commonnexus 1.9.0__py2.py3-none-any.whl → 1.9.2__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__ = '1.9.0'
+__version__ = '1.9.2'

commonnexus/blocks/characters.py

@@ -549,10 +549,15 @@ class Charstatelabels(Payload):
                 if isinstance(w, Token) and w.text == '/':
                     in_states = True
                     continue
+                if name:
+                    raise ValueError(
+                        'Illegal token in charstatelabel: "{}{}"'.format(name, w))
                 name = w
             except StopIteration:
                 break
         if num:
+            if name and name in names:
+                duplicate_charlabel(name, 'CHARSTATELABELS', nexus)
             self.characters.append(types.SimpleNamespace(number=num, name=name, states=states))
         elif comma:  # There was a comma, but no new label.
             warnings.warn('Trailing comma in CHARSTATELABELS command')

commonnexus/nexus.py

@@ -5,6 +5,7 @@ import collections
 import dataclasses
 
 from .tokenizer import TokenType, iter_tokens, get_name
+from .util import log_or_raise
 from commonnexus.command import Command
 from commonnexus.blocks import Block
 
@@ -45,7 +46,7 @@ class Config:
 
 class Nexus(list):
     """
-    A NEXUS object implemented as list of tokens with methods to access newick constituents.
+    A NEXUS object implemented as list of commands with methods to read and write blocks.
 
     From the spec:
 
@@ -95,6 +96,7 @@ class Nexus(list):
         """
         self.cfg = config or Config(**kw)
         self.trailing_whitespace = []
+        self.leading = []
         self.block_implementations = {}
         for cls in Block.__subclasses__():
             self.block_implementations[cls.__name__.upper()] = cls
@@ -120,7 +122,10 @@ class Nexus(list):
                     if token.is_semicolon:
                         commands.append(Command(tuple(tokens)))
                         tokens = []
-            self.trailing_whitespace = tokens
+            if commands:
+                self.trailing_whitespace = tokens
+            else:
+                self.leading = tokens
             s = commands
         list.__init__(self, s)
 
@@ -211,6 +216,7 @@ class Nexus(list):
             END;
         """
         return NEXUS \
+            + ''.join(str(t) for t in self.leading) \
             + ''.join(''.join(str(t) for t in cmd) for cmd in self) \
             + ''.join(str(t) for t in self.trailing_whitespace)
 
@@ -224,8 +230,10 @@ class Nexus(list):
         p.write_text(text, encoding=self.cfg.encoding)
 
     def iter_comments(self):
+        yield from (t for t in self.leading if t.type == TokenType.COMMENT)
         for cmd in self:
             yield from (t for t in cmd if t.type == TokenType.COMMENT)
+        yield from (t for t in self.trailing_whitespace if t.type == TokenType.COMMENT)
 
     @property
     def comments(self) -> typing.List[str]:
@@ -262,6 +270,8 @@ class Nexus(list):
 
     def validate(self, log=None):
         valid = True
+        if any(t.type not in {TokenType.WHITESPACE, TokenType.COMMENT} for t in self.leading):
+            log_or_raise('Invalid token in preamble', log=log)
         for block in self.iter_blocks():
             #
             # FIXME: we can do a lot of validation here! If block.__commands__ is a list, there is
@@ -269,6 +279,9 @@ class Nexus(list):
             # If Payload.__multivalued__ == False, only one command instance is allowed, ...
             #
             valid = valid and block.validate(log=log)
+        if any(t.type not in {TokenType.WHITESPACE, TokenType.COMMENT}
+               for t in self.trailing_whitespace):
+            log_or_raise('Invalid token in text after the last command', log=log)
         return valid
 
     def get_numbers(self, object_name, items):

commonnexus/tokenizer.py

@@ -11,7 +11,7 @@ Punctuation
 
     The following punctuation marks have special properties: [ ] do not break a
     word; + and - are allowed as state symbols, but none of the rest are allowed; - is
-    considered punctuation except were it is the minus sign in a negative number.
+    considered punctuation except where it is the minus sign in a negative number.
 """
 import enum
 import itertools

