commonnexus 1.4.0__py2.py3-none-any.whl → 1.6.0__py2.py3-none-any.whl

commonnexus/__init__.py

@@ -1,4 +1,4 @@
 from .nexus import Nexus, Config  # noqa: F401
 from commonnexus.blocks import Block  # noqa: F401
 
-__version__ = '1.4.0'
+__version__ = '1.6.0'

commonnexus/blocks/characters.py

@@ -1109,9 +1109,10 @@ class Characters(Block):
             DIMENSIONS NCHAR=3;
             FORMAT DATATYPE=STANDARD MISSING=? GAP=- SYMBOLS="01";
             MATRIX
-                t1 100
-                t2 010
-                t3 001;
+            t1 100
+            t2 010
+            t3 001
+            ;
             END;
 
         :param matrix: A matrix as returned by :meth:`Characters.get_matrix()`, with unlabeled \
@@ -1149,7 +1150,7 @@ class Characters(Block):
                     row.append('{{{}}}'.format(''.join(sorted(symbol(c) for c in entry))))
                 else:
                     row.append(symbol(entry))
-            rows.append('\n    {} {}'.format(tlabels[taxon].ljust(maxlen), ''.join(row)))
+            rows.append('\n{} {}'.format(tlabels[taxon].ljust(maxlen), ''.join(row)))
 
         symbols = ''.join(sorted([s for s in symbols if s not in [None, GAP]]))
         if missing in symbols or (gap in symbols):
@@ -1176,7 +1177,7 @@ class Characters(Block):
                 ) for n, l in charlabels.items())))
         if taxlabels:
             cmds.append(('TAXLABELS', ' '.join(tlabels.values())))
-        cmds.append(('MATRIX', ''.join(rows)))
+        cmds.append(('MATRIX', ''.join(rows) + '\n'))
         return cls.from_commands(cmds, nexus=nexus, TITLE=TITLE, ID=ID, LINK=LINK)
 
 

commonnexus/blocks/distances.py

@@ -383,5 +383,5 @@ class Distances(Block):
             row = [tlabels[taxon].ljust(maxlen)]
             row.extend(['?' if v is None else str(v) for v in dists.values()])
             mrows.append(' '.join(row))
-        cmds.append(('MATRIX', ''.join('\n    ' + row for row in mrows)))
+        cmds.append(('MATRIX', ''.join('\n' + row for row in mrows) + '\n'))
         return cls.from_commands(cmds, nexus=nexus, TITLE=TITLE, LINK=LINK, ID=ID)

commonnexus/command.py

@@ -27,7 +27,7 @@ class Command(tuple):
         tokens = []
         if comment:
             tokens.extend([Token('\n', TokenType.WHITESPACE), Token(comment, TokenType.COMMENT)])
-        tokens.append(Token('\n\t' if in_block else '\n', TokenType.WHITESPACE))
+        tokens.append(Token('\n', TokenType.WHITESPACE))
         name = list(iter_tokens(iter(name)))
         assert len(name) == 1 and name[0].type == TokenType.WORD
         tokens.extend(name)

commonnexus/tools/normalise.py

@@ -11,10 +11,15 @@ Normalisation includes
 
 In addition, after normalisation, the following assumptions hold:
 
-- All commands start on a new line, preceded by "\t" if the command is within a block.
+- All commands start on a new line.
 - All command names (**not** block names) are in uppercase with no "in-name-comment",
   like "MA[c]TRiX"
+- The ";" terminating MATRIX commands is on a separate line, allowing more simplistic parsing
+  of matrix rows.
 """
+import typing
+import collections
+
 from commonnexus import Nexus
 from commonnexus.blocks.characters import Data
 from commonnexus.blocks import Taxa, Distances, Characters, Trees
@@ -22,12 +27,14 @@ from commonnexus.blocks import Taxa, Distances, Characters, Trees
 
 def normalise(nexus: Nexus,
               data_to_characters: bool = False,
-              strip_comments: bool = False) -> Nexus:
+              strip_comments: bool = False,
+              remove_taxa: typing.Optional[typing.Container[str]] = None) -> Nexus:
     """
     :param nexus: A `Nexus` object to be normalised in-place.
     :param data_to_characters: Flag signaling whether DATA blocks should be converted to CHARACTER \
     blocks.
     :param strip_comments: Flag signaling whether to remove all non-command comments.
+    :param remove_taxa: Container of taxon labels specifying taxa to remove from relevant blocks.
     :return: The modified `Nexus` object.
 
