to_nexus() does not emit compliant Nexus #1785
Description
As per the above, the Nexus string emitted by tskit is not standard Nexus and cannot be parsed by most software.
There are a two different problems.
(1) The TAXA block requires a DIMENSIONS NTAX=## statement.
(2) The list of taxon labels should be separated by SPACES, NOT by commas.
The TreeSequence.to_nexus() statement currently emits:
#NEXUS
BEGIN TAXA;
TAXLABELS tsk_0_1,tsk_1_1,tsk_2_1,tsk_3_1,tsk_4_0,tsk_5_0,tsk_6_0;
END;
BEGIN TREES;
TREE tree0.00000000000000_100.00000000000000 = ((tsk_1_1:0.07086621911734,tsk_2_1:0.07086621911734)tsk_4_0:4.21281381972469,(tsk_0_1:1.20610709140453,tsk_3_1:1.20610709140453)tsk_5_0:3.07757294743750)tsk_6_0;
END;
Multiple program fail to parse this, including PAUP, BEAST, and, in fact, most phylogenetic programs and processing libraries.
Adding a DIMENSIONS statement and separating the taxon labels using spaces makes this not only compliant with standard Nexus, but also readable by most Nexus-supporting phylogenetic programs (whatever else their idiosyncracies! :) )
#NEXUS
BEGIN TAXA;
DIMENSIONS NTAX=7;
TAXLABELS tsk_0_1 tsk_1_1 tsk_2_1 tsk_3_1 tsk_4_0 tsk_5_0 tsk_6_0;
END;
BEGIN TREES;
TREE tree0.00000000000000_100.00000000000000 = ((tsk_1_1:0.07086621911734,tsk_2_1:0.07086621911734)tsk_4_0:4.21281381972469,(tsk_0_1:1.20610709140453,tsk_3_1:1.20610709140453)tsk_5_0:3.07757294743750)tsk_6_0;
END;