drime

commit 867643e14e7f8f50d58a7d05cfd3fe6723641947
parent 4da377ba38cfb840d3c45f3e3bc50e711e214823
Author: Antoine Amarilli <a3nm@a3nm.net>
Date:   Thu, 29 Sep 2011 15:20:48 +0200

remove syllable count from source (too crappy)

Diffstat:
make_db.py
 | 
15
+++++++--------
make_db.sh
 | 
2
+-

2 files changed, 8 insertions(+), 9 deletions(-)
diff --git a/make_db.py b/make_db.py
@@ -89,13 +89,13 @@ class Word:
         return True
     return False
 
-  def __init__(self, word, phon, base, kind, freq, nsyl):
+  def __init__(self, word, phon, base, kind, freq):
     self.word = word
     self.phon = phon
     self.base = base
     self.kind = kind
     self.freq = freq
-    self.nsyl = [nsyl, nsyl]
+    self.nsyl = None
     self.redundant = False
     self.do_extends()
 
@@ -113,8 +113,10 @@ class Word:
       self.extend(self.align_sum(align[0]))
 
   def extend(self, item):
-    self.nsyl = [min(self.nsyl[0], item),
-        max(self.nsyl[1], item)]
+    if self.nsyl == None:
+      self.nsyl = [item, item]
+    else:
+      self.nsyl = [min(self.nsyl[0], item), max(self.nsyl[1], item)]
 
 bases = {} #TODO transitive to the topmost base
 
@@ -148,11 +150,8 @@ while True:
   freq = float(l.pop(0))
   #print ("DBG for %s: %d and %d" % (word, int(l[0]), 1+len([x for x in l[1]
     #if x == ' ' or x == '-'])))
-  nsyl = max(int(l[0]), 1+len([x for x in l[1] if x == ' ' or x == '-']))
-  l.pop(0)
-  l.pop(0)
   assert(len(l) == 0)
-  w = Word(word, phon, base, kind, freq, nsyl)
+  w = Word(word, phon, base, kind, freq)
   if word not in bases.keys():
     bases[word] = []
   bases[word].append(base)
diff --git a/make_db.sh b/make_db.sh
@@ -4,5 +4,5 @@ cd "$( dirname "$0" )"
 
 cat - additions | # add custom exceptions
   cut -f 1,2,3,4,7,8,9,10,24,28 | # select relevant fields
-  awk '{FS="	"; OFS="	"; print $1, $2, $3, $4, ($5+$6)/2 + 100*($7+$8)/2, $9, $10}' #| # aggregate frequencies
+  awk '{FS="	"; OFS="	"; print $1, $2, $3, $4, ($5+$6)/2 + 100*($7+$8)/2}' #| # aggregate frequencies
   #./make_db.py