feat(cubestore): Enable ZSTD compression for Parquet file

[kunalpatange]

Problem

Cube Store writes Parquet files without explicit compression, relying on datafusion's default behavior. This results in
larger file sizes, which impacts performance when:

• Transferring pre-aggregation results from remote storage (S3, GCS, etc.)
• Storing intermediate computation results locally

Changes Made

• Add .set_compression(datafusion::parquet::basic::Compression::ZSTD) to Parquet writer properties
• Single-line change in ParquetTableStore::writer_props() method
• No breaking changes or API modifications