Zero trained m-values can lead to math domain error

[ADBond]

Given the right circumstances in the data, trained m-values can end up as 0 (probably "really" just smaller than floating point precision). This leads to math domain error when we try to take the log of the (also 0) Bayes factor, in ComparisonLevel._as_detailed_record, in e.g. charts such as match_weights_chart or m_u_parameters_chart.

Here is a not-very-elegant reprex:

import splink.comparison_library as cl
from splink import DuckDBAPI, Linker, SettingsCreator, block_on, splink_datasets

settings = SettingsCreator(
    "dedupe_only",
    comparisons=[
        cl.LevenshteinAtThresholds("first_name"),
        cl.LevenshteinAtThresholds("surname"),
        cl.ExactMatch("city"),
        cl.LevenshteinAtThresholds("dob"),
        cl.LevenshteinAtThresholds("email"),
        cl.ExactMatch("cluster"),
        cl.ExactMatch("cluster_1"),
        cl.ExactMatch("cluster_2"),
        cl.ExactMatch("cluster_3"),
        cl.ExactMatch("non_match_cat"),
    ],
    blocking_rules_to_generate_predictions=[
        block_on("dob"),
        block_on("city"),
    ]
)

df = splink_datasets.fake_1000
df["cluster_1"] = df["cluster"]
df["cluster_2"] = df["cluster"]
df["cluster_3"] = df["cluster"]

# specially chosen non-matchy things
df["non_match_cat"] = None
cats = {
    9: 1,
    192: 1,
    10: 1,
    7: 1,
    21: 2,
    287: 2,
    263: 6,
    273: 6,
    500: 7,
    729: 7,
}
for id_n, cat in cats.items():
    df["non_match_cat"][df["unique_id"] == id_n] = cat

db_api = DuckDBAPI()
linker = Linker(df, settings, db_api)

linker.training.estimate_probability_two_random_records_match(
    block_on("first_name", "surname", "dob"), recall=0.7
)
linker.training.estimate_u_using_random_sampling(max_pairs=1e8)

linker.training.estimate_parameters_using_expectation_maximisation(block_on("cluster"))
linker.training.estimate_parameters_using_expectation_maximisation(block_on("dob"))
linker.training.estimate_parameters_using_expectation_maximisation(block_on("city"))

linker.misc.save_model_to_json("mde.json", overwrite=True)

ch = linker.visualisations.match_weights_chart()

Related is #1889, which has the same root cause, and can be reproduced if in the above we call linker.inference.predict() instead of the match weights chart. Opening this separately though, as that issue occurs during SQL execution, while here we hit the issue in python, and so these will potentially require different solutions.