{commonnexus-1.9.0.dist-info → commonnexus-1.9.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: commonnexus
-Version: 1.9.0
+Version: 1.9.2
 Summary: A nexus (phylogenetics) file reader and writer (.nex, .trees)
 Home-page: https://github.com/dlce-eva/commonnexus
 Author: Robert Forkel
@@ -99,14 +99,15 @@ and writing NEXUS
 >>> print(Nexus.from_blocks(Data.from_data(nex.CHARACTERS.get_matrix())))
 #NEXUS
 BEGIN DATA;
-	DIMENSIONS NCHAR=10;
-	FORMAT DATATYPE=STANDARD MISSING=? GAP=- SYMBOLS="01";
-	MATRIX 
-    t1 1001010000
-    t2 0101000100
-    t3 0011101010
-    t4 0001100001
-    t5 0001100001;
+DIMENSIONS NCHAR=10;
+FORMAT DATATYPE=STANDARD MISSING=? GAP=- SYMBOLS="01";
+MATRIX 
+t1 1001010000
+t2 0101000100
+t3 0011101010
+t4 0001100001
+t5 0001100001
+;
 END;
 ```
 

{commonnexus-1.9.0.dist-info → commonnexus-1.9.2.dist-info}/RECORD

@@ -1,14 +1,14 @@
-commonnexus/__init__.py,sha256=WHq2NYClXqtJOmfAtyBC1X55uxYBcTcbRg17ywBEpKk,121
+commonnexus/__init__.py,sha256=c8gR9z5hNsSngL9DzrmgqmzpJYXivE4XNKRmbA8Z94M,121
 commonnexus/__main__.py,sha256=k07DsoCHCRw2Wi0HzXCmvNCSt8cddDEGCBlnHZorm6k,4361
 commonnexus/cli_util.py,sha256=pmqd3Fs2Wef7hRf0Z6koZ1Bg6-16FZUMOYZ8cMTY5Ig,3710
 commonnexus/command.py,sha256=5ZbW3wQY1GO6fChK7aqOzwjKBRxkPlKdUf09JbYEy9c,4284
-commonnexus/nexus.py,sha256=00pUVi1sBaVViI0r_P8YUfEeHw0h_10_sSGeIWn905o,17661
-commonnexus/tokenizer.py,sha256=8G2nP-xNN1gBd2a_7qxGDhHloCzoH4YE063p01FLTuk,11459
+commonnexus/nexus.py,sha256=rsfxX_cM4aOxS5p-lu39U2ZEcHswDoXPGGiVBiiVAMA,18380
+commonnexus/tokenizer.py,sha256=a_zcTLhEc2KW2KCbVFXF-KoZlJg5TTfN9PIh7qnnIuk,11460
 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=-sMfVczf7gWQ-hc9eV2pOgDgZUs9osrbDu_LbTPTgGM,8562
-commonnexus/blocks/characters.py,sha256=PvNHn11UdwCSNpAWZo_McowAzLlCqVhhxGRUF08T6Xs,59877
+commonnexus/blocks/characters.py,sha256=5zvGPUi6J9ES9hc-1flKb_OJzkhdqXn_1lbL_QbSfag,60130
 commonnexus/blocks/codons.py,sha256=Gox96r1Y3y8wmJIP1KhchfAepMC8BZB4Y7O2BTOcybc,7365
 commonnexus/blocks/distances.py,sha256=E1XnkOhn4NKwhKQS3APhcKJ6HR6B33O2q6rdzQ0n4wU,16018
 commonnexus/blocks/notes.py,sha256=6YctkCTgGAtOWKQ08LgzKXW_X0E1Oe5OgkqcF6iR5LY,8502
@@ -28,9 +28,9 @@ commonnexus/tools/__init__.py,sha256=X8dv4VrjAo5C8VJdUXMrKviDLUrwXbawE5SJiQ--2uQ
 commonnexus/tools/combine.py,sha256=5vPQKcqA2acFji2lZc_xNZr1Xm2F3TPmfQk-fa9GA60,3048
 commonnexus/tools/matrix.py,sha256=3e2n_beug3MZyzXDgawjxxh1QfiM7KVp757OepCjQR0,11492
 commonnexus/tools/normalise.py,sha256=RdAasznob1aIY4B5z0Ir3jm9T5Yj_nbFzdtFtUM553U,6131
-commonnexus-1.9.0.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-commonnexus-1.9.0.dist-info/METADATA,sha256=mNQkc5C10gAI0q9LRfJOsYIBFnZvQFhXzBPsBbyxT6k,5379
-commonnexus-1.9.0.dist-info/WHEEL,sha256=bb2Ot9scclHKMOLDEHY6B2sicWOgugjFKaJsT7vwMQo,110
-commonnexus-1.9.0.dist-info/entry_points.txt,sha256=0CbsOaqe6jaHnG1rpheuIISCZvFp1G1oBgDym82rpAo,58
-commonnexus-1.9.0.dist-info/top_level.txt,sha256=SiWuQrmGciG5Ivhrqz6ueU1Sxv-fzudflsMof81-I54,12
-commonnexus-1.9.0.dist-info/RECORD,,
+commonnexus-1.9.2.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+commonnexus-1.9.2.dist-info/METADATA,sha256=VLz6kYpG539mRNneegc18DiE-bq_7x9LyM6BSzUUEbI,5357
+commonnexus-1.9.2.dist-info/WHEEL,sha256=bb2Ot9scclHKMOLDEHY6B2sicWOgugjFKaJsT7vwMQo,110
+commonnexus-1.9.2.dist-info/entry_points.txt,sha256=0CbsOaqe6jaHnG1rpheuIISCZvFp1G1oBgDym82rpAo,58
+commonnexus-1.9.2.dist-info/top_level.txt,sha256=SiWuQrmGciG5Ivhrqz6ueU1Sxv-fzudflsMof81-I54,12
+commonnexus-1.9.2.dist-info/RECORD,,