Keep SEQid to a single field in the output tables

[alexlancaster]

I've rebased #41 against master.

Here is @ssiddhantsharma original comment:

Sequence names are copied verbatim from the FASTA header into the SEQid column of both output tables, which are tab-delimited. A tab inside a header therefore split SEQid across several fields and shifted every column after it. Nothing failed: the header row kept its 41 columns while affected data rows gained one field per tab, so anything reading the table by column index silently read values from the wrong columns. MitoCarta and several other curated FASTA distributions ship tab-separated headers, which is how this is usually met.

A related inconsistency came from fastareader.nextfasta(), which discarded the result of its trim(). The first record of a file takes its name from hasmorefastas(), which does trim, while later records took it from nextfasta(), which did not. A file whose headers carried trailing whitespace could therefore produce a different column count for its first record than for the rest.

Map tabs, newlines and other control characters in the name to spaces and trim it, so SEQid is always exactly one field, and assign the trimmed name in nextfasta(). Names without control characters or surrounding whitespace are unchanged, embedded spaces included, so the MOT3 golden output is byte-identical.

cli/tests/test_seqid_field.sh covers a tab inside a header, agreement between the tab- and space-separated forms of the same header, trailing tabs and spaces across a multi-record file, the per-residue table under -p, and that ordinary names are left alone. Five of its six checks fail without this change. Wired into cli.yml alongside the MOT3 and input-validation tests.

[alexlancaster]

hi @ssiddhantsharma looks like you deleted your plaac fork? are you OK with me still testing/merging this? we really appreciate your contributions!