- remove_children_if optimized

DHParser/transform.py

@@ -212,14 +212,13 @@ def traverse(root_node: Node,
         table = processing_table
         cache = processing_table['__cache__']
     else:
-        # normalize processing_table entries by turning single values into lists
-        # with a single value
+        # normalize processing_table entries by turning single values
+        # into lists with a single value
         table = {name: smart_list(call) for name, call in list(processing_table.items())}
         table = expand_table(table)
         cache = table.setdefault('__cache__', {})  # type: Dict[str, List[Callable]]
-        # change processing table in place, so that table expansion does not get lost
-        # between calls
-        processing_table.clear();
+        # change processing table in place, so its already expanded and cache filled next time
+        processing_table.clear()
         processing_table.update(table)
 
     # assert '__cache__' in processing_table
@@ -238,8 +237,9 @@ def traverse(root_node: Node,
                 context.pop()
 
         key = key_func(node)
-        sequence = cache.get(key, None)
-        if sequence is None:
+        try:
+            sequence = cache[key]
+        except KeyError:
             sequence = table.get('+', []) + \
                        table.get(key, table.get('*', [])) + \
                        table.get('~', [])
@@ -494,6 +494,14 @@ def keep_children(context: List[Node], section: slice = slice(None)):
 
 @transformation_factory(Callable)
 def remove_children_if(context: List[Node], condition: Callable, section: slice = slice(None)):
+    """Removes all children for which `condition()` returns `True`."""
+    node = context[-1]
+    if node.children:
+        node.result = tuple(c for c in node.children if not condition(context + [c]))
+
+
+@transformation_factory(Callable)
+def remove_children(context: List[Node], condition: Callable, section: slice = slice(None)):
     """Removes all nodes from a slice of the result field if the function
     `condition(child_node)` evaluates to `True`."""
     node = context[-1]

examples/LaTeX/LaTeX.ebnf

@@ -71,15 +71,16 @@ verbatim            = "\begin{verbatim}" sequence §"\end{verbatim}"
 tabular             = "\begin{tabular}" tabular_config { tabular_row } §"\end{tabular}"
 tabular_row         = (multicolumn | tabular_cell) { "&" (multicolumn | tabular_cell) }
                       "\\" ( hline | { cline } )
-tabular_cell        = { text_element //~ }
+tabular_cell        = { line_element //~ }
 tabular_config      = "{" /[lcr|]+/~ §"}"
 
 #### paragraphs and sequences of paragraphs ####
 
 block_of_paragraphs = "{" [sequence] §"}"
 sequence            = { (paragraph | block_environment ) [PARSEP] }+
-paragraph           = { !blockcmd (text_element | LINEFEED) //~ }+
-text_element        = text | block | inline_environment | command
+paragraph           = { !blockcmd text_element //~ }+
+text_element        = line_element | LINEFEED
+line_element        = text | block | inline_environment | command
 
 #### inline enivronments ####
 

examples/LaTeX/LaTeXCompiler.py

@@ -120,15 +120,16 @@ class LaTeXGrammar(Grammar):
     tabular             = "\begin{tabular}" tabular_config { tabular_row } §"\end{tabular}"
     tabular_row         = (multicolumn | tabular_cell) { "&" (multicolumn | tabular_cell) }
                           "\\" ( hline | { cline } )
-    tabular_cell        = { text_element //~ }
+    tabular_cell        = { line_element //~ }
     tabular_config      = "{" /[lcr|]+/~ §"}"
     
     #### paragraphs and sequences of paragraphs ####
     
     block_of_paragraphs = "{" [sequence] §"}"
     sequence            = { (paragraph | block_environment ) [PARSEP] }+
-    paragraph           = { !blockcmd (text_element | LINEFEED) //~ }+
-    text_element        = text | block | inline_environment | command
+    paragraph           = { !blockcmd text_element //~ }+
+    text_element        = line_element | LINEFEED
+    line_element        = text | block | inline_environment | command
     
