diff options
| author | mjdenkowski <michael.j.denkowski@gmail.com> | 2014-04-10 16:42:50 -0400 |
|---|---|---|
| committer | mjdenkowski <michael.j.denkowski@gmail.com> | 2014-04-10 16:42:50 -0400 |
| commit | 83b450a6cd7c472783d4dcdb0239de9f297cd5b5 (patch) | |
| tree | 2141c3e19cf1e0af475616fe912bf90d0f4b6f23 /python/cdec/sa/features.py | |
| parent | 24f0fdb8a646f7aff501c757abe16ca8585444f8 (diff) | |
New feature: working implementation (online bilex)
Diffstat (limited to 'python/cdec/sa/features.py')
| -rw-r--r-- | python/cdec/sa/features.py | 25 |
1 files changed, 23 insertions, 2 deletions
diff --git a/python/cdec/sa/features.py b/python/cdec/sa/features.py index 1779f2f9..e19a18c0 100644 --- a/python/cdec/sa/features.py +++ b/python/cdec/sa/features.py @@ -40,13 +40,31 @@ def CoherenceProb(ctx): # c(f) / sample c(f) prob = (ctx.fcount + ctx.online.fcount) / (ctx.fsample_count + ctx.online.fsample_count) return -math.log10(prob) +# Not a feature, used for MaxLex +# bilex get_score for multiple instances +def get_lex_online(f, e, dir, bilex_list): + num = 0 + denom = 0 + for bilex in bilex_list: + if dir == 0: + denom += bilex.f.get(f, 0) + else: + denom += bilex.e.get(e, 0) + num += bilex.fe.get((f, e), 0) + if (not num) or (not denom): + return None + return num / denom + def MaxLexEgivenF(ttable): def MaxLexEgivenF(ctx): fwords = ctx.fphrase.words fwords.append('NULL') maxOffScore = 0.0 for e in ctx.ephrase.words: - maxScore = max(ttable.get_score(f, e, 0) for f in fwords) + if ctx.online: + maxScore = max(get_lex_online(f, e, 0, (ttable, ctx.online.bilex)) for f in fwords) + else: + maxScore = max(ttable.get_score(f, e, 0) for f in fwords) maxOffScore += -math.log10(maxScore) if maxScore > 0 else MAXSCORE return maxOffScore return MaxLexEgivenF @@ -57,7 +75,10 @@ def MaxLexFgivenE(ttable): ewords.append('NULL') maxOffScore = 0.0 for f in ctx.fphrase.words: - maxScore = max(ttable.get_score(f, e, 1) for e in ewords) + if ctx.online: + maxScore = max(get_lex_online(f, e, 1, (ttable, ctx.online.bilex)) for e in ewords) + else: + maxScore = max(ttable.get_score(f, e, 1) for e in ewords) maxOffScore += -math.log10(maxScore) if maxScore > 0 else MAXSCORE return maxOffScore return MaxLexFgivenE |