Changelog

v1.0.0

Handle surface IDs >= 1000 as implicitly defined, if not given in the input file.

Surface IDs >= 1000 have a special meaning in MCNP. They are interpreted as (surf_id + 1000*cell_id), where cell_id refers to a cell card and surf_id refers to another surface card. For example, surface ID 123456 refers to cell 123 and surface 456. The meaning of this syntax is that surface 123456 is obtained by applying the transformation defined in the TRCL card of cell 123 to surface 456.
Print the real MCNP cell number in the T4 comments whenever possible.

Python 3.6 required.
Add caching for cell definitions. This reduces the redundance of the generated TRIPOLI-4 output files and usually results in faster conversion and faster TRIPOLI-4 execution.
Add CLI options to control cell inlining.
Fix conversion of isotopes starting with a zero.
Handle complement of lattice base cells.
Fix the conversion of cells with TRCL and FILL.
Move to poetry for package creation.

Implement conversion of SQ surfaces (fixes issue #11)
Improve error reporting (works towards fixing issue #23, although I doubt this issue can ever be closed).
Correctly support MCNP input files containing tabs.
Support rotation matrices in short form (3 parameters, 5 parameters, 6 parameters).
Largely reduce the memory footprint of the conversion process; this is crucial for largish (a few GB) output files.
Add a flag (--skip-deduplication) to skip surface deduplication.
Support references to macrobody facets in cell definitions.
Fix a few bugs.

Implement conversion of the LIKE n BUT syntax (fixes issue #26)
Implement conversion of hexagonal lattices (LAT=2, fixes issue #22)
Implement conversion of macrobodies (fixes issue #25)
Add a command-line option (-e) to specify the encoding of the input file (fixes issue #27)
Handle the MCNP data card shortcuts (14R, J, I, etc., fixes issue #24)
Support importance keywords for multiple particles
Support “message:” MCNP cards
Support affine transformations with 13 parameters
Improve error message on TatSu parse errors (partially fixes issue #23)
Add integration tests with MCNP and Oracle (fixes issue #6)
Add documentation for the Oracle tooling
Add a message with the list of the cells that were skipped during the conversion
Bug fixes:
- Fix parsing of consecutive complement operators (#1#2, fixes issue #33)
- Fix handling of TRCL keyword on cells defined using the complement operator (fixes issue #32)
- Fix several issues with line continuations (fixes issue #29)
- Fix parsing of input files without a newline at the end (fixes issue #28)
- Fix conversion of material cards containing keywords (ntab, ptab, etc.)
Lots of refactoring and cleaning

Support conversion of planes defined by three points (fixes issue #1).
Add support for 0 as a cone sheet specifier (fixes issue #2).
Add support for elliptic tori in MIP (fixes issue #3).
Handle the specification of FILL transformations by ID (fixes issue #4).
Partially handle conversion of lattices in fully-specified form (fixes issue #16).
Fix parsing of cell card options starting with *FILL in MIP.
Fix detection of line continuation.
Preserve the precision of isotope concentrations in the MCNP file.
Make pytest runnable from the package root directory.
Fix conversion of materials specified by density and atomic fractions (fixes issue #8).
Fix conversion of materials specified by total atomic concentration and atomic fractions (fixes issue #14).
Support spaces between the # operator and its arguments (fixes issue #10).
Handle the conversion of cards with a TRCL keyword (fixes issue #21).
Remove duplicated surfaces (fixes issue #19).
Do not use fixed IDs for the helper surface used in UNION cells (fixes issue #15).
Do not emit TRIPOLI-4 cells where the same surface appears twice.
Some linting and refactoring.