Changelog

All notable changes to this project will be documented in this file.

[0.2.0] - 2025-05-11

Added

• lastIndexWithData in Fits for tracking which HDU’s data is currently cached
• ensureHeadersParsed in FitsReader to prevent redundant header parsing
• makePadding utility for consistent block padding in FitsWriter

Changed

• Use nullish coalescing (??=) when initializing loadPromise in Fits
• Unified FitsData into three explicit interfaces: ImageData, TableData, and HeaderData
• FitsHeader.get / .getNumber / .getString / .getComment now return null instead of throwing on missing keys
• Refactored addHDU, getData, async iterators (entries), and serialization (toBytes) to handle optional data cleanly
• Overhauled header/data parsing in FitsIO, simplified offset math, removed vestigial error blocks
• Enhanced ImageIO and TableIO to operate on TypedArrays and fully support all FITS BITPIX values
• Improved formatHeaderLine to align numeric vs string values within 20-char fields and pad comments correctly
• Updated stats util to accept Iterable<number>, filter out non-finite values, and compute robust statistics

Fixed

• Brought all tests in line with new APIs: switched to TypedArrays in fixtures, updated expected error messages, and renamed spec files

[0.1.1] - 2025-04-07

Fixed

• Detecting correct data type based on header keywords and values

[0.1.0] - 2025-04-04

Added

• High-level Fits interface supporting both reading and writing of FITS files with a simple API.
• Full implementation of FitsWriter and FitsReader, supporting round-trip serialization.
• Support for automatically generating minimal FITS headers when missing.
• Auto-population of required header fields like NAXIS1, NAXIS2, and TFIELDS.
• New statistics utility: getStats() for tables and image HDUs.
• Extended FitsHeader typing and validation.
• Added entries() and entriesArray() for iteration over HDUs with lazy data loading.
• Robust test coverage for ImageIO, TableIO, HeaderIO, FitsReader, FitsWriter, and high-level Fits.
• Optional gzipped input support via pako.

Changed

• Refactored most of the FITS I/O code for clarity, testability, and separation of concerns.
• Rewrote FitsHeader parsing to handle modern FITS quirks more reliably.
• Ensured all exported data types support both Node.js and browser environments.
• Significantly improved coverage and reliability of unit tests.
• Internal refactor to simplify and unify how FitsHDU is constructed and handled.

Fixed

• Corrected parsing of fields like TFORMn, TTYPEn, and TUNITn in tables.
• Fixed multiple endian-ness and byte offset issues in data parsing.
• Ensured BSCALE/BZERO default handling in both image reader and writer paths.

Removed

• Browser-first assumptions: Switched to a Node-first setup
• Deprecated Blob-specific assumptions in internal file reader logic.

[0.0.1] - 2025-03-28

Added

• Initial development version of fits-ts, a TypeScript library for handling .fits files.
• Core functionality for parsing FITS headers, tables, and image data.
• Support for 8-bit, 16-bit, 32-bit, and floating-point image types.
• Application of BSCALE and BZERO for scaling image data.
• Parsing of FITS tables with support for multiple column types (L, B, I, J, E, D, A).
• Utility functions for handling FITS data types and field parsing.
• Unit tests for core parsers (Image, Table, Header).
• Support for reading FITS files from both gzipped and non-gzipped buffers.
• Preliminary documentation and usage examples.

Known Issues

• No support for compressed images beyond gzip.
• Limited handling of complex FITS structures or extensions.
• No support for hierarchical keywords or advanced metadata parsing.
• Limited error handling and validation for malformed FITS files.