declension table correction facility. See #11

nominals/pysanskritv2/tables/correction_inventory.txt

@@ -0,0 +1 @@
+corrections/corrections_j.txt

nominals/pysanskritv2/tables/corrections.py

@@ -0,0 +1,72 @@
+""" corrections.py
+"""
+import sys,re,codecs
+import tableread
+
+def add_manual_tables(d,filein):
+ """ modify d
+ """
+ recs = tableread.init_table(filein)
+ for rec in recs:
+  # a Table object
+  key = (rec.model,rec.key2)
+  # Do not allow possibility of duplicate 'keys'
+  #if key in d:
+  # print('add_manual_tables WARNING: duplicate key',key,"in file",filein)
+  #else:
+  # d[key] = rec.table
+
+  # Allow possibility of duplicate 'keys'
+  if key not in d:
+   d[key] = []
+  else:
+   print('manual duplicate key',key,filein)
+  #d[key].append(rec.table)
+  d[key] = rec
+
+def init_manual_tables(filein):
+ with codecs.open(filein,"r","utf-8") as f:
+  filenames = [x.rstrip() for x in f if not x.startswith(';')]
+ d = {}
+ for filename in filenames:
+  add_manual_tables(d,filename)
+ return d
+
+def process(line,d,iline):
+ """  If line has a correction in d, change and return new line
+   otherwise, return line.
+   line is assumed to be in format written by decline_file
+     '%s\t%s\t%s\t%s' %(rec.model,rec.key2,rec.refs,rec.inflection)
+ """
+ try:
+  model,key2,refs,inflection = line.split('\t')
+ except:
+  print('process ERROR. wrong format of old line #',iline+1)
+  print(line)
+  exit(1)
+ dkey = (model,key2)  # consistent with add_manual_tables
+ if dkey not in d:
+  return line
+ rec = d[dkey]
+ new_inflection = rec.tabstring
+ newline = '%s\t%s\t%s\t%s' % (model,key2,refs,new_inflection)
+ rec.nused = rec.nused + 1
+ print('new inflection for',dkey)
+ return newline
+if __name__ == "__main__":
+ filein = sys.argv[1]   # tables produced by decline_file
+ filein1 = sys.argv[2]  # inventory of correction files
+ fileout = sys.argv[3]  # output tables, same format as filein
+
+ d = init_manual_tables(filein1)
+ with codecs.open(filein,"r","utf-8") as f:
+  with codecs.open(fileout,"w","utf-8") as fout:
+   for iline,line in enumerate(f):
+    line = line.rstrip()
+    lineout = process(line,d,iline)
+    fout.write(lineout+'\n')
+ # check unused tables
+ dupkeys = [k for k in d.keys() if d[k].nused == 0]
+ print(len(d.keys()),' = # of correction records')
+ print(len(dupkeys),'= # of unused corrections')
+

