Languedoc Changeset - 54f61e475ab8

Changeset - 54f61e475ab8

Parent rev.

Child rev.

[Not reviewed]

default

0 1 0

Laman - 3 years ago 2022-09-29 21:19:32

merged all ngrams into a single frequencies field

1 file changed with 9 insertions and 15 deletions:

languedoc.py

0 comments (0 inline, 0 general)

languedoc.py

➞

Show inline comments

@@ @@ -46,24 +46,21 @@ def merge_ngram_freqs(freqs): @@
 class Sample:
 	def __init__(self, language="??", text=""):
 		self.language = language
-		self.frequencies = [dict(), dict(), dict()]
 		self.frequencies = dict()
 		if text:
 			self._extract(text)
 	def _extract(self, text):
 		for k in range(1, 4):
-			self.frequencies[k-1] = extract_ngram_freqs(text, k)
+			self.frequencies.update(extract_ngram_freqs(text, k))
 	@staticmethod
 	def merge(samples):
 		assert len({x.language for x in samples}) == 1
 		res = Sample(samples[0].language)
 		res.frequencies = []
 		for freqs in zip(*(x.frequencies for x in samples)):
 			res.frequencies.append(merge_ngram_freqs(freqs))
 		res.frequencies = merge_ngram_freqs([x.frequencies for x in samples])
 		return res
@@ @@ -72,16 +69,13 @@ class Sample: @@
 		use frequencies x order
 		use letter, digrams, trigrams
 		use absolute x square"""
 		own_frequencies = self.frequencies[0] | self.frequencies[1] | self.frequencies[2]
 		other_frequencies = other.frequencies[0] | other.frequencies[1] | other.frequencies[2]
 		ordered_own_ngrams = sorted(self.frequencies.items(), key=lambda kv: -kv[1])[:TOP_NGRAM_COUNT]
 		ordered_other_ngrams = sorted(other.frequencies.items(), key=lambda kv: -kv[1])[:TOP_NGRAM_COUNT]
 		ranked_own_ngrams = dict(zip([key for (key, freq) in ordered_own_ngrams], itertools.count(0)))
 		ranked_other_ngrams = dict(zip([key for (key, freq) in ordered_other_ngrams], itertools.count(0)))
 		ordered_own_trigrams = sorted(own_frequencies.items(), key=lambda kv: -kv[1])[:TOP_NGRAM_COUNT]
 		ordered_other_trigrams = sorted(other_frequencies.items(), key=lambda kv: -kv[1])[:TOP_NGRAM_COUNT]
 		ranked_own_trigrams = dict(zip([key for (key, freq) in ordered_own_trigrams], itertools.count(0)))
 		ranked_other_trigrams = dict(zip([key for (key, freq) in ordered_other_trigrams], itertools.count(0)))
 		res = sum(abs(v-ranked_other_trigrams.get(k, TOP_NGRAM_COUNT)) for (k, v) in ranked_own_trigrams.items()) + \
 					sum(abs(v-ranked_own_trigrams.get(k, TOP_NGRAM_COUNT)) for (k, v) in ranked_other_trigrams.items())
 		res = sum(abs(v-ranked_other_ngrams.get(k, TOP_NGRAM_COUNT)) for (k, v) in ranked_own_ngrams.items()) + \
 					sum(abs(v-ranked_own_ngrams.get(k, TOP_NGRAM_COUNT)) for (k, v) in ranked_other_ngrams.items())
 		return res

0 comments (0 inline, 0 general)