`md_eval_22_multifile` emits a logging traceback for mismatched RTTM recording IDs

[alexanderwerning]

When reference and hypothesis RTTMs contain different recording IDs,
meeteval.der.md_eval.md_eval_22_multifile emits a logging error.

Current code:

keys = set(r.keys()) & set(h.keys())
missing = set(r.keys()) ^ set(h.keys())
if len(missing) > 0:
    logging.warning(f"Missing {len(missing)} filenames:", missing)
    logging.warning(f"Found {len(keys)} filenames:", keys)

Because the logging message does not contain a % placeholder, Python's
logging formatter fails when it receives missing or keys as the second
argument:

TypeError: not all arguments converted during string formatting
Message: 'Found 32 filenames:'
Arguments: ({...},)

Minimal reproduction:

from tempfile import NamedTemporaryFile
from meeteval.der.api import dscore

reference = "SPEAKER rec 1 0 1 <NA> <NA> speaker <NA> <NA>\n"
hypothesis = ""

with NamedTemporaryFile(mode="w", suffix=".rttm") as ref_file, \
     NamedTemporaryFile(mode="w", suffix=".rttm") as hyp_file:
    ref_file.write(reference)
    hyp_file.write(hypothesis)
    ref_file.flush()
    hyp_file.flush()
    dscore(reference=[ref_file.name], hypothesis=[hyp_file.name])

Expected behavior:

• Emit a normal warning, without a logging traceback.

[alexanderwerning]

Fixed in #131.