Optimize import memory and stabilize import logs - #24

Summary

• optimize import memory usage by streaming CSV batches and reusing offsets
• reduce dry-run log memory growth and transient payload size
• stabilize import log tab behavior and preserve final logs
• fix CSV quote parsing for backslash-escaped double-quote edge cases
• allow test release builds from the current branch

Validation

• dry run completed successfully on staging for a large CSV
• actual import progressed stably without memory spikes or related-site impact during observation
• UI log tab issues and CSV parsing edge case were verified during testing