     .. code-block:: python
@@ -67,22 +74,26 @@ def normalise(nexus: Nexus,
         DIMENSIONS NCHAR=3;
         FORMAT DATATYPE=STANDARD MISSING=? GAP=- SYMBOLS="01";
         MATRIX
-            t1 100
-            t2 010
-            t3 001;
+        t1 100
+        t2 010
+        t3 001
+        ;
         END;
         BEGIN DISTANCES;
         DIMENSIONS NTAX=3;
         FORMAT TRIANGLE=BOTH MISSING=?;
         MATRIX
-            t1 0 1.0 2.0
-            t2 1.0 0 3.0
-            t3 2.0 3.0 0;
+        t1 0 1.0 2.0
+        t2 1.0 0 3.0
+        t3 2.0 3.0 0
+        ;
         END;
         BEGIN TREES;
         TREE 1 = (t1,t2,t3);
         END;
     """
+    remove_taxa = remove_taxa or []
+
     if strip_comments:
         nexus = Nexus([cmd.without_comments() for cmd in nexus], config=nexus.cfg)
     nexus = Nexus([cmd.with_normalised_whitespace() for cmd in nexus], config=nexus.cfg)
@@ -91,6 +102,7 @@ def normalise(nexus: Nexus,
     if nexus.characters:
         matrix = nexus.characters.get_matrix()
         taxlabels = list(matrix.keys())
+        matrix = collections.OrderedDict((k, v) for k, v in matrix.items() if k not in remove_taxa)
         characters = nexus.DATA or nexus.CHARACTERS
         cls = Data if characters.name == 'DATA' and not data_to_characters else Characters
         nexus.replace_block(
@@ -103,19 +115,26 @@ def normalise(nexus: Nexus,
             assert set(matrix.keys()).issubset(taxlabels)
         else:
             taxlabels = list(matrix.keys())
+        matrix = collections.OrderedDict(
+            (k, collections.OrderedDict((kk, vv) for kk, vv in v.items() if kk not in remove_taxa))
+            for k, v in matrix.items() if k not in remove_taxa)
         nexus.replace_block(nexus.DISTANCES, Distances.from_data(matrix))
 
     if nexus.TREES:
         trees = []
         for tree in nexus.TREES.trees:
             nwk = nexus.TREES.translate(tree) if nexus.TREES.TRANSLATE else tree.newick
+            if remove_taxa:
+                nwk.prune_by_names(remove_taxa)
             trees.append((tree.name, nwk, tree.rooted))
         nexus.replace_block(nexus.TREES, Trees.from_data(*trees))
 
     if taxlabels:
+        taxa = Taxa.from_data([t for t in taxlabels if t not in remove_taxa])
         if nexus.TAXA:
             assert nexus.TAXA.DIMENSIONS.ntax == len(taxlabels)
             assert set(nexus.TAXA.TAXLABELS.labels.values()) == set(taxlabels)
+            nexus.replace_block(nexus.TAXA, taxa)
         else:
-            nexus.prepend_block(Taxa.from_data(taxlabels))
+            nexus.prepend_block(taxa)
     return nexus

{commonnexus-1.4.0.dist-info → commonnexus-1.6.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: commonnexus
-Version: 1.4.0
+Version: 1.6.0
 Summary: A nexus (phylogenetics) file reader and writer (.nex, .trees)
 Home-page: https://github.com/dlce-eva/commonnexus
 Author: Robert Forkel

