plint

commit dbd295c4969b36ec4e14c3282bc0d83c72110405
parent b793213e9354089bd56be0daa31d42bf4a267cc2
Author: Antoine Amarilli <a3nm@a3nm.net>
Date:   Tue,  2 Aug 2016 09:50:40 +0200

Merge branch 'master' of a3nm.net:git/plint

Diffstat:
TODO
 | 
4
++++
common.py
 | 
2
++
verse.py
 | 
4
++--
versetest.py
 | 
17
+++++++++++++++++

4 files changed, 25 insertions(+), 2 deletions(-)
diff --git a/TODO b/TODO
@@ -8,6 +8,10 @@ Concluera l'examen. Venez, je vous invite !
 /usr/lib/python3.5/re.py:203: FutureWarning: split() requires a non-empty pattern match.
 
 Remerciez-le plutôt, et félicitez-vous
+
+- fuzz testing
+- log exceptions separately for plint_web
+
 - faire marcher oublieux
 - change additions to give numerical syllable count
 - document expected errors of test.sh and commit necessary files
diff --git a/common.py b/common.py
@@ -52,6 +52,8 @@ def rm_punct(text, rm_all=False, rm_apostrophe=False, rm_apostrophe_end=True):
   if rm_apostrophe_end:
     text = re.sub("'*$", "", text) # apostrophes at end of line
   text = re.sub("[‒–—―⁓⸺⸻]", " ", text) # no weird dashes
+  text = re.sub("^--*\s", " ", text) # no isolated dashes
+  text = re.sub("--*\s", " ", text) # no trailing dashes
 
   #TODO rather: keep only good chars
   if not rm_all:
diff --git a/verse.py b/verse.py
@@ -351,12 +351,12 @@ class Verse:
     cs = re.split(self.hyphen_regexp, word)
     miss = ""
     for i in range(len(cs)):
-      if re.match("^-*$", cs[i]):
+      if re.match("^-*$", cs[i]) or re.match("^ *$", cs[i]):
         if len(pre_chunks2) > 0:
           pre_chunks2[-1] = (pre_chunks2[-1][0], pre_chunks2[-1][1] + cs[i])
           continue
         else:
-          miss = cs[i]
+          miss += cs[i]
           continue
       if is_consonants(normalize(cs[i])):
         pre_chunks2.append((False if i < len(cs) - 1 else True, miss + cs[i]))
diff --git a/versetest.py b/versetest.py
@@ -1,5 +1,6 @@
 #!/usr/bin/python3
 
+import common
 import template
 import diaeresis
 import common
@@ -26,6 +27,22 @@ class SanityCheck(unittest.TestCase):
     v.parse()
     self.assertEqual(text, v.line)
 
+  def testLeadingSpaceHyphenVowel(self):
+    text = " -a"
+    v = verse.Verse(text, template.Template(), template.Pattern("12"))
+    v.parse()
+    self.assertEqual(text, v.line)
+
+  def testLeadingSpaceHyphenConsonant(self):
+    text = " -c"
+    v = verse.Verse(text, template.Template(), template.Pattern("12"))
+    v.parse()
+    self.assertEqual(text, v.line)
+
+  def testLoneHyphens(self):
+    text = " - - -- --   - -  - --"
+    self.assertEqual(common.normalize(text), "")
+
 class Eliminate(unittest.TestCase):
   def testEliminateOneGue(self):
     text = "gue"