Parse edits

chempy/util/parsing.py

@@ -52,17 +52,17 @@ def _get_formula_parser():
 
     BNF for simple chemical formula (no nesting)
 
-        integer :: '0'..'9'+
+        real :: '0'..'9'+
         element :: 'A'..'Z' 'a'..'z'*
-        term :: element [integer]
+        term :: element [real]
         formula :: term+
 
 
     BNF for nested chemical formula
 
-        integer :: '0'..'9'+
+        real :: '0'..'9'+
         element :: 'A'..'Z' 'a'..'z'*
-        term :: (element | '(' formula ')') [integer]
+        term :: (element | '(' formula ')') [real]
         formula :: term+
 
     Notes
@@ -80,14 +80,14 @@ def _get_formula_parser():
     _p = __import__(parsing_library)
     Forward, Group, OneOrMore = _p.Forward, _p.Group, _p.OneOrMore
     Optional, ParseResults, Regex = _p.Optional, _p.ParseResults, _p.Regex
-    Suppress, Word, nums = _p.Suppress, _p.Word, _p.nums
+    Suppress, Word, nums, Combine = _p.Suppress, _p.Word, _p.nums, _p.Combine
 
     LPAR, RPAR = map(Suppress, "()")
-    integer = Word(nums)
+    real = Combine(Word(nums) + Optional('.' + Word(nums)))
 
-    # add parse action to convert integers to ints, to support doing addition
+    # add parse action to convert reals to floats, to support doing addition
     # and multiplication at parse time
-    integer.setParseAction(lambda t: int(t[0]))
+    real.setParseAction(lambda t: float(t[0]))
 
     # element = Word(alphas.upper(), alphas.lower())
     # or if you want to be more specific, use this Regex
@@ -101,7 +101,7 @@ def _get_formula_parser():
     formula = Forward()
 
     term = Group((element | Group(LPAR + formula + RPAR)("subgroup")) +
-                 Optional(integer, default=1)("mult"))
+                 Optional(real, default=1)("mult"))
 
     # add parse actions for parse-time processing
 
@@ -205,6 +205,7 @@ def _formula_to_parts(formula, prefixes, suffixes):
 def _parse_stoich(stoich):
     if stoich == 'e':  # special case, the electron is not an element
         return {}
+
     return {symbols.index(k)+1: n for k, n
             in _get_formula_parser().parseString(stoich)}
 
@@ -218,19 +219,22 @@ def _parse_stoich(stoich):
 _latex_mapping = {k + '-': '\\' + k + '-' for k in _greek_letters}
 _latex_mapping['epsilon-'] = '\\varepsilon-'
 _latex_mapping['omicron-'] = 'o-'
-_latex_mapping['.'] = '^\\bullet '
-_latex_infix_mapping = {'.': '\\cdot '}
+_latex_mapping['.'] = '.'
+_latex_mapping[':'] = '\\mathpunct{:} '
+_latex_infix_mapping = {':': '\\mathpunct{:} '}
 
 _unicode_mapping = {k + '-': v + '-' for k, v in zip(_greek_letters, _greek_u)}
 _unicode_mapping['.'] = u'⋅'
-_unicode_infix_mapping = {'.': u'·'}
+_unicode_mapping[':'] = u':'
+_unicode_infix_mapping = {':': u':'}
 
 _html_mapping = {k + '-': '&' + k + ';-' for k in _greek_letters}
 _html_mapping['.'] = '⋅'
+_html_mapping[':'] = ':'
 _html_infix_mapping = _html_mapping
 
 
-def _get_leading_integer(s):
+def _get_leading_coeff(s):
     m = re.findall(r'^\d+', s)
     if len(m) == 0:
         m = 1
@@ -254,30 +258,30 @@ def formula_to_composition(formula, prefixes=None,
     formula: str
         Chemical formula, e.g. 'H2O', 'Fe+3', 'Cl-'
     prefixes: iterable strings
-        Prefixes to ignore, e.g. ('.', 'alpha-')
+        Prefixes to ignore, e.g. ( 'alpha-')
     suffixes: tuple of strings
         Suffixes to ignore, e.g. ('(g)', '(s)')
 
