FITS structure

FITS

Object to hold a single .fits file.

The fields are

• .filnam:: the .fits filename (:String)
• .hdu: the collection of header-data-unit objects (::Vector{FITS_HDU}`)

FITS_filnam

mutable FITS object to hold the decomposed name of a .fits file.

The fields are: " .value: for p#.fits this is p#.fits (::String)

• .name: for p#.fits this is p# (::String)
• .prefix: for p#.fits this is p (::String)
• .numerator: for p#.fits this is #, a serial number (e.g., '3') or a range (e.g., '3-7') (::String)
• .extension: for p#.fits this is .fits (::String)

FITS_HDU

Object to hold a single "Header and Data Unit" (HDU).

The fields are

• .hduindex:: identifier (a file may contain more than one HDU) (:Int)
• .header: the header object (::FITS_header)
• .dataobject: the data object (::FITS_dataobject)

NB. An empty data block (.dataobject = nothing) conforms to the standard.

FITS_header

Object to hold the header information of a FITS_HDU.

The fields are:

• .card: the array of cards (::Vector{FITS_card})
• .map: Dictionary keyword => recordindex (::Dict{String, Int})

FITS_card

Object to hold the card information of the FITS_header object.

The fields are:

• .cardindex: identifier of the header record (::Int)
• .record: the full record on the card (::String)
• .keyword: name of the corresponding header record (::String)
• .val: value of the corresponding header record (::Any)
• .comment: comment on the corresponding header record (::String)

FITS_dataobject

Object to hold the data of the FITS_HDU of given hdutype.

The fields are:

• .hdutype: accepted types are 'PRIMARY', 'IMAGE' and 'TABLE' (::String)
• .data: in the from appropriate for the hdutype (::Any)

cast_FITS_filnam(filnam::String)

Create the FITS_filnam object to decompose filnam into its name, prefix, numerator and extension.

Example:

julia> filnam = "T23.01.fits";

julia> n = cast_FITS_filnam(filnam);

julia> n.name, n.prefix, n.numerator, n.extension
("T23.01", "T23.", "01", ".fits")

cast_FITS_dataobject(hdutype::String, data)

Create the FITS_dataobject object for given hduindex constructed from the data in accordance to the specified hdutype: PRIMARY, IMAGE, ARRAY, TABLE (ASCII table) or BINTABLE (binary table).

Example:

julia> data = [11,21,31,12,22,23,13,23,33];

julia> data = reshape(data,(3,3));

julia> d = cast_FITS_dataobject("image", data)
FITS_dataobject("'IMAGE   '", [11 12 13; 21 22 23; 31 23 33])

julia> d.data
3×3 Matrix{Int64}:
 11  12  13
 21  22  23
 31  23  33

julia> d.hdutype
"'IMAGE   '"

cast_FITS_header(dataobject::FITS_dataobject)

Create the FITS_header object from the dataobject. The dataobject-input mode is used by fits_create to ceate the header object as part of creating the FITS object starting from Julia data input.

Example:

julia> data = [11 21 31; 12 22 23; 13 23 33];

julia> d = cast_FITS_dataobject("image", data);

julia> h = cast_FITS_header(d);

julia> h.map
Dict{String, Int64} with 7 entries:
  "BITPIX"   => 2
  "NAXIS2"   => 5
  "XTENSION" => 1
  "NAXIS1"   => 4
  ""         => 36
  "NAXIS"    => 3
  "END"      => 6

cast_FITS_header(record::Vector{String})

Create the FITS_header object from a block of (a multiple of) 36 single-record strings (of 80 printable ASCII characters). The record-input mode is used by fits_read after reading the header records from disk (see casting diagram above).

Example:

julia> record = [rpad("KEYWORD$i",8) * "'" * rpad("$i",70) * "'" for i=1:3];

julia> blanks = [repeat(' ', 80) for i = 1:36-length(record)];

julia> append!(record, blanks);         # to conform to the FITS standard

julia> h = cast_FITS_header(record);

julia> h.map
Dict{String, Int64} with 4 entries:
  "KEYWORD3" => 3
  "KEYWORD2" => 2
  "KEYWORD1" => 144
  ""         => 36

cast_FITS_card(cardindex::Int, record::String)

Create the FITS_card object for record with index cardindex.

Example:

julia> record = "SIMPLE  =                    T / file does conform to FITS standard             ";

julia> card = cast_FITS_card(1, record);

julia> card.cardindex, card.keyword, card.value, card.comment
(1, "SIMPLE", true, "file does conform to FITS standard             ")

cast_FITS_HDU(hduindex::Int, header::FITS_header, data::FITS_dataobject)

Create the FITS_HDU object for given hduindex, header and data.

