digitalpalitools / web-ui Goto Github PK

goals

• v bdhrs upload translations files to github and localized strings show up with **0** ongoing development work.

• translation stores

  • web-ui repo: public/locales/...
  • inflection-generator repo: declensions & conjugations.xlxs -> abbreviations text

stage 0: decouple development work from v bdhrs' uploading process

• in i18n.ts en/translations.json -> translations.en.json (this just puts all files in the same directory, after this v bdhrs can simply upload files into this folder)
  • test: the give giveFeedback button should continue to be localized in all languages
• add 1 column each for all supported locales into the abbreviations table in the inflections.db (these will be empty now, but will light up when v bdhrs add them)

stage 1: testability => enable us to see what portions are still unlocalized

• enable psuedoloc locale (ID is 'xx')
  • settable only using devtools console i.e. localStorage.currentScript = 'xx'
  • shows up and english in the drop down
  • there will be a corresponding translations.xx.json file
  • each string here will be same as .en except with ^ appended and $ prepended
• calls to pls will take in a lcid parameter
  • test with lcid = xx for now, actual string will be '^' + en string + '$' (e.g. opt => ^opt$)
  • the en string is taken from from 'name' column of abbreviations table

stage 2: make it work for en

• pull out all strings from the UI just like 'giveFeedback'
  • naming conventions is 'controlName.shortId'
• pls: test will lcid = en, the values will be taken from 'name' column of abbreviations table
  • test: the tables should now show abbreviations all throughtout
• test the above is complete by setting to psuedoloc locale

stage 3: integration testing

• web-ui: have v bdhrs provide with an translated strings say translations.si.json
  • test 1: no work should be needed from your side
  • test 2: web-ui is localized to si
  • test 3: for strings not localized / not present in the translations.si.json, fallback value is from translations.en.json
• pls: have v bdhrs add at least 1 other column to abbreviations table say si
  • test 1: no work should be needed from your side
  • test 2: ensure tables show up with those string

stage 4: other strings in PLS

• the file that is downloaded with the selected node names
• the pattern (eti pr) in summary
• the word 'like' in summary
• feedback link
• pron_type in pron_x types

• remove empty rows in vocative in cardinals "eka card" "dvi card" etc. test with eka dvi catu
  
• replace top left cell with part of speech e.g. adj, masc, fem nt
  in the verb tables, the top left cell is the name of the tense. this is great
  
  in other tables, the top left cell is the name of the pattern - replace with pos
  
• there is an issue with the second line - sometimes it is required (if there is no other information, like in the web-ui) but in the dictionary it is an unnecessary 3rd repetition of information. suggestions? [@parthopdas 24/4 - lets make this configurable so it is present in webui (and DPR when we get to it) but absent in GD]
  
• same with indeclinables - they required in the web ui, but unnecessary repetition of information in the dictionary. [@parthopdas 24/4 - lets make this configurable so it is present in webui (and DPR when we get to it) but absent in GD]
  

• For non-Roman keyboards, inputs are honored
• For Roman keyboards, they are transliterated on the fly

Note that db interaction must always happen in RO.

there's a good version of this in dpr to copy. i think this is the same

aa ā
ii ī
uu ū
"n ṅ
~n ñ
.t ṭ
.d ḍ
.n ṇ
.l ḷ
.m ṃ

Current Note is
nūna 1 ind
NOTE: nūna" does not have inflection table as it is an indeclinable

Replace with
nūna 1 is indeclinable

AC

• On the WF page a dropdown in header similar to
  
• Should show all 17 scripts in the drop down
• Selecting a script should remember it across browser sessions
• Consume the converter API from pali-script-converter project - it does not have an NPM package so please add directly
  yarn add https://github.com/Path-Nirvana-Foundation/pali-script-converter

p0 urgent - causing problems

• requested feature - work offline, even after page open and close.
• in the drop-down list, change alphabetical order - space must comes before pāli letters, so short words come to the top of the list. at the moment the most common words are buried at the bottom of very long lists.
• spot a mistake? update text to
  spot a mistake in the inflection table? something missing? fix it here!
  (to clearly differentiate it from dictionary feedback link)
  users are already using the wrong link for inflection feedback
  

complete

• replace Roman pos e.g. (pr) with pos in chosen script.
• change initial word from "ābādheti" to "pāli 2"
• vocative in pronouns "ahaṃ pron 1st" "tvaṃ pron 2nd" "ta pron" etc. test with ahaṃ tvaṃ taṃ
• Pāli1 input in /inflect should be transliterated
  • For non-Roman keyboards, inputs are honored
  • For Roman keyboards, they are transliterated on the fly
  • Note that db interaction must always happen in RO.
• remove double space between (pos) and (meaning)

handle white screen crashes when

• clicking in input box before db has finished loading
• pasting text with a punctuation mark into input box. test with pātv'ākāsi

• refer to the spec, and to v @bdhrs for more details
• Implementation to be done in yarn add https://github.com/Path-Nirvana-Foundation/pali-script-converter
• Check that it lights up in /word-frequency and in /converter

changes to the inflection table heading.
right now in the dictionary it occupies too much visual space and contains duplicate information

• line 1 not bold
• line 1 syntax: pañña "a adj" like dīgha
• line 1 syntax for irreg: addasa 1 "addasa aor" irreg
• table can follow immediately afterwards

tables

• centre text horizontally and vertically.
• balance all table sizes with widest column in each section
• indeclinable can be short form "ind"

These require intrusive design changes and will need to be handled separately. v @bdhrs if any of these are super-urgent let me know.

[Moved from /issues/27]

• remove empty columns in "a1 card" e.g pañca
  
• remove empty columns in "dvi card" e.g. dvi
  
• masc fem nt nouns need only 1 header line
  

Current Note is
NOTE: ehisi" does not have inflection table as it is an inflected form.

Replace with
see ehiti for inflection table

the word "ehiti" will be found in the "pattern" column
stem !
pattern ehiti

AC

• Structure on the lines of /count or /sort
• Drop down in the middle must have all 17 languages listed
• Consume the converter API from pali-script-converter project - it does not have an NPM package so please add directly
  yarn add https://github.com/Path-Nirvana-Foundation/pali-script-converter

Note:

1. An example of a site using the same library is: https://tipitaka.app/misc/convert/index.html