     Examples
     --------
     >>> formula_to_composition('NH4+') == {0: 1, 1: 4, 7: 1}
     True
-    >>> formula_to_composition('.NHO-(aq)') == {0: -1, 1: 1, 7: 1, 8: 1}
+    >>> formula_to_composition(':NHO-(aq)') == {0: -1, 1: 1, 7: 1, 8: 1}
     True
-    >>> formula_to_composition('Na2CO3.7H2O') == {11: 2, 6: 1, 8: 10, 1: 14}
+    >>> formula_to_composition('Na2CO3:7H2O') == {11: 2, 6: 1, 8: 10, 1: 14}
     True
 
     """
     if prefixes is None:
         prefixes = _latex_mapping.keys()
     stoich_tok, chg_tok = _formula_to_parts(formula, prefixes, suffixes)[:2]
     tot_comp = {}
-    parts = stoich_tok.split('.')
+    parts = stoich_tok.split(':')
     for idx, stoich in enumerate(parts):
         if idx == 0:
             m = 1
         else:
-            m, stoich = _get_leading_integer(stoich)
+            m, stoich = _get_leading_coeff(stoich)
         comp = _parse_stoich(stoich)
         for k, v in comp.items():
             if k not in tot_comp:
@@ -321,7 +325,7 @@ def _parse_multiplicity(strings, substance_keys=None):
         elif len(items) == 2:
             if items[1] not in result:
                 result[items[1]] = 0
-            result[items[1]] += float(items[0]) if '.' in items[0] or 'e' in items[0] else int(items[0])
+            result[items[1]] += float(items[0]) if ':' in items[0] or 'e' in items[0] else int(items[0])
         else:
             raise ValueError("To many parts in substring")
     if substance_keys is not None:
@@ -398,14 +402,14 @@ def to_reaction(line, substance_keys, token, Cls, globals_=None, **kwargs):
 def _formula_to_format(sub, sup, formula, prefixes=None,
                        infixes=None, suffixes=('(s)', '(l)', '(g)', '(aq)')):
     parts = _formula_to_parts(formula, prefixes.keys(), suffixes)
-    stoichs = parts[0].split('.')
+    stoichs = parts[0].split(':')
     string = ''
     for idx, stoich in enumerate(stoichs):
         if idx == 0:
             m = 1
         else:
-            m, stoich = _get_leading_integer(stoich)
-            string += _subs('.', infixes)
+            m, stoich = _get_leading_coeff(stoich)
+            string += _subs(':', infixes)
         if m != 1:
             string += str(m)
         string += re.sub(r'([0-9]+)', lambda m: sub(m.group(1)), stoich)
@@ -445,8 +449,8 @@ def formula_to_latex(formula, prefixes=None, infixes=None, **kwargs):
     'Fe(CN)_{6}^{2+}'
     >>> formula_to_latex('Fe(CN)6+2(aq)')
     'Fe(CN)_{6}^{2+}(aq)'
-    >>> formula_to_latex('.NHO-(aq)')
-    '^\\bullet NHO^{-}(aq)'
+    >>> formula_to_latex(':NHO-(aq)')
+    '\\mathpunct{:} NHO^{-}(aq)'
     >>> formula_to_latex('alpha-FeOOH(s)')
     '\\alpha-FeOOH(s)'
 

chempy/util/tests/test_parsing.py

@@ -37,12 +37,12 @@ def test_formula_to_composition():
     assert formula_to_composition('SO4-2(aq)') == {0: -2, 8: 4, 16: 1}
 
