#2311 feat(loaders): add target-s3 (plinxore variant) Pull request opened by plinxore Summary Adds the

plinxore

variant of

target-s3

, a fork of `crowemi/target-s3` matured for production use: • Real gzip compression (the original wrote plaintext behind a misleading

.gz

extension) • Corrected JSONL datetime serialization • Configurable handling of source dates the SDK can't parse (e.g. legacy MySQL/MyISAM zero-dates) • A collision-safe per-batch filename scheme (monotonic counter + UUID) so concurrent or repeated runs can never silently overwrite each other's files • A Parquet writer with a schema kept stable across batches (derived once from the Singer SCHEMA rather than re-inferred per batch, so multi-file glob reads never hit a schema mismatch) and exact decimal precision for DECIMAL-shaped columns (

decimal128

, not

float64

) Package: `plinxore-target-s3` on PyPI (currently

0.1.2

). Repo: https://github.com/plinxore/target-s3 Settings were scraped directly from the published package's

--about --format json

output via

hub-utils

, and validated with

plugin_schema_validate.py

,

yaml_check_names.py

, and

maintainers_validate.py

. Test plan •

plugin_schema_validate.py

passes for all plugin definitions •

yaml_check_names.py

passes •

maintainers_validate.py

passes • Settings scraped from a real

pip install plinxore-target-s3==0.1.2

in a fresh venv, confirming

format_type

includes

parquet

and

format.format_parquet.get_schema_from_tap

reflects the real default/description meltano/hub