Metadata-Version: 2.4
Name: areteos
Version: 0.2.0
Summary: Python SDK for the Areteos Workflow Runtime API
Project-URL: Homepage, https://github.com/aretecp/areteos
Project-URL: Documentation, https://github.com/aretecp/areteos/tree/main/sdks/python
Project-URL: Repository, https://github.com/aretecp/areteos
Author-email: Arete <dev@arete.io>
License-Expression: MIT
Classifier: Development Status :: 3 - Alpha
Classifier: Intended Audience :: Developers
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Typing :: Typed
Requires-Python: >=3.11
Requires-Dist: httpx<1.0,>=0.27
Requires-Dist: pydantic<3.0,>=2.7
Provides-Extra: pydantic-ai
Requires-Dist: pydantic-ai<2.0,>=1.0; extra == 'pydantic-ai'
Description-Content-Type: text/markdown

# Areteos Python SDK

Python client for the [Areteos Workflow Runtime API](../../docs/guides/workflow-runtime-api.md). Define workflows, trigger runs, and monitor execution with full type safety.

## Installation

```bash
pip install areteos
```

## Quick Start

```python
from areteos import AreteosClient, Step, Edge, InputMapping

client = AreteosClient(
    base_url="https://areteos.example.com",
    api_key="areteos_your_key",
    engagement_id="your-engagement-uuid",
)

# Register a workflow
workflow = client.create_workflow(
    "Document Pipeline",
    steps=[
        Step.python("extract", "Extract Text", code="""
import json, sys
data = json.load(open(sys.argv[1])) if len(sys.argv) > 1 else {}
print(json.dumps({"text": data.get("raw_text", "")}))
"""),
        Step.agent(
            "summarize", "Summarize",
            system_prompt="Summarize the document concisely.",
            input_mappings=[
                InputMapping(source_step_id="extract", source_field="text", target_field="document")
            ],
        ),
    ],
    edges=[Edge(from_step_id="extract", to_step_id="summarize")],
)

# Trigger and wait
run = client.trigger(workflow.current_version_id, input={"raw_text": "..."})
result = client.wait_for_run(run.id, timeout=300)
print(result.output)
```

## Async Usage

```python
from areteos import AreteosAsyncClient, Step

async with AreteosAsyncClient(
    base_url="https://areteos.example.com",
    api_key="areteos_your_key",
    engagement_id="your-engagement-uuid",
) as client:
    workflow = await client.create_workflow("My Workflow", steps=[...])
    run = await client.trigger(workflow.current_version_id)
    result = await client.wait_for_run(run.id)
```

## Step Types

### Agent (LLM)

```python
Step.agent(
    "analyze", "Analyze",
    model="capable",           # fast, capable, thinking
    strategy="react",          # react, adaptive, cot, cod, tot, got, trm
    system_prompt="You are an analyst.",
    max_turns=10,
)
```

### Python (Inline Code)

```python
Step.python(
    "transform", "Transform",
    code="import json, sys\ndata = json.load(open(sys.argv[1]))\nprint(json.dumps({'count': len(data)}))",
    timeout_ms=30000,
)
```

### Human-in-the-Loop

```python
Step.hitl(
    "review", "Manager Review",
    request_type="approval",   # approval, input, review
    timeout_minutes=1440,      # 24 hours
)
```

## Data Flow

Connect steps with edges and input mappings:

```python
steps = [
    Step.python("extract", "Extract", code="..."),
    Step.agent(
        "summarize", "Summarize",
        input_mappings=[
            InputMapping(source_step_id="extract", source_field="text", target_field="document"),
            InputMapping(source_step_id=None, source_field="config", target_field="settings"),  # from run input
        ],
    ),
]
edges = [Edge(from_step_id="extract", to_step_id="summarize")]
```

## Error Handling

```python
from areteos import AreteosError, ApiError, ForbiddenError, NotFoundError, ValidationError, TimeoutError

try:
    run = client.trigger(version_id)
    result = client.wait_for_run(run.id, timeout=60)
except ValidationError as e:
    print(f"Invalid workflow: {e}")
except ForbiddenError:
    print("Access denied — check your API key scope")
except TimeoutError:
    print("Run did not complete in time")
except ApiError as e:
    print(f"API error {e.status_code}: {e}")
```

## Auto-Approve HITL

For CI/CD pipelines, auto-approve all human-in-the-loop requests:

```python
result = client.wait_for_run(run.id, auto_approve=True)
```

## Development

```bash
cd sdks/python
uv sync
uv run pytest
uv run ruff check .
uv run mypy src/
```