     # prefixes and suffixes
-    assert formula_to_composition('.NO2(g)') == {7: 1, 8: 2}
-    assert formula_to_composition('.NH2') == {1: 2, 7: 1}
+    assert formula_to_composition(':NO2(g)') == {7: 1, 8: 2}
+    assert formula_to_composition(':NH2') == {1: 2, 7: 1}
     assert formula_to_composition('ONOOH') == {1: 1, 7: 1, 8: 3}
-    assert formula_to_composition('.ONOO') == {7: 1, 8: 3}
-    assert formula_to_composition('.NO3/2-') == {0: -2, 7: 1, 8: 3}
-    assert formula_to_composition('.NO3-2') == {0: -2, 7: 1, 8: 3}
+    assert formula_to_composition(':ONOO') == {7: 1, 8: 3}
+    assert formula_to_composition(':NO3/2-') == {0: -2, 7: 1, 8: 3}
+    assert formula_to_composition(':NO3-2') == {0: -2, 7: 1, 8: 3}
 
     with pytest.raises(ValueError):
         formula_to_composition('F-F')
@@ -51,7 +51,7 @@ def test_formula_to_composition():
     assert formula_to_composition('epsilon-Zn(OH)2(s)') == {1: 2, 8: 2, 30: 1}
 
     # crystal water
-    assert formula_to_composition('Na2CO3.7H2O(s)') == {11: 2, 6: 1, 8: 10, 1: 14}
+    assert formula_to_composition('Na2CO3:7H2O(s)') == {11: 2, 6: 1, 8: 10, 1: 14}
 
 
 @requires(parsing_library)
@@ -123,17 +123,17 @@ def test_formula_to_latex():
     assert formula_to_latex('NaCl(s)') == 'NaCl(s)'
     assert formula_to_latex('e-(aq)') == 'e^{-}(aq)'
     assert formula_to_latex('Ca+2(aq)') == 'Ca^{2+}(aq)'
-    assert formula_to_latex('.NO2(g)') == r'^\bullet NO_{2}(g)'
-    assert formula_to_latex('.NH2') == r'^\bullet NH_{2}'
+    assert formula_to_latex(':NO2(g)') == r'\mathpunct{:} NO_{2}(g)'
+    assert formula_to_latex(':NH2') == r'\mathpunct{:} NH_{2}'
     assert formula_to_latex('ONOOH') == 'ONOOH'
-    assert formula_to_latex('.ONOO') == r'^\bullet ONOO'
-    assert formula_to_latex('.NO3/2-') == r'^\bullet NO_{3}^{2-}'
-    assert formula_to_latex('.NO3-2') == r'^\bullet NO_{3}^{2-}'
+    assert formula_to_latex(':ONOO') == r'\mathpunct{:} ONOO'
+    assert formula_to_latex(':NO3/2-') == r'\mathpunct{:} NO_{3}^{2-}'
+    assert formula_to_latex(':NO3-2') == r'\mathpunct{:} NO_{3}^{2-}'
     assert formula_to_latex('alpha-FeOOH(s)') == r'\alpha-FeOOH(s)'
     assert formula_to_latex('epsilon-Zn(OH)2(s)') == (
         r'\varepsilon-Zn(OH)_{2}(s)')
-    assert formula_to_latex('Na2CO3.7H2O(s)') == r'Na_{2}CO_{3}\cdot 7H_{2}O(s)'
-    assert formula_to_latex('Na2CO3.1H2O(s)') == r'Na_{2}CO_{3}\cdot H_{2}O(s)'
+    assert formula_to_latex('Na2CO3:7H2O(s)') == r'Na_{2}CO_{3}\mathpunct{:} 7H_{2}O(s)'
+    assert formula_to_latex('Na2CO3:1H2O(s)') == r'Na_{2}CO_{3}\mathpunct{:} H_{2}O(s)'
 
