Memory Optimization Skill for Streamlit + Supabase Apps

Reduce Streamlit memory bloat by tuning cache parameters, downcasting DataFrame data types, and offloading heavy data filtering to Supabase to stay within Community Cloud resource limits. This directly addresses unbounded cache growth and excessive session state usage.

Streamlit apps crash loading large pandas DataFrames due to excessive memory copies and unbounded caching. Resolve this by loading only necessary columns, downcasting data types, and applying targeted cache parameter tuning to minimize the memory footprint.

Reduce Streamlit session state memory by leveraging Supabase for server-side processing. Use Supabase RPC calls and cursor-based pagination to handle data aggregation remotely, preventing large datasets from bloating the client app.

Optimize DataFrame memory in a Streamlit caching workflow by downcasting data types, loading only necessary columns, and avoiding redundant data copies. Tune caching parameters to prevent unbounded cache growth and limit session state usage.

Yes, this approach works without migrating data away from Supabase. It leverages existing Supabase infrastructure for server-side filtering, aggregation via functions, and cursor-based pagination to efficiently reduce client-side memory load.

Streamlit caching limitations include unbounded cache growth and excessive session state memory bloat if parameters are not tuned. Mitigate these constraints by offloading heavy data operations to Supabase RPC calls and optimizing pandas DataFrames.