def make_cal_pos_hashtable(cutoff=0):
obj = {}
with open(full_cal_db_location, 'rb') as cal:
for line in cal:
try:
lineObj = cal_tools.parseCalLine(line, False, False)
except IndexError:
print line
continue
word = lineObj["word"]
pos = lineObj["POS"]
if not word in obj:
obj[word] = []
#pos_set = set(obj[word])
#pos_set.add(pos)
obj[word].append(pos)
num_one_pos_words = 0
total_num_pos = 0
for word, pos in reversed(obj.items()):
pos_counts = {}
for p in pos:
if not p in pos_counts:
pos_counts[p] = 0
pos_counts[p] += 1
obj[word] = pos_counts
if len(pos_counts) < cutoff:
del obj[word]
continue
total_num_pos += len(pos_counts)
if len(pos_counts) == 1:
num_one_pos_words += 1
print "Percent Words With 1 POS", round(
100.0 * num_one_pos_words / len(obj), 3)
print "Avg Num POS per word", round(1.0 * total_num_pos / len(obj), 3)
cal_tools.saveUTFStr(obj, "cal_pos_hashtable.json")
f = codecs.open("double_pos_before_eng.txt", "wb", encoding='utf8')
for word, pos in obj.items():
f.write(u'{} ~-~ {}\n'.format(word, str(pos)))
f.close()
def make_cal_pos_hashtable(cutoff=0):
obj = {}
with open(full_cal_db_location,'rb') as cal:
for line in cal:
try:
lineObj = cal_tools.parseCalLine(line,False,False)
except IndexError:
print line
continue
word = lineObj["word"]
pos = lineObj["POS"]
if not word in obj:
obj[word] = []
#pos_set = set(obj[word])
#pos_set.add(pos)
obj[word].append(pos)
num_one_pos_words = 0
total_num_pos = 0
for word,pos in reversed(obj.items()):
pos_counts = {}
for p in pos:
if not p in pos_counts:
pos_counts[p] = 0
pos_counts[p] += 1
obj[word] = pos_counts
if len(pos_counts) < cutoff:
del obj[word]
continue
total_num_pos += len(pos_counts)
if len(pos_counts) == 1:
num_one_pos_words += 1
print "Percent Words With 1 POS",round(100.0*num_one_pos_words/len(obj),3)
print "Avg Num POS per word",round(1.0*total_num_pos/len(obj),3)
cal_tools.saveUTFStr(obj,"cal_pos_hashtable.json")
f = codecs.open("double_pos_before_eng.txt","wb",encoding='utf8')
for word,pos in obj.items():
f.write(u'{} ~-~ {}\n'.format(word,str(pos)))
f.close()
def make_cal_segments(mesechta):
def get_daf_str(daf_num, daf_side_num):
return '{}{}'.format(daf_num, 'a' if daf_side_num == 1 else 'b')
cal_gem_lines = []
with open("{}{}.txt".format(mesechta_cal_db_location, mesechta),
"rb") as f:
temp_gem_line = []
curr_gem_line_num = -1
curr_daf = ''
for line in f:
line_obj = cal_tools.parseCalLine(line, True, False)
line_obj["daf"] = get_daf_str(
line_obj['pg_num'], line_obj['side']) #add a daf str prop
line_obj["word"] = line_obj["word"].replace("'", '"')
if len(line_obj["word"]) > 1 and line_obj["word"][-1] == '"':
line_obj["word"] = line_obj["word"][0:
-1] #remove abbreviations
if line_obj["line_num"] != curr_gem_line_num:
if len(temp_gem_line) > 0:
small_gem_lines = [temp_gem_line]
has_big_lines = True
#recursively split up big lines until they're not big
while has_big_lines:
has_big_lines = False
new_small_gem_lines = []
for gem_line in small_gem_lines:
if len(gem_line) > 5:
has_big_lines = True
cut_index = len(gem_line) / 2
new_small_gem_lines.append(
gem_line[:cut_index])
new_small_gem_lines.append(
gem_line[cut_index:])
else:
new_small_gem_lines.append(gem_line)
small_gem_lines = new_small_gem_lines
for gem_line in small_gem_lines:
cal_gem_lines.append(gem_line)
temp_gem_line = [line_obj]
curr_gem_line_num = line_obj["line_num"]
else:
temp_gem_line.append(line_obj)
'''
#clean up lines with only 1 or 2 words
