FITS standard: Kaitai Struct description of any FITS file (FITS Standard 4.0)
StructureRepresentationInformation
· http://example.org/archive/structure-representation-information-ecbe8900
Write it out
Download the data description →
Its Representation Information, with all it needs, as a zip: description.ttl and each structure description and view specification (DFDL, Kaitai Struct, DRB, ...) as a file.
Its Archival Information Package
Properties
| comment | The Kaitai Struct equivalent of the DFDL description of any FITS file. |
| label | FITS standard: Kaitai Struct description of any FITS file (FITS Standard 4.0) |
| specificationLanguage | Kaitai Struct |
| specificationText | meta: id: fits title: FITS (Flexible Image Transport System), FITS Standard 4.0 file-extension: - fits - fit - fts endian: be encoding: ASCII doc: | Any FITS file, as the FITS Standard 4.0 (IAU FITS Working Group, 2018) defines it: a primary HDU, then any extensions, each HDU a header of 80-character keyword records ending with END, filled with blank records to a 2880-byte block, then its data, filled with zeros to a 2880-byte block. Sizes come from the mandatory keywords at their fixed positions: SIMPLE or XTENSION, BITPIX, NAXIS, NAXIS1 ... NAXISn, and in an extension PCOUNT and GCOUNT. The data has |BITPIX| x (NAXIS1 x ... x NAXISm) bits in a primary HDU (Eq. 1) and |BITPIX| x GCOUNT x (PCOUNT + NAXIS1 x ... x NAXISm) bits in an extension (Eq. 2); the product is written out for up to 9 axes. The primary array and IMAGE extensions are read as values of BITPIX's type, big-endian, the first axis varying fastest; a BINTABLE's rows as bytes, then its heap; a TABLE's rows as text; any other extension as bytes. A binary table's columns (its TFORMn keywords) and random groups (GROUPS, PCOUNT, GCOUNT) are given by keywords that can be anywhere in the header, so they aren't read here. seq: - id: hdus type: hdu repeat: eos types: hdu: seq: - id: first doc: SIMPLE (the primary HDU) or XTENSION (an extension). type: keyword_record - id: bitpix type: integer_record - id: naxis type: integer_record - id: naxisn type: integer_record repeat: expr repeat-expr: naxis.value - id: pcount type: integer_record if: is_extension - id: gcount type: integer_record if: is_extension - id: records doc: The other keyword records, and END. type: keyword_record repeat: until repeat-until: _.keyword == "END " - id: header_fill doc: Blank records filling the header's last 2880-byte block (36 records). size: '((36 - (3 + naxis.value + (is_extension ? 2 : 0) + records.size) % 36) % 36) * 80' - id: data size: data_bytes type: switch-on: layout cases: 1: array 2: binary_table 3: ascii_table - id: data_fill doc: Zeros filling the data's last 2880-byte block. size: (2880 - data_bytes % 2880) % 2880 instances: is_extension: value: first.keyword == "XTENSION" elements: doc: NAXIS1 x NAXIS2 x ... x NAXISm, or 0 when NAXIS is 0. value: >- naxis.value == 0 ? 0 : naxisn[0].value * (naxis.value >= 2 ? naxisn[1].value : 1) * (naxis.value >= 3 ? naxisn[2].value : 1) * (naxis.value >= 4 ? naxisn[3].value : 1) * (naxis.value >= 5 ? naxisn[4].value : 1) * (naxis.value >= 6 ? naxisn[5].value : 1) * (naxis.value >= 7 ? naxisn[6].value : 1) * (naxis.value >= 8 ? naxisn[7].value : 1) * (naxis.value >= 9 ? naxisn[8].value : 1) data_bytes: value: >- (bitpix.value < 0 ? -bitpix.value : bitpix.value) / 8 * (is_extension ? gcount.value * (pcount.value + elements) : elements) layout: doc: 1 for an array (primary or IMAGE), 2 a binary table, 3 an ASCII table, 0 anything else. value: >- not is_extension ? 1 : first.text.substring(2, 8) == "'IMAGE" ? 1 : first.text.substring(2, 11) == "'BINTABLE" ? 2 : first.text.substring(2, 8) == "'TABLE" ? 3 : 0 keyword_record: doc: One 80-character keyword record (Sect. 4.1); bytes 9 to 80 as text. seq: - id: keyword type: str size: 8 - id: text type: str size: 72 integer_record: doc: A keyword record whose value is a fixed-format integer (Sect. 4.2.3), e.g. BITPIX or NAXIS1. seq: - id: keyword type: str size: 8 - id: indicator type: str size: 2 - id: value_field size: 20 type: fixed_integer - id: comment type: str size: 50 instances: value: value: value_field.value fixed_integer: doc: | A fixed-format integer, right-justified in its 20 bytes (Sect. 4.2.3): the spaces before it, then the integer itself, with its sign if it has one. Only worked out when it's used, so other values are left as text. seq: - id: lead doc: The spaces before the integer, and its first character. type: u1 repeat: until repeat-until: _ != 0x20 or _io.eof instances: text: pos: lead.size - 1 size-eos: true type: str value: value: text.to_i array: doc: The primary array or an IMAGE extension's, as BITPIX says. seq: - id: values type: switch-on: _parent.bitpix.value cases: 8: u1 16: s2 32: s4 64: s8 -32: f4 -64: f8 repeat: expr repeat-expr: _parent.elements binary_table: doc: NAXIS2 rows of NAXIS1 bytes, then the heap (PCOUNT bytes). seq: - id: rows size: _parent.naxisn[0].value repeat: expr repeat-expr: _parent.naxisn[1].value - id: heap size-eos: true ascii_table: doc: NAXIS2 rows of NAXIS1 characters. seq: - id: rows type: str size: _parent.naxisn[0].value repeat: expr repeat-expr: _parent.naxisn[1].value |