Replace unwrap calls with filter_map

[jmagnuson]

Fixes an issue where accessing cgroups for ephemeral systemd services results in a panic.

Dec 03 21:11:35 user systemd[1]: Started Prometheus PSI Exporter.
Dec 03 21:11:35 user psi_exporter[1087]: Listening address: 127.0.0.1:9093
Dec 04 01:45:09 user psi_exporter[1087]: thread 'main' panicked at src/main.rs:187:45:
Dec 04 01:45:09 user psi_exporter[1087]: called `Result::unwrap()` on an `Err` value: Error { depth: 2, inner: Io { path: Some("/sys/fs/cgroup/system.slice/a-timer-service.service"), err: Os { code: 2, kind: NotFound, message…e or directory" } } }
Dec 04 01:45:09 user psi_exporter[1087]: note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
Dec 04 01:45:09 user systemd[1]: psi-exporter.service: Main process exited, code=exited, status=101/n/a
Dec 04 01:45:09 user systemd[1]: psi-exporter.service: Failed with result 'exit-code'.
Dec 04 01:45:09 user systemd[1]: psi-exporter.service: Consumed 3.543s CPU time.