compatibility with DataFrames mini-language

[tbeason]

I use @chain from Chain.jl quite a bit, but it doesn't seem like @mfalse plays nicely with it. Do you know if it might be possible to get them to work well together?

[pdeffebach]

Can you show an MWE?

[tbeason]

Welp, perhaps it isn't about @chain at all! Does @mfalse not always work with filter?

Setup:

julia> using DataFrames, Chain, MissingsAsFalse

julia> df = DataFrame(x=rand(4),inds = [true, false, true, missing])
4×2 DataFrame
 Row │ x          inds
     │ Float64    Bool?
─────┼────────────────────
   1 │ 0.392307      true
   2 │ 0.20751      false
   3 │ 0.940361      true
   4 │ 0.0766065  missing

The actual problem:

julia> @mfalse filter("inds" => ==(true),df)
ERROR: TypeError: non-boolean (Missing) used in boolean context
...

julia> @mfalse df.inds.==true
4-element BitVector:
 1
 0
 1
 0

# I guess this works!?
julia> @mfalse filter(r -> r.inds==true,df)
2×2 DataFrame
 Row │ x         inds
     │ Float64   Bool?
─────┼─────────────────
   1 │ 0.392307   true
   2 │ 0.940361   true

What I want to actually do (which doesn't work)

# option 1 (preferred!)
@chain df begin
       @mfalse filter("inds" => ==(true),_)
       end

# option 2
@mfalse @chain df begin
       filter("inds" => ==(true),_)
       end

[tbeason]

Yes clearly now this is definitely about filter and not about @chain. Changing the issue title.

julia> @chain df begin
       @mfalse filter(r-> r.inds ==(true),_)
       end
2×2 DataFrame
 Row │ x         inds
     │ Float64   Bool?
─────┼─────────────────
   1 │ 0.392307   true
   2 │ 0.940361   true

[pdeffebach]

The problem is with ==(true). @mfalse isn't looking for the use of ==(true) in that way. I can add something that fixes it.

[pdeffebach]

@tbeason I have created a PR to fix this in #5 . Do you want to review?

[tbeason]

gonna be a few days, it's the end of the semester you know! I'm also quite unskilled in macro wizardry but perhaps that won't be necessary for review