feat: upgrade manual interpolation pipeline to generate perfectly projected PNGs in-memory

fix: restore manual AI frame rendering logic by decoupling array-to-png process

fix: resolve missing area definition crash in SatPy by correctly applying monkey patch

fix: resolve AI frame projection stretching by inheriting raw Geostationary Scene during prebake

fix: resolve AI frame misalignment by completely bypassing pyresample geostationary unrolling

fix: resolve map squishing by falling back to INSAT full disk bounding box and patching area extraction

fix: restore missing hard-reset endpoint after file overwrite

fix: resolve AI frame rendering crash by squeezing extra time dimension from SatPy CF writer

fix: resolve AI frame parsing crash by updating NetCDF parser identifier

fix: resolve SatPy regex parser crash by preserving original filename for local caching

fix: resolve stale SSE state by forcing fresh load from bucket

fix: resolve redundant bucket downloads during sequence interpolation

fix: resolve asyncio event loop conflict in scheduler and disk cache absolute path crash

feat: add hard-reset endpoint to clear hf bucket

feat: implement pre-baked PNG ETL pipeline to optimize memory and cloud storage

feat: implement SSE streaming and physical bucket deletion for scheduler

fix: change default MOSDAC dataset ID to 3SIMG_L1B_STD

feat: complete automated mosdac to ai animation pipeline

feat(fullstack): enhance visual clarity and add restart workflows

feat(fullstack): add interactive colorful maps and heat gradients

fix(backend): correct scientific evaluation logic to prevent inflated accuracy metrics

feat(fullstack): final production-readiness overhaul

refactor(backend): purge obsolete validation service to resolve architectural flaws

perf(backend): massive optimization of visualization mapping pipeline

fix(backend): correct cloud mask overlay logic

feat(fullstack): add upload progress tracking and interpolated files list

fix(backend): use unified file resolver for metadata extraction

fix(backend): align metrics API response payload with frontend expectations

fix(backend): remove india geographical hard-locks on AI files

chore(backend): add pyresample explicitly to dependencies

Merge frontend changes into backend-fixes branch

feat(backend): organize huggingface bucket folders and add geographical reprojection

fix(backend): optimize visualization service for AI netcdf and legacy h5

chore: Sync backend from HF

Merge branch 'siddhant385:testing' into testing

universal saver done

added isro parser

fixed some errors

improving h5 parser

improving h5 parser

worked on integration (partial errors)

improved ai file reading and resolving error

improved ai file reading and resolving error

improved export and visualization

improved get file path

improved visualization

improved visualization

recitfied visualization

solving git conflicts

added sse events in fastapi