Example:

julia> data = [11 21 31; 12 22 23; 13 23 33];

julia> d = cast_FITS_dataobject("image", data);

julia> h = cast_FITS_header(d);

julia> hdu = cast_FITS_HDU(1, h, d);

julia> hdu.dataobject.data
3×3 Matrix{Int64}:
 11  21  31
 12  22  23
 13  23  33

FITS(filnam::String, hdu::Vector{FITS_HDU})

Object to hold a single .fits file.

The fields are

• .filnam: filename of the corresponding .fits file (::String)
• .hdu: array of FITS_HDUs (::Vector{FITS_HDU})

Example:

julia> data = [11 21 31; 12 22 23; 13 23 33];

julia> d = cast_FITS_dataobject("image", data);

julia> h = cast_FITS_header(d);

julia> hdu = cast_FITS_HDU(1, h, d);

julia> f = cast_FITS("test.fits", [hdu]);

julia> f.hdu[1].dataobject.data
3×3 Matrix{Int64}:
 11  21  31
 12  22  23
 13  23  33

fits_zero_offset(T::Type)

Zero offset a as used in linear scaling equation

f(x) = a + b x,

where b is the scaling factor.

The default value is a = 0.0 for Real numeric types. For non-real types a = nothing.

Example:

julia> T = Type[Any, Bool, Int8, UInt8, Int16, UInt16, Int32, UInt32,
                  Int64, UInt64, Float16, Float32, Float64];

julia> o = (0.0, 0.0, -128, 0.0, 0.0, 32768,
                   0.0, 2147483648, 0.0, 9223372036854775808, 0.0, 0.0, 0.0);

julia> sum([fits_zero_offset(T[i]) == o[i] for i ∈ eachindex(T)]) == 13
true

fits_apply_offset(data)

Shift the UInt range of values onto the Int range by substracting from the data the appropriate integer offset value as specified by the BZERO keyword.

NB. Since the FITS format does not support a native unsigned integer data type (except UInt8), unsigned values of the types UInt16, UInt32 and UInt64, are stored as native signed integers of the types Int16, Int32 and Int64, respectively, after substracting the appropriate integer offset specified by the (positive) BZERO keyword value. For the byte data type (UInt8), the converse technique can be used to store signed byte values (Int8) as native unsigned values (UInt) after subtracting the (negative) BZERO offset value.

This method is included and used in storing of data to ensure backward compatibility with software not supporting native values of the types Int8, UInt16, UInt32 and UInt64.

Example:

julia> fits_apply_offset(UInt32[0])
1-element Vector{Int32}:
 -2147483648

julia> fits_apply_offset(Int8[0])
1-element Vector{UInt8}:
 0x80

julia> Int(0x80)
128

fits_remove_offset(data, bzero::Real)

Shift the Int range of values onto the UInt range by adding to the data the appropriate integer offset value as specified by the BZERO keyword.

NB. Since the FITS format does not support a native unsigned integer data type (except UInt8), unsigned values of the types UInt16, UInt32 and UInt64, are recovered from stored native signed integers of the types Int16, Int32 and Int64, respectively, by adding the appropriate integer offset specified by the (positive) BZERO keyword value. For the byte data type (UInt8), the converse technique can be used to recover the signed byte values (Int8) from the stored native unsigned values (UInt) by adding the (negative) BZERO offset value.

This method is included and used in reading stored data to ensure backward compatibility with software not supporting native values of the types Int8, UInt16, UInt32 and UInt64.

Example:

julia> fits_remove_offset(Int32[-2147483648])
1-element Vector{UInt32}:
 0x00000000

julia> Int(0x00000000)
0

julia> fits_remove_offset(UInt8[128])
1-element Vector{Int8}:
 0

FORTRAN_format

Object to hold a FORTRAN format specifier decomposed in its fields.

Accepted datatype specifiers are: Aw, Iw, Fw.d, Ew.d, Dw.d

Accepted output formating specifiers are: Aw, Iw.m, Bw.m, Ow.m, Zw.m, Fw.d, Ew.dEe, ENw.d, ESw.d, Gw.dEe, Dw.dEe. Notation: w - width, m (optional) - minimum number of digits, d - number of digits to right of decimal, e - number of digits in exponent N/S (optional) indicates engineering/scientific formating of the E type.

The fields are:

• .datatype: primary FORTRAN datatype (::String)
• .char: primary FORTRAN datatype character (::Char)
• .EngSci: secundary datatype character - N for engineering/ S for scientific (::Union{Char,Nothing})
• .width: width of numeric field (::Int)
• .nmin: minimum number of digits displayed (::Int)
• .ndec: number of digits to right of decimal (::Int)
• .nexp: number of digits in exponent (::Int)

cast_FORTRAN_format(format::String)

Decompose the format specifier format into its fields and cast this into the FORTRAN_format object. Allowed format specifiers are of the types: Aw, Iw.m, Bw.m, Ow.m, Zw.m, Fw.d, Ew.dEe, ENw.d, ESw.d, Gw.dEe, Dw.dEe, with: w - width, m(optional) - minimum number of digits, d - number of digits to right of decimal, e - number of digits in exponent; N/S (optional) indicates engineering/scientific formating of the E type.

Examples:

julia> cast_FORTRAN_format("I10")
FORTRAN_format("Iw", 'I', nothing, 10, 0, 0, 0)

julia> cast_FORTRAN_format("I10.12")
FORTRAN_format("Iw.m", 'I', nothing, 10, 12, 0, 0)

julia> F = cast_FORTRAN_format("E10.5E3")
FORTRAN_format("Ew.dEe", 'E', nothing, 10, 0, 5, 3)

julia> F.Type, F.TypeChar, F.EngSci, F.width, F.nmin, F.ndec, F.nexp
("Ew.dEe", 'E', nothing, 10, 0, 5, 3)  

FORTRAN_eltype_char(T::Type)

FORTRAN datatype description character for julia type T:

Bool => 'L', UInt8 => 'B', Int16 => 'I', UInt16 => 'I', Int32 => 'J', UInt32 => 'J', Int64 => 'K', UInt64 => 'K', Float32 => 'E', Float64 => 'D', ComplexF32 => 'C', ComplexF64 => 'M'

The character '-' is returned for non-primitive FORTRAN datatypes and for primitive datatypes not included in the FITS standard.

Examples:

julia> T = Type[Bool, Int8, UInt8, Int16, UInt16, Int32, UInt32, Int64, UInt64];

julia> print([FORTRAN_eltype_char(T[i]) for i ∈ eachindex(T)])
Int8: datatype not part of the FITS standard
['L', '-', 'B', 'I', 'I', 'J', 'J', 'K', 'K']

julia> T = [Float16, Float32, Float64, ComplexF32, ComplexF64];

julia> print([FORTRAN_eltype_char(T[i]) for i ∈ eachindex(T)])
Float16: datatype not part of the FITS standard
['-', 'E', 'D', 'C', 'M']

julia> T = [String, Vector{Char}, FITS];

julia> print([FORTRAN_eltype_char(T[i]) for i ∈ eachindex(T)])
Vector{Char}: not a FORTRAN datatype
FITS: not a FORTRAN datatype
['A', 'A', '-', '-']

parse_FITS_TABLE(hdu::FITS_HDU)

Parse FITS_TABLE (ASCII table) into a Vector of its columns for further processing by the user. Default formatting in ISO 2004 FORTRAN data format specified by keys "TFORMS1" - "TFORMSn"). Display formatting in ISO 2004 FORTRAN data format ("TDISP1" - "TDISPn") prepared for user editing.

Example:

strExample = "example.fits"
data = [10, 20, 30]
fits_create(strExample, data; protect=false)

t1 = Float16[1.01E-6,2.0E-6,3.0E-6,4.0E-6,5.0E-6]
t2 = [0x0000043e, 0x0000040c, 0x0000041f, 0x0000042e, 0x0000042f]
t3 = [1.23,2.12,3.,4.,5.]
t4 = ['a','b','c','d','e']
t5 = ["a","bb","ccc","dddd","ABCeeaeeEEEEEEEEEEEE"]
data = [t1,t2,t3,t4,t5]
fits_extend(strExample, data, "TABLE")

f = fits_read(strExample)
d = f[2].header.dict
d = [get(d,"TFORM\$i",0) for i=1:5]; println(strip.(d))
  SubString{String}["'E6.1    '", "'I4      '", "'F4.2    '", "'A1      '", "'A20     '"]

f[2].dataobject.data                            # this is the table hdu
  5-element Vector{String}:
   "1.0e-6 1086 1.23 a a                    "
   "2.0e-6 1036 2.12 b bb                   "
   "3.0e-6 1055 3.0  c ccc                  "
   "4.0e-6 1070 4.0  d dddd                 "
   "5.0e-6 1071 5.0  e ABCeeaeeEEEEEEEEEEEE "

parse_FITS_TABLE(f[2])
  5-element Vector{Vector{T} where T}:
   [1.0e-6, 2.0e-6, 3.0e-6, 4.0e-6, 5.0e-6]
   [1086, 1036, 1055, 1070, 1071]
   [1.23, 2.12, 3.0, 4.0, 5.0]
   ["a", "b", "c", "d", "e"]
   ["a                   ", "bb                  ", "ccc                 ", "dddd                ", "ABCeeaeeEEEEEEEEEEEE"]