T1-T4: tests de resilience, SDK PyPI ready, worker docker-compose, fix compat tests

sdk/README.md

@@ -8,6 +8,22 @@ The official Python client for interacting with the Imago backend API.
 pip install imago-client
 ```
 
+For development:
+```bash
+cd sdk
+pip install -e ".[dev]"
+```
+
+## Publishing to PyPI
+
+```bash
+cd sdk
+pip install build twine
+python -m build
+twine check dist/*
+twine upload dist/*
+```
+
 ## Basic Usage
 
 ```python
@@ -25,7 +41,7 @@ async def main():
             stream = client.images.stream_pipeline(image.id, "your-api-key")
             async for event in stream.stream_events():
                 print(f"Event: {event}")
-        
+
         # Get processed data
         processed = await client.images.get(image.id)
         if processed.ai:
@@ -34,3 +50,49 @@ async def main():
 if __name__ == "__main__":
     asyncio.run(main())
 ```
+
+## API Reference
+
+### HubClient
+
+```python
+client = HubClient(
+    base_url="http://localhost:8000",  # Imago API URL
+    api_key="your-api-key",            # Imago API key
+    timeout=30.0,                      # Request timeout (seconds)
+)
+```
+
+### Images Resource
+
+```python
+# Upload
+image = await client.images.upload(data, filename, content_type="image/jpeg")
+
+# Get
+detail = await client.images.get(image_id)
+
+# List
+images = await client.images.list(page=1, page_size=20, tag="nature")
+
+# Delete
+await client.images.delete(image_id)
+
+# Reprocess
+await client.images.reprocess(image_id)
+
+# WebSocket pipeline stream
+stream = client.images.stream_pipeline(image_id, api_key)
+async for event in stream.stream_events():
+    print(event)
+```
+
+### AI Resource
+
+```python
+# Summarize URL
+result = await client.ai.summarize("https://example.com")
+
+# Draft task
+task = await client.ai.draft_task("Organiser les photos de vacances")
+```