Teiphy

This release incorporates contributions from catsmith to ensure compatibility with Python versions 3.9 through 3.12. (As a result of these changes, Python 3.8 is no longer supported.) To accommodate software like PAUP* and edmondac's fork of MrBayes (https://github.com/edmondac/MrBayes), the symbol set for NEXUS outputs has been extended to 62 symbols (0-9, a-z, and A-Z). This release also adds support for the use of `--table distance` and `--table similarity` options (along with the `--proportion` and `--show-ext` flags) with outputs in PHYLIP (.phy and .ph) format to produce PHYLIP-formatted distance or similarity matrices.

This release introduces the `--table similarity` option, which produces a tabular output with counts of pairwise agreements between witnesses (or, if the `--proportion` flag is specified, proportions of agreements among variation units where both witnesses have non-ambiguous readings), as well as the `--show-ext` flag, which adds the number of variation units where both witnesses have non-ambiguous readings to each cell's value (e.g., 47/50 or 0.94/50). This option can also be used with distance matrices specified with `--table distance`.

With this release, you can exclude fragmentary witnesses from your collation by specifying the `--fragmentary-threshold` command-line option, followed by a number between 0 and 1 indicating the proportion of variation units at which a witness must be extant (i.e., have a non-missing reading according to the reading type(s) specified with the `-m` option) to be included in the output. Thus, `--fragmentary-threshold 0.7` will exclude all witnesses with more than 30 percent of their readings missing, while `--fragmentary-threshold 1.0` will exclude witnesses with any missing readings. (Note that this check is performed after correctors' hands have been filled in, if you have supplied the `--fill-correctors` option.)

In principle, any number of states should theoretically be permissible in BEAST 2.7 XML inputs, since the states are specified as sequences of probabilities rather than with one-character symbols. But even with sequences encoded in this way, BEAST 2 still requires code maps (for some reason), so we are limited by the space of allowable single-character symbols. Previously, teiphy restricted the set of BEAST state symbols to 0-9 and a-z. This release adds A-Z to the symbol set.

Previously, teiphy assumed that each variation unit (i.e., an `app` element) would be uniquely identified by its `xml:id` attribute or its `n` attribute alone. While this assumption holds in the case of `xml:id` attributes (which, by definition, must be unique), it does not hold for `n` attributes. In practice, TEI XML collations assign `app` elements in the same larger passage of text (e.g., a verse) the same `n` value as that larger passage and then assign the `app` elements additional `from` and `to` attributes specifying word indices, so as to specify the unique location of the variation unit within that larger passage. To this end, the `VariationUnit` class of teiphy now checks for `from` and `to` attributes in addition to an `n` attribute and combines them to form a unique ID for the variation unit.

This release provides a new feature for the convenience of users who have derived their collation data and witness date ranges from different sources: a CSV file containing witness IDs and (potentially empty) minimum and maximum dates can be specified with the `--dates-file` command-line option. For witnesses in the CSV file, the specified date range will overwrite any existing date range in the TEI XML collation.