revert init_model.py back to pre-german state (because it makes more sense)

simplify token.n_rights and token.n_lefts
This commit is contained in:
Wolfgang Seeker 2016-03-21 16:10:25 +01:00
parent 5e2e8e951a
commit 5080077097
2 changed files with 5 additions and 18 deletions

View File

@ -98,7 +98,7 @@ def _read_probs(loc):
return probs, probs['-OOV-']
def _read_freqs(loc, max_length=100, min_doc_freq=0, min_freq=200):
def _read_freqs(loc, max_length=100, min_doc_freq=5, min_freq=200):
if not loc.exists():
print("Warning: Frequencies file not found")
return {}, 0.0
@ -125,8 +125,7 @@ def _read_freqs(loc, max_length=100, min_doc_freq=0, min_freq=200):
doc_freq = int(doc_freq)
freq = int(freq)
if doc_freq >= min_doc_freq and freq >= min_freq and len(key) < max_length:
# word = literal_eval(key)
word = key
word = literal_eval(key)
smooth_count = counts.smoother(int(freq))
log_smooth_count = math.log(smooth_count)
probs[word] = math.log(smooth_count) - log_total
@ -166,7 +165,7 @@ def setup_vocab(get_lex_attr, tag_map, src_dir, dst_dir):
clusters = _read_clusters(src_dir / 'clusters.txt')
probs, oov_prob = _read_probs(src_dir / 'words.sgt.prob')
if not probs:
probs, oov_prob = _read_freqs(src_dir / 'freqs.txt')
probs, oov_prob = _read_freqs(src_dir / 'freqs.txt.gz')
if not probs:
oov_prob = -20
else:

View File

@ -179,23 +179,11 @@ cdef class Token:
property n_lefts:
def __get__(self):
cdef int n = 0
cdef const TokenC* ptr = self.c - self.i
while ptr != self.c:
if ptr + ptr.head == self.c:
n += 1
ptr += 1
return n
return self.c.l_kids
property n_rights:
def __get__(self):
cdef int n = 0
cdef const TokenC* ptr = self.c + (self.array_len - self.i)
while ptr != self.c:
if ptr + ptr.head == self.c:
n += 1
ptr -= 1
return n
return self.c.r_kids
property lefts:
def __get__(self):