2nd of this month is incorrectly extracted

[facebook]

2nd of this month is incorrectly extracted

[patapizza]

Hi @linuxjain,

You're right, Duckling doesn't fully parse 2nd of this month:

ghc> debug (makeLocale EN $ Just US) "2nd of this month" [This Time]
this|last|next <cycle> (this month)
-- regex (this)
-- month (grain) (month)
-- -- regex (month)
[Entity {dim = "time", body = "this month", value = Object (fromList [("values",Array [Object (fromList [("value",String "2013-02-01T00:00:00.000-02:00"),("grain",String "month"),("type",String "value")])]),("value",String "2013-02-01T00:00:00.000-02:00"),("grain",String "month"),("type",String "value")]), start = 7, end = 17}]

However, it does parse on the 2nd of this month:

ghc> debug (makeLocale EN $ Just US) "on the 2nd of this month" [This Time]
on <date> (on the 2nd of this month)
-- regex (on)
-- intersect by ",", "of", "from", "'s" (the 2nd of this month)
-- -- the <day-of-month> (ordinal) (the 2nd)
-- -- -- regex (the)
-- -- -- ordinal (digits) (2nd)
-- -- -- -- regex (2nd)
-- -- regex (of)
-- -- this|last|next <cycle> (this month)
-- -- -- regex (this)
-- -- -- month (grain) (month)
-- -- -- -- regex (month)
[Entity {dim = "time", body = "on the 2nd of this month", value = Object (fromList [("values",Array [Object (fromList [("value",String "2013-02-02T00:00:00.000-02:00"),("grain",String "day"),("type",String "value")])]),("value",String "2013-02-02T00:00:00.000-02:00"),("grain",String "day"),("type",String "value")]), start = 0, end = 24}]
it :: [Entity]

And the 2nd of this month:

ghc> debug (makeLocale EN $ Just US) "the 2nd of this month" [This Time]
intersect by ",", "of", "from", "'s" (the 2nd of this month)
-- the <day-of-month> (ordinal) (the 2nd)
-- -- regex (the)
-- -- ordinal (digits) (2nd)
-- -- -- regex (2nd)
-- regex (of)
-- this|last|next <cycle> (this month)
-- -- regex (this)
-- -- month (grain) (month)
-- -- -- regex (month)
[Entity {dim = "time", body = "the 2nd of this month", value = Object (fromList [("values",Array [Object (fromList [("value",String "2013-02-02T00:00:00.000-02:00"),("grain",String "day"),("type",String "value")])]),("value",String "2013-02-02T00:00:00.000-02:00"),("grain",String "day"),("type",String "value")]), start = 0, end = 21}]
it :: [Entity]

To parse, 2nd of this month we would need to update a bit the rules.
Can you try to get a look at it?

[saurabhhssaurabh]

I have tried to look at the code but problem is I am not familiar with Clojure. I was facing problem in understanding functions definition and duckling flow. I will be needed guidance from a person who understands clojure well.

[patapizza]

@linuxjain Duckling is now in Haskell. ;) I'll try to see if I can get it going this week.

[saurabhhssaurabh]

Thank u.

[saurabhhssaurabh]

@patapizza I have updated issue with more failure cases. Can we load duckling only to parse date, not time? Currently if I load duckling it parses hour, minute and second along with year, month and day.

[patapizza]

@linuxjain What do you mean to only parse dates? Can you provide more context on what you are trying to do?

[saurabhhssaurabh]

@patapizza currently duckling parses a phrase and gives output as 2017-11-17T13:00:00.000+05:30 where 2017-11-17 represents date, 13:00:00.000 represents time and +05:30 represents time zone. So if I query like "show all movies released on 23.", 23 is parsed as time i.e 11 PM (23:00:00). So is there any flag or something else so that I can tell duckling to parse only date not time and time zone.

[patapizza]

@linuxjain I can't repro this - it doesn't return any Time for show all movies released on 23.. It does return a Time entity on on the 23th.

[saurabhhssaurabh]

@patapizza Ran this query on duckling demo page, It is considering 23 as hour (11 PM)-> Tuesday, 21 November 2017 at 23:00:00 -0800 (-08:00) :hour.

[patapizza]

@linuxjain We still need to update duckling.wit.ai to use the Haskell library (it's still running the Clojure version).

[chessai]

This is no longer an issue.

> debug (makeLocale EN Nothing) "2nd of this month" [Seal Time]
<day-of-month> (ordinal or number) of <month> (2nd of this month)
-- ordinal (digits) (2nd)
-- -- regex (2nd)
-- regex (of)
-- this|last|next <cycle> (this month)
-- -- regex (this)
-- -- month (grain) (month)
-- -- -- regex (month)
[Entity {dim = "time", body = "2nd of this month", value = RVal Time (TimeValue (SimpleValue (InstantValue {vValue = 2013-02-02 00:00:00 -0200, vGrain = Day})) [SimpleValue (InstantValue {vValue = 2013-02-02 00:00:00 -0200, vGrain = Day})] Nothing), start = 0, end = 17, latent = False, enode = Node {nodeRange = Range 0 17, token = Token Time TimeData{latent=False, grain=Day, form=Nothing, direction=Nothing, holiday=Nothing, hasTimezone=False}, children = [Node {nodeRange = Range 0 3, token = Token Ordinal (OrdinalData {value = 2}), children = [Node {nodeRange = Range 0 3, token = Token RegexMatch (GroupMatch ["2","nd"]), children = [], rule = Nothing}], rule = Just "ordinal (digits)"},Node {nodeRange = Range 4 6, token = Token RegexMatch (GroupMatch [""]), children = [], rule = Nothing},Node {nodeRange = Range 7 17, token = Token Time TimeData{latent=False, grain=Month, form=Nothing, direction=Nothing, holiday=Nothing, hasTimezone=False}, children = [Node {nodeRange = Range 7 11, token = Token RegexMatch (GroupMatch ["this","",""]), children = [], rule = Nothing},Node {nodeRange = Range 12 17, token = Token TimeGrain Month, children = [Node {nodeRange = Range 12 17, token = Token RegexMatch (GroupMatch []), children = [], rule = Nothing}], rule = Just "month (grain)"}], rule = Just "this|last|next <cycle>"}], rule = Just "<day-of-month> (ordinal or number) of <month>"}}]

❯ curl -XPOST http://0.0.0.0:5406/parse --data 'locale=en_US&text=2nd of this month' | jq
[
  {
    "body": "2nd of this month",
    "start": 0,
    "value": {
      "values": [
        {
          "value": "2021-04-02T00:00:00.000-07:00",
          "grain": "day",
          "type": "value"
        }
      ],
      "value": "2021-04-02T00:00:00.000-07:00",
      "grain": "day",
      "type": "value"
    },
    "end": 17,
    "dim": "time",
    "latent": false
  }
]