 
 @requires(parsing_library)
@@ -151,16 +151,16 @@ def test_formula_to_unicoce():
     assert formula_to_unicode('NaCl(s)') == u'NaCl(s)'
     assert formula_to_unicode('e-(aq)') == u'e⁻(aq)'
     assert formula_to_unicode('Ca+2(aq)') == u'Ca²⁺(aq)'
-    assert formula_to_unicode('.NO2(g)') == u'⋅NO₂(g)'
-    assert formula_to_unicode('.NH2') == u'⋅NH₂'
+    assert formula_to_unicode(':NO2(g)') == u':NO₂(g)'
+    assert formula_to_unicode(':NH2') == u':NH₂'
     assert formula_to_unicode('ONOOH') == u'ONOOH'
-    assert formula_to_unicode('.ONOO') == u'⋅ONOO'
-    assert formula_to_unicode('.NO3/2-') == u'⋅NO₃²⁻'
-    assert formula_to_unicode('.NO3-2') == u'⋅NO₃²⁻'
+    assert formula_to_unicode(':ONOO') == u':ONOO'
+    assert formula_to_unicode(':NO3/2-') == u':NO₃²⁻'
+    assert formula_to_unicode(':NO3-2') == u':NO₃²⁻'
     assert formula_to_unicode('alpha-FeOOH(s)') == u'α-FeOOH(s)'
     assert formula_to_unicode('epsilon-Zn(OH)2(s)') == u'ε-Zn(OH)₂(s)'
-    assert formula_to_unicode('Na2CO3.7H2O(s)') == u'Na₂CO₃·7H₂O(s)'
-    assert formula_to_unicode('Na2CO3.1H2O(s)') == u'Na₂CO₃·H₂O(s)'
+    assert formula_to_unicode('Na2CO3:7H2O(s)') == u'Na₂CO₃:7H₂O(s)'
+    assert formula_to_unicode('Na2CO3:1H2O(s)') == u'Na₂CO₃:H₂O(s)'
 
 
 @requires(parsing_library)
@@ -177,14 +177,14 @@ def test_formula_to_html():
     assert formula_to_html('NaCl(s)') == 'NaCl(s)'
     assert formula_to_html('e-(aq)') == 'e<sup>-</sup>(aq)'
     assert formula_to_html('Ca+2(aq)') == 'Ca<sup>2+</sup>(aq)'
-    assert formula_to_html('.NO2(g)') == r'&sdot;NO<sub>2</sub>(g)'
-    assert formula_to_html('.NH2') == r'&sdot;NH<sub>2</sub>'
+    assert formula_to_html(':NO2(g)') == r'&#58;NO<sub>2</sub>(g)'
+    assert formula_to_html(':NH2') == r'&#58;NH<sub>2</sub>'
     assert formula_to_html('ONOOH') == 'ONOOH'
-    assert formula_to_html('.ONOO') == r'&sdot;ONOO'
-    assert formula_to_html('.NO3/2-') == r'&sdot;NO<sub>3</sub><sup>2-</sup>'
-    assert formula_to_html('.NO3-2') == r'&sdot;NO<sub>3</sub><sup>2-</sup>'
+    assert formula_to_html(':ONOO') == r'&#58;ONOO'
+    assert formula_to_html(':NO3/2-') == r'&#58;NO<sub>3</sub><sup>2-</sup>'
+    assert formula_to_html(':NO3-2') == r'&#58;NO<sub>3</sub><sup>2-</sup>'
     assert formula_to_html('alpha-FeOOH(s)') == r'&alpha;-FeOOH(s)'
     assert formula_to_html('epsilon-Zn(OH)2(s)') == (
         r'&epsilon;-Zn(OH)<sub>2</sub>(s)')
-    assert formula_to_html('Na2CO3.7H2O(s)') == 'Na<sub>2</sub>CO<sub>3</sub>&sdot;7H<sub>2</sub>O(s)'
-    assert formula_to_html('Na2CO3.1H2O(s)') == 'Na<sub>2</sub>CO<sub>3</sub>&sdot;H<sub>2</sub>O(s)'
+    assert formula_to_html('Na2CO3:7H2O(s)') == 'Na<sub>2</sub>CO<sub>3</sub>&#58;7H<sub>2</sub>O(s)'
+    assert formula_to_html('Na2CO3:1H2O(s)') == 'Na<sub>2</sub>CO<sub>3</sub>&#58;H<sub>2</sub>O(s)'