The number conversion depends on the table only in the sense that certain entries are marked with a # if they represent > numbers or are part of the context in which a number would appear. This is a fairly short list, which I paste below:
〇,ling#0
零,ling#0
一,yi#1
二,er#2
兩,liang#2
两,liang#2
三,san#3
四,si#4
五,wu#5
六,liu#6
七,qi#7
八,ba#8
九,jiu#9
十,shi#10
廾,gong#20
廿,nian#20
卅,sa#30
卌,xi#40
百,bai#100
千,qian#1000
万,wan#10000
萬,wan#10000
亿,yi#100000000
億,yi#100000000
及,ji#
至,zhi#
年,nian#
月,yue#
日,ri#
第,di#
When the "#" is followed by a number, that represents the numeric value of the character. The other entries are: 及 and 至 ("to" when used in a date range), 年 (year), 月(month), 日 (day), and 第 (ordinal marker, used before a number to indicate 1st, 2nd, 3rd, etc).
As far as the format for date ranges, the code looks for any sequence of characters containing only numbers and/or the 5 date-related characters 及至 年 月 日. As long as that sequence contains both (年 and 月) or (月 and 日), then it considered a date and any numerical characters in that entire sequence are converted to numbers. This covers a variety of different formats that a date range could be in.
This can be done by reverse-engineering @tventimi 's
processNumbers()Parallelogram function:https://github.com/pulibrary/parallelogram/blob/main/cloudapp/src/app/pinyin.service.ts#L160
From earlier correspondence: