feat: enhance CDF variable display and refactor byte encoding

src/dataset.jl

@@ -38,6 +38,8 @@ function CDFDataset(filename)
     end
 end
 
+is_big_endian_encoding(cdf::CDFDataset) = is_big_endian_encoding(cdf.cdr.encoding)
+
 is_compressed(magic_numbers::UInt32) = magic_numbers != 0x0000FFFF
 majority(cdf::CDFDataset) = majority(cdf.cdr)
 
@@ -119,7 +121,9 @@ function Base.show(io::IO, m::MIME"text/plain", cdf::CDFDataset)
     println(io, "variables:")
     for key in keys(cdf)
         var = cdf[key]
-        print(io, "  $key : ", size(var), " ", DataType(var.vdr.data_type))
+        print(io, "  ", key, " : ", size(var), " ")
+        printstyled(io, variable_type(var); bold = true)
+        print(io, " ", DataType(var.vdr.data_type))
         !isempty(var) && print(io, " [", var[1], " … ", var[end], "]")
         println(io)
     end

src/enums.jl

@@ -40,9 +40,6 @@ end
     CDF_UCHAR = 52
 end
 
-DataType(x::DataType) = x
-DataType(x::UInt32) = DataType(Int32(x))
-
 Base.:(==)(x::RecordType, y::T) where {T <: Integer} = T(x) == y
 Base.:(==)(x::T, y::RecordType) where {T <: Integer} = x == T(y)
 Base.:(==)(x::DataType, y::T) where {T <: Integer} = T(x) == y

src/loading/variable.jl

@@ -43,10 +43,8 @@ function variable(cdf::CDFDataset, name)
     T = julia_type(vdr.data_type, vdr.num_elems)
     dims = (record_sizes(vdr)..., vdr.max_rec + 1)
     N = vdr isa VDR ? vdr.num_dims + 1 : length(dims)
-    byte_swap = is_big_endian_encoding(cdf.cdr.encoding)
-
     return CDFVariable{T, N, typeof(vdr), typeof(cdf)}(
-        name, vdr, cdf, dims, byte_swap
+        name, vdr, cdf, dims
     )
 end
 
@@ -132,7 +130,7 @@ function DiskArrays.readblock!(var::CDFVariable{T, N}, dest::AbstractArray{T}, r
             end
         end
     end
-    var.byte_swap && _btye_swap!(dest)
+    is_big_endian_encoding(var) && _btye_swap!(dest)
     return dest
 end
 

src/parsing.jl

@@ -1,9 +1,6 @@
 # CDF parsing utilities
 # Low-level binary reading and record parsing functions
 
-@inline read_be(io::IO, T) = ntoh(read(io, T))
-@inline read_be(io::IO, n, T) = ntuple(i -> read_be(io, T), n)
-
 # Buffer-based reading functions for zero-copy access
 # https://github.com/JuliaLang/julia/issues/31305
 @inline function read_be(v::Vector{UInt8}, i, T)
@@ -13,7 +10,7 @@
     end
 end
 
-@inline read_be(v::Vector{UInt8}, i, ::Type{RInt32}) = RInt32()
+@inline read_be(::Vector{UInt8}, i, ::Type{RInt32}) = RInt32()
 
 @inline function read_be(p::Ptr{T}, i) where {T}
     return ntoh(unsafe_load(p + (i - 1) * sizeof(T)))
@@ -143,16 +140,6 @@ function get_offsets!(offsets, buffer::Vector{UInt8}, pos, FieldSizeType)
 end
 get_offsets(args...) = get_offsets!(Int[], args...)
 
-
-# Big-endian readers (CDF uses big-endian for most fields)
-"""
-    read_uint32_be(io::IO)
-
-Read a 32-bit unsigned integer in big-endian byte order.
-CDF format uses big-endian for record fields.
-"""
-read_uint32_be(io::IO) = ntoh(read(io, UInt32))
-
 """
     is_cdf_v3(magic_bytes)
 

src/variable.jl

@@ -13,12 +13,12 @@ struct CDFVariable{T, N, V, P} <: AbstractVariable{T, N}
     vdr::V
     parentdataset::P
     dims::NTuple{N, Int}
-    byte_swap::Bool
 end
 
 Base.size(var::CDFVariable) = var.dims
 
 @inline majority(var::CDFVariable) = majority(var.parentdataset)
+@inline is_big_endian_encoding(var::CDFVariable) = is_big_endian_encoding(var.parentdataset)
 
 function dst_src_ranges(first, last, entry)
     overlap_first = max(first, entry.first)
@@ -87,10 +87,14 @@ end
 attrib(var::CDFVariable, name::String) = vattrib(var.parentdataset, var.vdr.num, name)
 attrib(var::CDFVariable) = vattrib(var.parentdataset, var.vdr.num)
 
-function CPR(var::CDFVariable)
-    vdr = var.vdr
-    cdf = var.parentdataset
-    return CPR(parent(cdf), vdr.cpr_or_spr_offset, recordsize_type(cdf))
-end
-
 is_record_varying(v::CDFVariable) = is_record_varying(v.vdr)
+variable_type(v::CDFVariable) = get(v.attrib, "VAR_TYPE", "unknown")
+
+function Base.show(io::IO, m::MIME"text/plain", var::CDFVariable)
+    summary(io, var)
+    println(io)
+    println(io, var.vdr)
+    print(io, "attributes: ")
+    show(io, m, var.attrib)
+    return
+end

test/runtests.jl

@@ -29,6 +29,7 @@ end
     @test first(ds) == ds["var"]
     @test occursin("Version: 3.9.0", string(ds))
     display(ds)
+    display(ds["var"])
 end
 
 @testset "Compressed cdf file (gzip)" begin