[BUG] compressed event logs are not processed

[gerashegalov]

Describe the bug
The profiling tool path filter out file names containing a . . Thus lz4 event log files produced with the config spark.eventLog.compress=true which is standard for production deployments are skipped. The remainder of the logic relying on the SHS event log reader supports lz4 (default), snappy, zstd , lzf

Steps/Code to reproduce bug
Run some integration tests e.g.:

 SPARK_HOME=~/dist/spark-3.1.1-bin-hadoop3.2 \
 NUM_LOCAL_EXECS=0 \
./integration_tests/run_pyspark_from_build.sh -k large_orderby_nested

This will generate compressed event logs under integration_tests/target/run_dir/eventlog_gw0

Then run:

~/dist/spark-3.1.1-bin-hadoop3.2/bin/spark-submit --class com.nvidia.spark.rapids.tool.profiling.ProfileMain ~/gits/NVIDIA/spark-rapids/tools/target/rapids-4-spark-tools_2.12-21.08.0-SNAPSHOT.jar -o /tmp/prof1/ ~/gits/NVIDIA/spark-rapids/integration_tests/target/run_dir/eventlog_gw0

The tool silently quits: I think we need an actionable log message for legitimately empty input directories

Expected behavior
The tool should generate reports for SHS-supported compression codecs. I verified that it works if you modify the filter https://github.com/NVIDIA/spark-rapids/blame/3e303c0dae26d1a8ef01fa271ee5f0f43759e865/tools/src/main/scala/com/nvidia/spark/rapids/tool/profiling/ProfileMain.scala#L73

and

https://github.com/NVIDIA/spark-rapids/blame/3e303c0dae26d1a8ef01fa271ee5f0f43759e865/tools/src/main/scala/com/nvidia/spark/rapids/tool/profiling/ProfileMain.scala#L91

Environment details (please complete the following information)
local dev

Additional context
N/A

[gerashegalov]

Fixed but not auto-closed by #2626