nominals/pysanskritv2/tables/corrections/corrections_j.txt

@@ -0,0 +1,40 @@
+; Corrections for declensions of words ending in 'j'
+;Declension of f_1_j sfj C=sfk, H=sfw
+;Declension of m_1_j sfj C=sfk, H=sfw
+;Declension of n_1_j sfj C=sfk=H 
+; MW viSva—sfj (nom. -sfk or incorrectly -sfw
+; Goldman, p. 378  -sfw
+; Follow Goldman 10-24-2020
+Declension of m_1_j viSva-sfj
+Case 1:  viSvasfw viSvasfjO viSvasfjaH
+Case 2:  viSvasfjam viSvasfjO viSvasfjaH
+Case 3:  viSvasfjA viSvasfqByAm viSvasfqBiH
+Case 4:  viSvasfje viSvasfqByAm viSvasfqByaH
+Case 5:  viSvasfjaH viSvasfqByAm viSvasfqByaH
+Case 6:  viSvasfjaH viSvasfjoH viSvasfjAm
+Case 7:  viSvasfji viSvasfjoH viSvasfwsu
+Case 8:  viSvasfk viSvasfjO viSvasfjaH
+; Declension of m_1_j aDi-rAj  [H 3 stems]
+; Declension of m_1_j vEdya-rAj  [Huet 2 stems]
+; Declension of f_1_j asfj asfk
+; Declension of m_1_j asfj asfk
+; Declension of n_1_j asfj asfk
+; Declension of f_1_j KaYj ejf error? (MW,PWG Kan)
+; Declension of m_1_j KaYj
+; Declension of n_1_j KaYj
+; Declension of f_1_j naBrAj  (C=naBrAk, H=naBrAw=PW
+; Declension of m_1_j naBrAj
+; Declension of n_1_j naBrAj
+; Declension of m_1_j mftaBraj C=mftaBrak, H=mftaBraw
+; Declension of f_1_j mftaBraj C=mftaBrak=H ?
+; Declension of f_1_j parivrAj ?? PW vrAw = H
+; Declension of m_1_j parivrAj H has two stems
+; Declension of n_1_j pravrAj  k/w
+; Declension of f_1_j yuj [Huet 2 stems in 1,2,8]
+; Declension of m_1_j yuj [Huet 2 stems in 1,2,8]
+; Declension of n_1_j yuj   [Huet 2 stems in 1,2,8]
+; sraj MW  ifc. nom. sraw;   sraj.f. nom. srak
+; there are 30 compounds ending in -sraj in MW
+; Declension of f_1_j sraj Note H=sragByAm 
+; Declension of m_1_j sraj  C=sragByAm, H=sraqByAm etc.
+; Declension of n_1_j sraj  H=sragByAm

nominals/pysanskritv2/tables/redo.sh

@@ -1,2 +1,6 @@
-
-python3 decline_file.py ../stems/calc_stems.txt calc_tables.txt
+echo "REDO DECLENSION TABLES BEGINS"
+echo "calc_tables0"
+python3 decline_file.py ../stems/calc_stems.txt calc_tables0.txt
+echo "calc_tables (after corrections)"
+python3 corrections.py calc_tables0.txt correction_inventory.txt calc_tables.txt
+echo "REDO DECLENSION TABLES ENDS"

nominals/pysanskritv2/tables/tableread.py

@@ -0,0 +1,49 @@
+"""table.py
+  Read conjugation table and construct dictionary
+"""
+class Table(object):
+ def __init__(self,model,key2,table):
+  self.model = model
+  self.key2 = key2
+  self.table = table
+  self.nused = 0 
+  #print(len(table))
+  assert len(self.table) == 24  # declension table is 8*3 = 24
+  self.tabstring = ':'.join(self.table)
+def init_table(filein0):
+ from os.path import dirname, abspath
+ import os,codecs,re
+ curdir = dirname(abspath(__file__))
+ filein = os.path.join(curdir,filein0)
+ with codecs.open(filein,"r","utf-8") as f:
+  lines = [line.rstrip() for line in f if not line.startswith(';')]
+  #recs = [Table(line) for line in f if not line.startswith(';')]
+ recs = []
+ #for iline,line in enumerate(lines):
+ nlines = len(lines)
+ for iline in range(0,nlines,9):
+  # 1st line of form 'Conjugation of <model> <key2>'
+  m = re.search(r'^Declension of (.*?) (.*?)$',lines[iline + 0])
+  if not m:
+   print('tableread ERROR1 @ line:',lines[iline + 0])
+   exit(1)
+  model = m.group(1)
+  key2 = m.group(2)
+  # initialize declension table
+  tab = []
+  for icase in [1,2,3,4,5,6,7,8]:
+   # 2nd-9th line of form 'Case c:  x y z'
+   m = re.search(r'^Case %s:  (.*?) (.*?) (.*?)$'%icase,lines[iline + icase])
+   if not m:
+    print('tableread ERROR2 @ line:',lines[iline + icase])
+    exit(1)
+   tab.append(m.group(1))
+   tab.append(m.group(2))
+   tab.append(m.group(3))
+  # replace missing values ('_' or '?') with empty string
+  tab = [x.replace('_','') for x in tab]
+  tab = [x.replace('?','') for x in tab]
+  # generate a Table record
+  rec = Table(model,key2,tab)
+  recs.append(rec)
+ return recs