Ensure FilePaths work (#511)

* Ensure FilePaths work

* Use FilePathsBase instead of FilePaths

* Enables mmap on SystemPaths

* Adds test for FilePaths with DataFrame

* Adds compat for FilePathsBase

* Switches to dispatch-based approach

Project.toml

@@ -7,6 +7,7 @@ version = "0.5.13"
 CategoricalArrays = "324d7699-5711-5eae-9e2f-1d82baa6b597"
 DataFrames = "a93c6f00-e57d-5684-b7b6-d8193f3e46c0"
 Dates = "ade2ca70-3891-5945-98fb-dc099432e06a"
+FilePathsBase = "48062228-2e41-5def-b9a4-89aafe57970f"
 LazyArrays = "5078a376-72f3-5289-bfd5-ec5146d43c02"
 Mmap = "a63ad114-7e13-5084-954f-fe012c677804"
 Parsers = "69de0a69-1ddd-5017-9359-2bf0b02dc9f0"
@@ -19,11 +20,12 @@ WeakRefStrings = "ea10d353-3f73-51f8-a26c-33c1cb351aa5"
 [compat]
 CategoricalArrays = "0.5,0.6,0.7"
 DataFrames = "0.18,0.19,0.20"
+FilePathsBase = "0.6"
+LazyArrays = "0.12"
 Parsers = "0.3"
 PooledArrays = "0.5"
 Tables = "0.1,0.2"
 WeakRefStrings = "0.5,0.6"
-LazyArrays = "0.12"
 julia = "1"
 
 [extras]

src/CSV.jl

@@ -3,7 +3,7 @@ module CSV
 # stdlib
 using Mmap, Dates, Unicode
 using Parsers, Tables
-using PooledArrays, CategoricalArrays, WeakRefStrings, DataFrames, LazyArrays
+using PooledArrays, CategoricalArrays, WeakRefStrings, DataFrames, FilePathsBase, LazyArrays
 
 function validate(fullpath::Union{AbstractString,IO}; kwargs...)
     Base.depwarn("`CSV.validate` is deprecated. `CSV.read` now prints warnings on misshapen files.", :validate)
@@ -119,7 +119,7 @@ Read a UTF-8 CSV input (a filename given as a String or FilePaths.jl type, or an
 
 Opens the file and uses passed arguments to detect the number of columns and column types, unless column types are provided
 manually via the `types` keyword argument. Note that passing column types manually can increase performance and reduce the
-memory use for each column type provided (column types can be given as a `Vector` for all columns, or specified per column via 
+memory use for each column type provided (column types can be given as a `Vector` for all columns, or specified per column via
 name or index in a `Dict`). For text encodings other than UTF-8, see the [StringEncodings.jl](https://github.com/JuliaStrings/StringEncodings.jl)
 package for re-encoding a file or IO stream.
 The returned `CSV.File` object supports the [Tables.jl](https://github.com/JuliaData/Tables.jl) interface

src/utils.jl

@@ -175,27 +175,25 @@ function slurp(source)
     return final
 end
 
+getsource(source::Vector{UInt8}, ::Any) = source
+getsource(source::Cmd, ::Any) = Base.read(source)
+getsource(source::AbstractPath, ::Any) = Base.read(open(source))
+getsource(source::IO, ::Any) = slurp(source)
+getsource(source::SystemPath, use_mmap) = getsource(string(source), use_mmap)
 function getsource(source, use_mmap)
-    if source isa Vector{UInt8}
-        return source
-    elseif source isa Cmd
-        return Base.read(source)
-    elseif use_mmap && !isa(source, IO)
-        return Mmap.mmap(source)
-    elseif !isa(source, IO)
-        m = Mmap.mmap(source)
-        m2 = Mmap.mmap(Vector{UInt8}, length(m))
-        copyto!(m2, 1, m, 1, length(m))
-        finalize(m)
-        return m2
-    else
-        return slurp(source isa IO ? source : open(String(source)))
+    m = Mmap.mmap(source)
+    if use_mmap
+        return m
     end
+    m2 = Mmap.mmap(Vector{UInt8}, length(m))
+    copyto!(m2, 1, m, 1, length(m))
+    finalize(m)
+    return m2
 end
 
 getname(buf::Vector{UInt8}) = "<raw buffer>"
 getname(cmd::Cmd) = string(cmd)
-getname(str) = String(str)
+getname(str) = string(str)
 getname(io::I) where {I <: IO} = string("<", I, ">")
 
