From 72f45e8e2c7385f41f2f8b1f063e7b5daa6dca94 Mon Sep 17 00:00:00 2001 From: Daniel Cartwright Date: Fri, 27 Aug 2021 12:10:58 -0700 Subject: [PATCH] Recognise hh:mm:ss as not an interval Summary: An interval regex was overzealous and matching too much, so `hh:mm:ss` was getting parsed as an interval instead of a time. Reviewed By: patapizza Differential Revision: D30608223 fbshipit-source-id: b24c18146070f15ada80b9401e67f0c0aefef7d8 --- Duckling/Ranking/Classifiers/EN_AU.hs | 160 ++++++++++++------------- Duckling/Ranking/Classifiers/EN_BZ.hs | 160 ++++++++++++------------- Duckling/Ranking/Classifiers/EN_CA.hs | 163 ++++++++++++-------------- Duckling/Ranking/Classifiers/EN_GB.hs | 162 ++++++++++++------------- Duckling/Ranking/Classifiers/EN_IE.hs | 160 ++++++++++++------------- Duckling/Ranking/Classifiers/EN_IN.hs | 160 ++++++++++++------------- Duckling/Ranking/Classifiers/EN_JM.hs | 160 ++++++++++++------------- Duckling/Ranking/Classifiers/EN_NZ.hs | 160 ++++++++++++------------- Duckling/Ranking/Classifiers/EN_PH.hs | 160 ++++++++++++------------- Duckling/Ranking/Classifiers/EN_TT.hs | 160 ++++++++++++------------- Duckling/Ranking/Classifiers/EN_US.hs | 163 ++++++++++++-------------- Duckling/Ranking/Classifiers/EN_XX.hs | 163 ++++++++++++-------------- Duckling/Ranking/Classifiers/EN_ZA.hs | 160 ++++++++++++------------- Duckling/Time/EN/Corpus.hs | 3 + Duckling/Time/EN/Rules.hs | 2 +- 15 files changed, 1018 insertions(+), 1078 deletions(-) diff --git a/Duckling/Ranking/Classifiers/EN_AU.hs b/Duckling/Ranking/Classifiers/EN_AU.hs index 6b3a71d1..7fec8cc2 100644 --- a/Duckling/Ranking/Classifiers/EN_AU.hs +++ b/Duckling/Ranking/Classifiers/EN_AU.hs @@ -83,11 +83,11 @@ classifiers likelihoods = HashMap.fromList [("", 0.0)], n = 3}}), ("integer (numeric)", Classifier{okData = - ClassData{prior = -0.53208562319284, unseen = -6.186208623900494, + ClassData{prior = -0.5357197883070597, unseen = -6.186208623900494, likelihoods = HashMap.fromList [("", 0.0)], n = 484}, koData = - ClassData{prior = -0.8852249122992647, unseen = -5.834810737062605, - likelihoods = HashMap.fromList [("", 0.0)], n = 340}}), + ClassData{prior = -0.8800742478577515, unseen = -5.84354441703136, + likelihoods = HashMap.fromList [("", 0.0)], n = 343}}), (" (bc|ad)", Classifier{okData = ClassData{prior = 0.0, unseen = -1.9459101490553135, @@ -1845,8 +1845,8 @@ classifiers likelihoods = HashMap.fromList [], n = 0}}), ("hh:mm:ss", Classifier{okData = - ClassData{prior = 0.0, unseen = -1.791759469228055, - likelihoods = HashMap.fromList [("", 0.0)], n = 4}, + ClassData{prior = 0.0, unseen = -1.9459101490553135, + likelihoods = HashMap.fromList [("", 0.0)], n = 5}, koData = ClassData{prior = -infinity, unseen = -0.6931471805599453, likelihoods = HashMap.fromList [], n = 0}}), @@ -2187,8 +2187,7 @@ classifiers likelihoods = HashMap.fromList [], n = 0}}), ("hh(:mm) - am|pm", Classifier{okData = - ClassData{prior = -0.10536051565782628, - unseen = -3.1354942159291497, + ClassData{prior = 0.0, unseen = -3.1354942159291497, likelihoods = HashMap.fromList [("time-of-day (latent)", -1.0116009116784799), @@ -2196,13 +2195,8 @@ classifiers ("minute", -1.9924301646902063)], n = 9}, koData = - ClassData{prior = -2.3025850929940455, - unseen = -1.9459101490553135, - likelihoods = - HashMap.fromList - [("time-of-day (latent)", -1.0986122886681098), - ("hour", -1.0986122886681098)], - n = 1}}), + ClassData{prior = -infinity, unseen = -1.6094379124341003, + likelihoods = HashMap.fromList [], n = 0}}), ("this|last|next ", Classifier{okData = ClassData{prior = -0.42050298547270487, @@ -2434,19 +2428,19 @@ classifiers likelihoods = HashMap.fromList [], n = 0}}), ("time-of-day (latent)", Classifier{okData = - ClassData{prior = -0.6970764586998348, unseen = -4.867534450455582, + ClassData{prior = -0.7009903580209712, unseen = -4.867534450455582, likelihoods = HashMap.fromList [("integer (numeric)", -9.763846956391606e-2), ("integer (0..19)", -2.374905754573672)], n = 127}, koData = - ClassData{prior = -0.689233281238809, unseen = -4.875197323201151, + ClassData{prior = -0.6853650401178903, unseen = -4.882801922586371, likelihoods = HashMap.fromList - [("integer (numeric)", -0.15800424914324832), - ("integer (0..19)", -1.923095471289142)], - n = 128}}), + [("integer (numeric)", -0.15669845190605702), + ("integer (0..19)", -1.9307583440347111)], + n = 129}}), ("beginning of year", Classifier{okData = ClassData{prior = 0.0, unseen = -1.791759469228055, @@ -2551,41 +2545,41 @@ classifiers likelihoods = HashMap.fromList [], n = 0}}), (" am|pm", Classifier{okData = - ClassData{prior = -0.2584829177595186, unseen = -5.697093486505405, + ClassData{prior = -0.262364264467491, unseen = -5.703782474656201, likelihoods = HashMap.fromList - [("from|since|after