new_cal_gem_lines = []
new_cal_gem_dafs = []
for i,clt in enumerate(zip(cal_gem_lines,cal_gem_line_nums,cal_gem_dafs)):
cal_line = clt[0], line_num = clt[1], daf = clt[2]
if i > 0 and cal_gem_dafs[i-1] == daf and line_num-cal_gem_line_nums[i-1] <= 1:
p_cal_line = cal_gem_lines[i-1]
else:
p_cal_line = None
if i < len(cal_gem_lines)-1 and cal_gem_dafs[i+1] == daf and cal_gem_line_nums[i+1]-line_num <= 1:
n_cal_line = cal_gem_lines[i+1]
else:
n_cal_line = None
if len(cal_line) <= 2
'''
#break up by daf, concat lines to strs
all_daf_lines = []
all_dafs = []
curr_daf = ''
curr_daf_lines = []
for iline, line in enumerate(cal_gem_lines):
if line[0]["daf"] != curr_daf:
if len(curr_daf_lines) > 0:
all_daf_lines.append(curr_daf_lines)
all_dafs.append(curr_daf)
curr_daf = line[0]["daf"]
curr_daf_lines = [line]
else:
curr_daf_lines.append(line)
# dont forget to add the last daf in
if iline == len(cal_gem_lines) - 1:
if len(curr_daf_lines) > 0:
all_daf_lines.append(curr_daf_lines)
all_dafs.append(curr_daf)
cal_tools.saveUTFStr({
"lines": all_daf_lines,
"dafs": all_dafs
}, "cal_lines_{}.json".format(mesechta))
def make_cal_segments(mesechta):
def get_daf_str(daf_num,daf_side_num):
return '{}{}'.format(daf_num,'a' if daf_side_num == 1 else 'b')
cal_gem_lines = []
with open("{}{}.txt".format(mesechta_cal_db_location, mesechta), "rb") as f:
temp_gem_line = []
curr_gem_line_num = -1
curr_daf = ''
for line in f:
line_obj = cal_tools.parseCalLine(line,True,False)
line_obj["daf"] = get_daf_str(line_obj['pg_num'],line_obj['side']) #add a daf str prop
line_obj["word"] = line_obj["word"].replace("'",'"')
if len(line_obj["word"]) > 1 and line_obj["word"][-1] == '"':
line_obj["word"] = line_obj["word"][0:-1] #remove abbreviations
if line_obj["line_num"] != curr_gem_line_num:
if len(temp_gem_line) > 0:
small_gem_lines = [temp_gem_line]
has_big_lines = True
#recursively split up big lines until they're not big
while has_big_lines:
has_big_lines = False
new_small_gem_lines = []
for gem_line in small_gem_lines:
if len(gem_line) > 5:
has_big_lines = True
cut_index = len(gem_line)/2
new_small_gem_lines.append(gem_line[:cut_index])
new_small_gem_lines.append(gem_line[cut_index:])
else:
new_small_gem_lines.append(gem_line)
small_gem_lines = new_small_gem_lines
for gem_line in small_gem_lines:
cal_gem_lines.append(gem_line)
temp_gem_line = [line_obj]
curr_gem_line_num = line_obj["line_num"]
else:
temp_gem_line.append(line_obj)
'''
#clean up lines with only 1 or 2 words
new_cal_gem_lines = []
new_cal_gem_dafs = []
for i,clt in enumerate(zip(cal_gem_lines,cal_gem_line_nums,cal_gem_dafs)):
cal_line = clt[0], line_num = clt[1], daf = clt[2]
if i > 0 and cal_gem_dafs[i-1] == daf and line_num-cal_gem_line_nums[i-1] <= 1:
p_cal_line = cal_gem_lines[i-1]
else:
p_cal_line = None
if i < len(cal_gem_lines)-1 and cal_gem_dafs[i+1] == daf and cal_gem_line_nums[i+1]-line_num <= 1:
n_cal_line = cal_gem_lines[i+1]
else:
n_cal_line = None
if len(cal_line) <= 2
'''
#break up by daf, concat lines to strs
all_daf_lines = []
all_dafs = []
curr_daf = ''
curr_daf_lines = []
for line in cal_gem_lines:
if line[0]["daf"] != curr_daf:
if len(curr_daf_lines) > 0:
all_daf_lines.append(curr_daf_lines)
all_dafs.append(curr_daf)
curr_daf = line[0]["daf"]
curr_daf_lines = [line]
else:
curr_daf_lines.append(line)
cal_tools.saveUTFStr({"lines":all_daf_lines,"dafs":all_dafs},"cal_lines_{}.json".format(mesechta))
