Individuals not associated with nodes can lead to incorrect VCF

[jeromekelleher]

Related to #2446 and #2257

ts = msprime.sim_ancestry(3, sequence_length=1e2, random_seed=1234)
ts = msprime.sim_mutations(ts, rate=0.01, random_seed=1234)
tables = ts.dump_tables()
tables.individuals.add_row()
ts = tables.tree_sequence()
print(ts.as_vcf())

gives

##fileformat=VCFv4.2
##source=tskit 0.5.2.dev0
##FILTER=<ID=PASS,Description="All filters passed">
##contig=<ID=1,length=100>
##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">
#CHROM  POS     ID      REF     ALT     QUAL    FILTER  INFO    FORMAT  tsk_0   tsk_1   tsk_2tsk_3
1       2       0       C       T       .       PASS    .       GT      1|0     1|0     1|0
1       15      1       G       T       .       PASS    .       GT      1|0     0|0     0|0
1       19      2       A       T       .       PASS    .       GT      1|0     0|0     0|0
1       29      3       C       A       .       PASS    .       GT      1|0     1|0     1|0
1       35      4       T       C       .       PASS    .       GT      0|0     0|0     0|1
1       37      5       G       C       .       PASS    .       GT      0|1     0|1     0|0
1       56      6       C       G       .       PASS    .       GT      0|1     0|1     0|0
1       61      7       T       G       .       PASS    .       GT      0|0     0|0     1|0
1       71      8       C       G       .       PASS    .       GT      0|1     0|1     0|0
1       81      9       G       C       .       PASS    .       GT      1|0     0|0     0|0

Note the extra sample tsk_3 in the header.

The current code essentially assumes that all individuals are associated with sample nodes, and it is possible to come up with other examples where incorrect data is output without raising an error.