{commonnexus-1.4.0.dist-info → commonnexus-1.6.0.dist-info}/RECORD

@@ -1,16 +1,16 @@
-commonnexus/__init__.py,sha256=5_gCJDmeiZzd1NGnKzB-ZV44Q-JyyUt-jizbL8V1kRI,121
+commonnexus/__init__.py,sha256=Tr84gK3MelAO7U__8dpEsLEXy_JmiX9NmCvX1wg8gzw,121
 commonnexus/__main__.py,sha256=k07DsoCHCRw2Wi0HzXCmvNCSt8cddDEGCBlnHZorm6k,4361
 commonnexus/cli_util.py,sha256=pmqd3Fs2Wef7hRf0Z6koZ1Bg6-16FZUMOYZ8cMTY5Ig,3710
-commonnexus/command.py,sha256=-qPv_WWrQEq_09lC4CmQP4kJHdYoj3zMnSNu7-clMMI,4308
+commonnexus/command.py,sha256=5ZbW3wQY1GO6fChK7aqOzwjKBRxkPlKdUf09JbYEy9c,4284
 commonnexus/nexus.py,sha256=VFuBOpujAWb551mMCvTmmMPd7A_ciL5Z8u9LyzQoIBE,17585
 commonnexus/tokenizer.py,sha256=8G2nP-xNN1gBd2a_7qxGDhHloCzoH4YE063p01FLTuk,11459
 commonnexus/util.py,sha256=WomyVUggYJON5T9uNuIZ_Bxm16UJHIhO737SjKcFYO0,168
 commonnexus/blocks/__init__.py,sha256=ehKEEc_E62zhlkvbp0p3asY56QkLuhqk7wYd1AqQuKI,602
 commonnexus/blocks/assumptions.py,sha256=SNURUIUNP86SgTTdfLVTnGQ0FJDnd3I6FTz3jT9A7aw,11215
 commonnexus/blocks/base.py,sha256=MzEkgBYl0dSIPGuFqZvdROGRr_gEz-HgHBOIQzBmGhI,7988
-commonnexus/blocks/characters.py,sha256=rP2bl_c_npvochu883-nVngEESyA8DI0AHb8qqGsGNo,59647
+commonnexus/blocks/characters.py,sha256=VK6T5enMJbG2n-oncbojRE3-q4oPcUmn3A5f_S7XIgQ,59651
 commonnexus/blocks/codons.py,sha256=Gox96r1Y3y8wmJIP1KhchfAepMC8BZB4Y7O2BTOcybc,7365
-commonnexus/blocks/distances.py,sha256=1Oeh2Ad8u2S2FZfupGCqOnmF4hd7iXx9WNGq8DcDHKk,15853
+commonnexus/blocks/distances.py,sha256=fRnV4BDFOL43x22sHIz7HjTwXdwWK9Tb4x8DLS92pLI,15856
 commonnexus/blocks/notes.py,sha256=Ukx2SNQYWaUSmFlCNE3MiiVpEQn_Bir1CPrYt92xE9M,8333
 commonnexus/blocks/sets.py,sha256=iLtOgDSAbgkh8JeVi7kdMK4lTuAZMXdZUaE7cU2Mrio,7699
 commonnexus/blocks/taxa.py,sha256=KU2GnZXBC3LGsJgzr9l48fi_Y-z740QWjr0rsOossWM,3040
@@ -27,10 +27,10 @@ commonnexus/commands/trees.py,sha256=aMN9WPvzOXfNgPcLvwmkJHtyvjq3LWWo2rpbGFCyczs
 commonnexus/tools/__init__.py,sha256=X8dv4VrjAo5C8VJdUXMrKviDLUrwXbawE5SJiQ--2uQ,213
 commonnexus/tools/combine.py,sha256=CZ0JsJa8azPv6XWuUAHLGOMI0N8EYPY11ipJfTCiRUs,3061
 commonnexus/tools/matrix.py,sha256=3e2n_beug3MZyzXDgawjxxh1QfiM7KVp757OepCjQR0,11492
-commonnexus/tools/normalise.py,sha256=G5Gm6a0f_GRXjU2uIwJqLPNHxfcfJfzmJTG_Wz2jGmk,4137
-commonnexus-1.4.0.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-commonnexus-1.4.0.dist-info/METADATA,sha256=vEJisLVvA43uIm0uI6x3orjsy1R9S99GK7URo38OPVQ,5328
-commonnexus-1.4.0.dist-info/WHEEL,sha256=bb2Ot9scclHKMOLDEHY6B2sicWOgugjFKaJsT7vwMQo,110
-commonnexus-1.4.0.dist-info/entry_points.txt,sha256=0CbsOaqe6jaHnG1rpheuIISCZvFp1G1oBgDym82rpAo,58
-commonnexus-1.4.0.dist-info/top_level.txt,sha256=SiWuQrmGciG5Ivhrqz6ueU1Sxv-fzudflsMof81-I54,12
-commonnexus-1.4.0.dist-info/RECORD,,
+commonnexus/tools/normalise.py,sha256=l7MNBkFQk0hyl2IA33AvRFS0mjAZwDYumpfw8pzzMa0,4924
+commonnexus-1.6.0.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+commonnexus-1.6.0.dist-info/METADATA,sha256=dRy1hWD10vl597dRpdzmI4OiSs0FkqE61KDC9ClZF5E,5328
+commonnexus-1.6.0.dist-info/WHEEL,sha256=bb2Ot9scclHKMOLDEHY6B2sicWOgugjFKaJsT7vwMQo,110
+commonnexus-1.6.0.dist-info/entry_points.txt,sha256=0CbsOaqe6jaHnG1rpheuIISCZvFp1G1oBgDym82rpAo,58
+commonnexus-1.6.0.dist-info/top_level.txt,sha256=SiWuQrmGciG5Ivhrqz6ueU1Sxv-fzudflsMof81-I54,12
+commonnexus-1.6.0.dist-info/RECORD,,