PCSM-201 Add some configurability to local environment scripts

[chupe]

The PR adds infrastructure improvements to support testing PCSM replication from a ReplicaSet source to a Sharded cluster target. This configuration validates that PCSM can replicate collections from RS topology to sharded topology.

• hack/rs/run.sh: Made replica set count configurable via RS_COUNT environment variable (default: 2, max: 2). Added dynamic service initialization based on count.

• hack/sh/run.sh: Added support for skipping source or target cluster creation via SRC_SHARDS=0 or TGT_SHARDS=0. This enables RS→Sharded testing by using hack/rs/run.sh for the source and hack/sh/run.sh with SRC_SHARDS=0 for the target.

• hack/cleanup.sh: Improved cleanup to remove networks and volumes created by both environments, ensuring clean teardown between test runs.

• tests/test_documents.py: Fixed typo in test name (trucated → truncated)

Testing:

RS→Sharded configuration validated locally with:

RS_COUNT=1 ./hack/rs/run.sh
SRC_SHARDS=0 TGT_SHARDS=2 ./hack/sh/run.sh

E2E test suite ran against MongoDB 6.0, 7.0, and 8.0.

[inelpandzic]

When I read, it is a bit strange:

RS_COUNT=1 ./hack/rs/run.sh
SRC_SHARDS=0 TGT_SHARDS=2 ./hack/sh/run.sh

RS_COUNT I read as number of replika sets, but it actually controls do we deploy source+target or just source target. The same thing goes for SRC_SHARDS=0 and TGT_SHARDS=2.

Since this task was just one-time testing that we support RS->SH sync, we don't have to think about improving this at the time. I'll close this one and when we start working on fully supporting RS-SH sync, we can address this and improve our scripts.