     #### inline enivronments ####
     
@@ -219,7 +220,7 @@ class LaTeXGrammar(Grammar):
     paragraph = Forward()
     tabular_config = Forward()
     text_element = Forward()
-    source_hash__ = "e493869bdc02eb835ec3ce1ebfb0a4ea"
+    source_hash__ = "1d9bad5194b49edf88a447f370541ed1"
     parser_initialization__ = "upon instantiation"
     COMMENT__ = r'%.*(?:\n|$)'
     WSP__ = mixin_comment(whitespace=r'[ \t]*(?:\n(?![ \t]*\n)[ \t]*)?', comment=r'%.*(?:\n|$)')
@@ -267,12 +268,13 @@ class LaTeXGrammar(Grammar):
     generic_inline_env = Series(begin_inline_env, RE(''), paragraph, Required(end_inline_env))
     known_inline_env = Synonym(inline_math)
     inline_environment = Alternative(known_inline_env, generic_inline_env)
-    text_element.set(Alternative(text, block, inline_environment, command))
-    paragraph.set(OneOrMore(Series(NegativeLookahead(blockcmd), Alternative(text_element, LINEFEED), RE(''))))
+    line_element = Alternative(text, block, inline_environment, command)
+    text_element.set(Alternative(line_element, LINEFEED))
+    paragraph.set(OneOrMore(Series(NegativeLookahead(blockcmd), text_element, RE(''))))
     sequence = OneOrMore(Series(Alternative(paragraph, block_environment), Optional(PARSEP)))
     block_of_paragraphs.set(Series(Token("{"), Optional(sequence), Required(Token("}"))))
     tabular_config.set(Series(Token("{"), RE('[lcr|]+'), Required(Token("}"))))
-    tabular_cell = ZeroOrMore(Series(text_element, RE('')))
+    tabular_cell = ZeroOrMore(Series(line_element, RE('')))
     tabular_row = Series(Alternative(multicolumn, tabular_cell),
                          ZeroOrMore(Series(Token("&"), Alternative(multicolumn, tabular_cell))),
                          Token("\\\\"), Alternative(hline, ZeroOrMore(cline)))

examples/LaTeX/tst_LaTeX_docs.py

@@ -19,14 +19,16 @@ See the License for the specific language governing permissions and
 limitations under the License.
 """
 
+import cProfile as profile
 import os
+import pstats
 import sys
 
+sys.path.extend(['../../', '../', './'])
+
 import DHParser.dsl
 from DHParser import toolkit
 
-sys.path.extend(['../../', '../', './'])
-
 if not DHParser.dsl.recompile_grammar('LaTeX.ebnf', force=False):  # recompiles Grammar only if it has changed
     print('\nErrors while recompiling "LaTeX.ebnf":\n--------------------------------------\n\n')
     with open('LaTeX_ebnf_ERRORS.txt') as f:
@@ -46,18 +48,24 @@ def fail_on_error(src, result):
             print(e)
         sys.exit(1)
 
-with toolkit.logging(True):
+
+with toolkit.logging(False):
     files = os.listdir('testdata')
     files.sort()
+    pr = profile.Profile()
+    pr.enable()
     for file in files:
-        if file.lower().endswith('.tex'):
+        if file.lower().endswith('.tex') and file.lower().find('error') < 0:
             with open(os.path.join('testdata', file), 'r') as f:
                 doc = f.read()
             print('\n\nParsing document: "%s"\n' % file)
             result = parser(doc)
             parser.log_parsing_history__()
             fail_on_error(doc, result)
-            ast = transformer(result)
-            fail_on_error(doc, ast)
-            print(ast.as_sxpr())
-
+            transformer(result)
+            fail_on_error(doc, result)
+            print(result.as_sxpr())
+    pr.disable()
+    st = pstats.Stats(pr)
+    st.strip_dirs()
+    st.sort_stats('time').print_stats(20)

examples/LaTeX/tst_LaTeX_grammar.py

@@ -21,10 +21,9 @@ limitations under the License.
 
 import sys
 
-import DHParser.dsl
-
 sys.path.extend(['../../', '../', './'])
 
+import DHParser.dsl
 from DHParser import testing
 from DHParser import toolkit