morphodict icon indicating copy to clipboard operation
morphodict copied to clipboard

Show generated English translations for each cell in basic/full paradigms

Open aarppe opened this issue 3 years ago • 1 comments

In the basic and full paradigms, show as a pop-up for any cell a generated English translation based on the CW English gloss, e.g.

niwâpamâw, ê-wâpamiyan (Prs+1Sg+2SgO) -> I see you, I witness you
nikî-wâpamâw, ê-kî-wâpamiyan (Prt+1Sg+2SgO) -> I saw you, I witnessed you
nika-wâpamâw, ê-ka-wâpamiyan  (Fut+1Sg+2SgO) -> I am going to to see you, I am going to to witness you
niwî-wâpamâw, ê-wî-wâpamiyan  (Int+1Sg+2SgO) -> I want to see you, I want to witness you
wâpamitâni (Cond+1Sg+2SgO) -> I want to see you, I want to witness you
ka-wâpamiyan, ta-wâpamiyan  (Inf+1Sg+2SgO) -> for me to see you, for me to witness you
wâpamin (Imm+1Sg+2SgO) -> see now me, witness now me
wâpamîhka (Del+1Sg+2SgO) -> see later me, witness later me

This could be based on the following:

  • [?] 0. A transcriptor taking the standard English gloss in CW and converting that according to 1) tense/aspect, actor and goal features for verbs, and 2) number/obviation/locative/distributive, diminutive, and possessive features for nouns.

First versions of the transcriptors can be found here (these are quite big, and should be shrinkable using flag-diacritics):

ls -l ../lang-crk/src/transcriptions/transcriptor-cw-english-*
-rw-r--r--  1 arppe  staff  3299 25 Oct 22:21 ../lang-crk/src/transcriptions/transcriptor-cw-english-noun-entries2inflected-phrases.regex
-rw-r--r--  1 arppe  staff  6935 27 Oct 02:14 ../lang-crk/src/transcriptions/transcriptor-cw-english-verb-entries2inflected-phrases.regex
  • [x] 1. Generating English translations for all cells in full paradigms at the time of dictionary XML import, using the transcriptor (or its python equivalent), and storing these generated English translations.

  • [ ] 2. Generating English translations for individual cells, using transcriptor or its python equivalent, upon (user) demand.

How the above is implemented might be relevant to another feature, namely how we might match searches using English phrases with individual Cree word-forms (at least those available in the full paradigms).

aarppe avatar Oct 26 '20 21:10 aarppe