 const RESERVED = Set(["local", "global", "export", "let",

src/write.jl

@@ -9,7 +9,7 @@ Supported keyword arguments include:
 * `quotechar::Char='"'`: ascii character to use for quoting text fields that may contain delimiters or newlines
 * `openquotechar::Char`: instead of `quotechar`, use `openquotechar` and `closequotechar` to support different starting and ending quote characters
 * `escapechar::Char='"'`: ascii character used to escape quote characters in a text field
-* `missingstring::String=""`: string to print for `missing` values 
+* `missingstring::String=""`: string to print for `missing` values
 * `dateformat=Dates.default_format(T)`: the date format string to use for printing out `Date` & `DateTime` columns
 * `append=false`: whether to append writing to an existing file/IO, if `true`, it will not write column names by default
 * `writeheader=!append`: whether to write an initial row of delimited column names, not written by default if appending
@@ -136,7 +136,7 @@ function with(f::Function, io::Union{Base.TTY, Base.Pipe, Base.PipeEndpoint, Bas
     f(io)
 end
 
-function with(f::Function, file::String, append)
+function with(f::Function, file::Union{AbstractString, AbstractPath}, append)
     open(file, append ? "a" : "w") do io
         f(io)
     end

test/runtests.jl

@@ -1,4 +1,4 @@
-using Test, CSV, Dates, Tables, DataFrames, CategoricalArrays, PooledArrays, CodecZlib
+using Test, CSV, Dates, Tables, DataFrames, CategoricalArrays, PooledArrays, CodecZlib, FilePathsBase
 
 const dir = joinpath(dirname(pathof(CSV)), "..", "test", "testfiles")
 
@@ -75,7 +75,7 @@ end
     v = f.X[1]
     @test v == "b"
     @test levels(v.pool) == ["a", "b", "c"]
-    
+
     f = CSV.read(IOBuffer("X\nb\nc\na\nc"), categorical=true, copycols=true)
     v = f.X[1]
     @test v == "b"

test/testfiles.jl

@@ -1,4 +1,4 @@
-function testfile(file, kwargs, expected_sz, expected_sch, testfunc)
+function testfile(file, kwargs, expected_sz, expected_sch, testfunc; dir=dir)
     println("testing $file")
     if file isa IO
         seekstart(file)
@@ -24,7 +24,7 @@ function testfile(file, kwargs, expected_sz, expected_sch, testfunc)
 end
 
 testfiles = [
-    # file, kwargs, expected_sz, expected_sch, testfunc = 
+    # file, kwargs, expected_sz, expected_sch, testfunc =
     ("test_utf8_with_BOM.csv", NamedTuple(),
         (3, 3),
         NamedTuple{(:col1, :col2, :col3),Tuple{Float64,Float64,Float64}},
@@ -605,3 +605,10 @@ testfiles = [
 for test in testfiles
     testfile(test...)
 end
+# Test file with FilePaths
+testfile("test_basic.csv", (types=Dict(2=>Float64),),
+    (3, 3),
+    NamedTuple{(:col1, :col2, :col3),Tuple{Int64,Float64,Int64}},
+    (col1 = [1, 4, 7], col2 = [2.0, 5.0, 8.0], col3 = [3, 6, 9]);
+    dir=Path(dir)
+)

test/write.jl

@@ -70,6 +70,11 @@ using CSV, Dates, WeakRefStrings, CategoricalArrays, Tables
     @test String(read(file)) == "col1,col2,col3\n1,4,7\n2,5,8\n3,6,9\n"
     rm(file)
 
+    filepath = Path(file)
+    (col1=[1,2,3], col2=[4,5,6], col3=[7,8,9]) |> CSV.write(filepath)
+    @test String(read(filepath)) == "col1,col2,col3\n1,4,7\n2,5,8\n3,6,9\n"
+    rm(filepath)
+
     open(file, "w") do io
         (col1=[1,2,3], col2=[4,5,6], col3=[7,8,9]) |> CSV.write(io)
     end
@@ -165,7 +170,7 @@ using CSV, Dates, WeakRefStrings, CategoricalArrays, Tables
 
     # validate char args: #369
     @test_throws ArgumentError (col1=[1,2,3], col2=[4,5,6], col3=[7,8,9]) |> CSV.write(io; escapechar='☃')
-    
+
     # custom float decimal: #385
     (col1=[1.1,2.2,3.3], col2=[4,5,6], col3=[7,8,9]) |> CSV.write(io; delim='\t', decimal=',')
     @test String(take!(io)) == "col1\tcol2\tcol3\n1,1\t4\t7\n2,2\t5\t8\n3,3\t6\t9\n"
@@ -203,4 +208,9 @@ using CSV, Dates, WeakRefStrings, CategoricalArrays, Tables
     (col1=[""],) |> CSV.write(io)
     @test String(take!(io)) == "col1\n\n"
 
+    # test with FilePath
+    mktmpdir() do tmp
+        CSV.write(tmp / "test.txt", df)
+        @test CSV.read(tmp / "test.txt") == df
+    end
 end # @testset "CSV.write"