feat: stream large submission exports directly from db without memory…

[AbuJulaybeeb]

closes #1971.

This PR addresses memory spikes caused by generating large submission exports for quests with thousands of participants by transitioning entirely away from offset-based, chunked data queries and removing memory buffering in our streaming service.

Changes:

True DB Streaming Cursors: Refactored getQueryIterator in StreamExportService to utilize TypeORM's native .stream() method. Previously, getQueryIterator achieved "chunking" via .skip(offset).take(chunkSize), which executes repeated offset queries causing degraded SQL performance for deep records and buffering arrays in memory.
Implement Network Backpressure: Rewrote streamAsCSV, streamAsJSONLines, and streamAsJSON in StreamExportService to evaluate the boolean return value from response.write(). If the underlying TCP buffer backs up (returns false), the stream now pauses via await new Promise(resolve => response.once('drain', resolve)) until the buffer clears, capping concurrency safely.
Testing & Benchmarks: Updated stream-export.service.spec.ts mocks to handle the new once and backpressure assertions. A new benchmark script (BackEnd/scripts/benchmark-export.ts) was added to demonstrate stable memory (less than a few MBs of variance) regardless of how large the requested dataset is.

[drips-wave]

@AbuJulaybeeb Great news! 🎉 Based on an automated assessment of this PR, the linked Wave issue(s) no longer count against your application limits.

You can now already apply to more issues while waiting for a review of this PR. Keep up the great work! 🚀

Learn more about application limits

[RUKAYAT-CODER]

Kindly resolve conflict

[RUKAYAT-CODER]

Thank you for contributing to the project.