Worksheet -> [[Text]] function?

[mb21]

This is not really urgent, but I was thinking of adding xlsx support (in addition to csv) to the pandoc-placetable filter, so I'd need a Worksheet -> [[Text]] function.

I guess I could come up with the fold Map (Int, Int) Cell -> [[Cell]] myself somehow (although there might be a neater way using Lens.Traversal?), but how about a Cell -> Text function that preserves some styling for numbers (like number of significant digits, currency, scientific notation etc.)?

[qrilka]

Currently xlsx does not support cell formatting - Styles is just a wrapper around a bytestring. I suppose it is required for your filter, right?

[mb21]

Yeah, I mean exporting to csv from a spreadsheet program usually preserves that kind of formatting... so without that I'm kind of hesitating to add xlsx as a dependency... thanks anyway :)

[qrilka]

I'll try to look into that area but no particular promises at the moment

[olafklinke]

FWIW, I happened to implement something along these lines, because the xls package has this and I wanted a unified Megaparsec parser for both xlsx and xls. Here is the code:

import Control.Arrow (first)
import Data.Function (on)
import Data.List (groupBy)
import Data.Map.Strict (toAscList)
cellRows :: CellMap -> [[Cell]]
cellRows = fmap (fillBlanks . fmap (first snd)) . groupBy ((==) `on` (fst.fst)) . toAscList where
    fillBlanks :: [(Int,Cell)] -> [Cell]
    fillBlanks = go 0
    blankCell = Cell Nothing Nothing Nothing Nothing
    go :: Int -> [(Int,Cell)] -> [Cell]
    go _ [] = []
    go i ((j,c):cs) = replicate (j-i-1) blankCell ++ (c : go j cs) 

[qrilka]

@olafklinke could we worth creating a PR if you have time for it

[olafklinke]

@olafklinke could we worth creating a PR if you have time for it

IMHO this would actually better be part of a separate package, see here.