ICP4D Updates for Python development guide

Developing IBM Streams Applications with Python (Versions 1.6+)

Proposed additions in bold

Completed additions in strikethrough

IBM Streams Python Support

1. Installing Python APIs
2. Developing for the IBM Streaming Analytics service
   • Developing for ICP4Data
3. Developing with an IBM Streams installation
4. Common Streams operations
   • 4.1 Connecting to known data sources
   • 4.x Aggregation
     • Creating a window using last() or batch()
     • Simple aggregation using Window.aggregate (use pandas)
       • Partitioned windows
   • 4.x New filter option for splitting streams in matching and non-matching streams
   • 4.x New split function for splitting streams
   • 4.x Data model - namedtuples vs streams schema etc
   • 4.x Using views
     • Using views in a notebook
5. API features: Scalablity, fault tolerance
   • UDP (parallel fns)
   • Fault tolerance (consistent regions)
6. The Python REST API
   • connect to CP4D examples
7. Using SPL Operators in Python
   • Using Aggregate, Functor, Deduplicate etc. from streamsx.standard
   • Invoke SPL operators: reference Streams-SPLToolkitsTutorial.ipynb from sample.starter_notebooks
8. AI - CP4D services
   • Connecting to Watson Machine learning (streamsx.wml)
   • Connecting to Speech to Text Gateway (streamsx.sttgateway)
   • Scoring a model in Python
9. Improving performance - best practice
   • Reduce latency using low_latency
   • Using isolate
   • resource tags?
   • Saving credentials in app configs - a best practice to make code reusable
10. Metrics/Logging