- Add reads_count=None to all GenotypedContig instantiations in test cases
- Update expected CSV output to include reads_count column
- Ensure all 406 contig stitcher tests pass with new field

- Add optional reads_count field to track read counts per contig
- Implement proportional distribution when splitting contigs in cut_query()
- Set reads_count to None when combining contigs in munge()
- Update AlignedContig.make() to preserve reads_count field
- Initialize reads_count=None in fasta_to_csv and read_contigs()
- Add reads_count column to CSV output in write_contigs()
- Update plot_contigs.py to include reads_count when creating fake Contig objects

Keep reads_count as internal-only field by removing it from the CSV
output in write_contigs(). The field is still tracked internally for
proportional distribution during contig operations.

- Add read_remap_counts() to parse remap_counts.csv and extract read counts
- Update read_contigs() to accept contig_read_counts dict and populate reads_count
- Make remap_counts_csv a required parameter for contig_stitcher()
- Update sample.py to pass unstitched_remap_counts_csv to contig_stitcher
- Update main() CLI to require remap_counts argument
- Create test data file exact_parts_contigs_remap_counts.csv
- Update tests to pass remap_counts parameter
- Fix expected output file to match new CSV format without reads_count column

This enables the contig stitcher to track read counts from the initial
remap step (before stitching) for proper proportional distribution when
splitting contigs.