tristaaan / saythisnumber Goto Github PK

Or some equivalent linter.

Numbers become innaccurate in human readable language (confirmed in English), around the ten quadrillion mark:

Correct:
http://www.saythisnumber.in/english/1000000000000099
Expected result: One quadrillion and ninety nine
Actual result: One quadrillion and ninety nine

Incorrect:
http://www.saythisnumber.in/english/10000000000000099
Expected result: ten quadrillion and ninety nine
Actual result: ten quadrillion and one hundred

A few of the numbers in the English parser are not spelled correctly.

'ninty' -> 'ninety'
'fourty' -> 'forty'

This would be more useful as an npm module than an application with some endpoint.

Numbers 16 - 19 (and 26 - 29, 36 - 39, etc.) have wrong the spacing. Instead of 'diezyseis' it should be 'diez y seis', and the same thing for all the others. Starting from 200, it becomes 'cientos' instead of 'ciento', so from there it should be 'doscientos', 'doscientos uno', and so on, later 'trescientos', 'cuatrocientos' until 'quinientos' (where the bug is fixed). Then again 600 is broken (should be 'seiscientos', 'setecientos', 'ochocientos' and finally 'novecientos'.

maybe it's a bit silly.

1024 => 1️⃣0️⃣2️⃣4️⃣

Only integers are supported currently.

As specified by ISO 639-3

There is a different word for 500, it's not 'cinco cincuenta'

Would be nice to not only have a normal page output but also in json, csv, text.

Example:

2-5

text/plain

two(newline)
three(newline)
four(newline)
five(newline)

csv

two,three,four,five;

json

{[
"2" : "two",
"3" : "three",
"4" : "four",
"5" : "five",
]}

Can I do that?

Calling http://www.saythisnumber.in/english/4 everything goes well.

http://www.saythisnumber.in/english/4[a-z] will always get a header that it cannot GET this output.

But http://www.saythisnumber.in/english/4% will get the following error:

Error: Failed to decode param '4%' at Route.match (/app/node_modules/express/lib/router/route.js:65:17) at Router.matchRequest (/app/node_modules/express/lib/router/index.js:244:17) at pass (/app/node_modules/express/lib/router/index.js:104:30) at Router._dispatch (/app/node_modules/express/lib/router/index.js:173:5) at Object.router (/app/node_modules/express/lib/router/index.js:33:10) at next (/app/node_modules/express/node_modules/connect/lib/proto.js:194:15) at Object.expressInit as handle at next (/app/node_modules/express/node_modules/connect/lib/proto.js:194:15) at Object.query as handle at next (/app/node_modules/express/node_modules/connect/lib/proto.js:194:15)

To prevent or uncover bugs when when adding new languages or features. Build into package.json so npm test

• French numbers from 1000-1999 should not have 'un' prefix: 1024 => mille vingt-quatre
• 200 is "deux cents"
• 80 is "quatre-vingts", even though 82 is indeed "quatre-vingt-deux"
• 81 is "quatre-vingt-un"

No transliterations, sub in correct characters.

If the numbers get too big, there should be a warning / error instead of getting an undefined answer.

Example: http://saythisnumberin.herokuapp.com/english/1000000000000000000

is currently: "one undefined"

Right now this is specific for Chinese (Pinyin, Simplified, Traditional) and Japanese (Hiragana, Kanji, Romaji)

Perhaps choose a default and allow user to specify a different one?

say.thisNumber(32).in('chinese'); // => defaults to simplified(?)
say.thisNumber(32).in('chinese', 'pinyin'); 

This could be used to specify numeral systems of the language too?

• Keep the font as Times New Roman because it has perfect international support.
• Keep it minimal.