     await response.aread()
-                    error_body = response.text
                     logger.error(
-                        f"Failed to run code. Status: {response.status_code}, Body: {error_body}"
-                    )
-                    raise SandboxApiException(
-                        message=f"Failed to run code. Status code: {response.status_code}, Body: {error_body}",
-                        status_code=response.status_code,
-                        request_id=extract_request_id(response.headers),
+                        f"Failed to run code. Status: {response.status_code}, Body: {response.text}"
                     )
+                    raise build_api_exception_from_httpx(response, "Run code")
 
                 dispatcher = ExecutionEventDispatcher(execution, handlers)
 

{opensandbox_code_interpreter-1.1.1 → opensandbox_code_interpreter-1.2.1}/src/code_interpreter/sync/adapters/code_adapter.py

@@ -27,19 +27,19 @@ from opensandbox.adapters.converter.exception_converter import (
     ExceptionConverter,
 )
 from opensandbox.adapters.converter.response_handler import (
-    extract_request_id,
+    build_api_exception_from_httpx,
     handle_api_error,
     require_parsed,
 )
 from opensandbox.config.connection_sync import ConnectionConfigSync
-from opensandbox.exceptions import InvalidArgumentException, SandboxApiException
+from opensandbox.exceptions import InvalidArgumentException
 from opensandbox.models.execd import Execution
 from opensandbox.models.execd_sync import ExecutionHandlersSync
 from opensandbox.models.sandboxes import SandboxEndpoint
 from opensandbox.sync.adapters.converter.execution_event_dispatcher import (
     ExecutionEventDispatcherSync,
 )
-from opensandbox.transport import unwrap_retry_transport
+from opensandbox.transport import streaming_transport
 
 from code_interpreter.models.code_sync import CodeContextSync, SupportedLanguageSync
 from code_interpreter.sync.services.code import CodesSync
@@ -80,7 +80,6 @@ class CodesAdapterSync(CodesSync):
     """
 
     RUN_CODE_PATH = "/code"
-    CREATE_CONTEXT_PATH = "/code/context"
 
     def __init__(
         self, execd_endpoint: SandboxEndpoint, connection_config: ConnectionConfigSync
@@ -123,9 +122,8 @@ class CodesAdapterSync(CodesSync):
             "Accept": "text/event-stream",
             "Cache-Control": "no-cache",
         }
-        # SSE bootstraps bypass the retry wrapper: request bodies are
-        # not replayable and a non-idempotent status opt-in would cause
-        # duplicate execution on a resent SSE POST.
+        # SSE bootstraps keep only the wake-timeout retry: a non-idempotent
+        # status opt-in would cause duplicate execution on a resent SSE POST.
         sse_transport = (
             self.connection_config.new_owned_transport()
             if self._owns_clients
@@ -136,7 +134,7 @@ class CodesAdapterSync(CodesSync):
             timeout=httpx.Timeout(
                 connect=timeout_seconds, read=None, write=timeout_seconds, pool=None
             ),
-            transport=unwrap_retry_transport(sse_transport),
+            transport=streaming_transport(sse_transport),
         )
 
     def _get_execd_url(self, path: str) -> str:
@@ -310,17 +308,13 @@ class CodesAdapterSync(CodesSync):
 
             url = self._get_execd_url(self.RUN_CODE_PATH)
             execution = Execution(id=None, execution_count=None, result=[], error=None)
+            execution._is_code_execution = True
             dispatcher = ExecutionEventDispatcherSync(execution, handlers)
 
             with self._sse_client.stream("POST", url, json=api_request) as response:
                 if response.status_code != 200:
                     response.read()
-                    error_body = response.text
-                    raise SandboxApiException(
-                        message=f"Failed to run code. Status code: {response.status_code}, Body: {error_body}",
-                        status_code=response.status_code,
-                        request_id=extract_request_id(response.headers),
-                    )
+                    raise build_api_exception_from_httpx(response, "Run code")
 
                 for line in response.iter_lines():
                     if not line or not line.strip():