Add implementation for bloom filters (predicate pushdown)

[kiloOhm]

As described in the specification

With the changes in this PR you can now add column-specific settings for bloom filters:

new ParquetOptions {
    BloomFilterOptionsByColumn = new Dictionary<string, ParquetOptions.BloomFilterOptions>() {
        { field.Name, new ParquetOptions.BloomFilterOptions { 
            EnableBloomFilters = true,
            BloomFilterFpp = 0.01f,
        } }
    }
},

During writing, the bloom filters are constructed and written to the data column header

While reading, you can then check if a value is definitely not or maybe present in the rowgroup column:

var reader = new DataColumnReader(field, ms, chunk, stats, footer, parquetOptions: new ParquetOptions {
    BloomFilterOptionsByColumn = new Dictionary<string, ParquetOptions.BloomFilterOptions>() {
            { field.Name, new ParquetOptions.BloomFilterOptions { EnableBloomFilters = true } }
        }
});

Assert.True(reader.MightMatchEquals("contained"));