Allow users to specify string matching case sensitivity behavior via configurations

[knassre-bodo]

Currently, the default in sqlite is that string matching via LIKE is case insensitive, but users may want to make it case sensitive (and frankly, this makes sense as the default). To that end, the goal of this issue is to add a new config property to the PyDoughConfigs class: case_sensitive_string_matching (default=True). When True, LIKE and similar operators (STARTSWITH, ENDSWITH, CONTAINS are all case sensitive. For sqlite, this means that during conversion, the 2nd argument to like should be wrapped in COLLATE BINARY. For example, if I have PyDough code with a predicate X LIKE 'A%", that should become SELECT ... FROM ... WHERE X LIKE 'A%' COLLATE BINARY, which is the following tree expression in sqlglot:

Select(
  expressions=[...],
  from=From(...),
  where=Where(
    this=Like(
      this=Column(
        this=Identifier(this=A, quoted=False)),
      expression=Collate(
        this=Literal(this=A%, is_string=True),
        expression=Var(this=BINARY)))))

When the config is False, like & similar functions should go back to being case-insensitive. We can keep the COLLATE but switch it from BINARY to NOCASE, which is the default. For other dialects, we can think about how to handle the translation, b ut if a dialect's default is case-sensitive then all we need to do is call LOWER on both arguments when case-insensitivity is requested According to ChatGPT, the dialects have the following behavior:

• Case-Insensitive by Default: MySQL, MariaDB, SQL Server, SQLite (for ASCII), Snowflake, CockroachDB.
• Case-Sensitive by Default: PostgreSQL, Oracle, Db2, Firebird, Informix, Sybase ASE, Teradata, Redshift, BigQuery, ClickHouse, DuckDB.
• Configurable via Collations or Settings: MySQL (utf8_bin for case-sensitive), SQL Server (collations), SQLite (PRAGMA case_sensitive_like = ON;), Snowflake (COLLATE 'BINARY').

The behavior of the config should be well documented in the usage guide and various README files.