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