- Maintenance
- update workflow to accommodate security-driven changes (#1445)
- turn off all fortran (spatial solvers and ENSDF processing) by default (#1444)
- update where to look for DAGMC CMake files (#1446)
- ensure NUC_DATA_PATH is a string (#1447)
- update CI to Ubuntu 20.04 (#1449)
- New Capabilities
- Adds method to write a MaterialLibrary as an OpenMC XML file.
- Change
- exclude docker related stuff from build_test (#1404 #1408)
- move changelog test in its own workflow (#1404)
- Adds 2 templates header files (one for Python one for CPP) to store PYNE version number, cmake will automatically configure proper version files for python and cpp based on the PYNE VERSION number located in the root CMakeFile.txt (#1415)
- PyNE CI docker container now target DAGMC:'stable' tag instead of develop. (#1415)
- Move canonical versioning to python (#1434)
Fix
- fixing VR test (#1399)
- fixing output from MaterialLibrary warning. (#1414)
- fixing docker publish CI Github Action (#1426)
- Maintenance
- removing old circle-ci related stuff (#1405)
- adds a manual trigger for docker build workflow (#1406)
- clarify the tag name for docker images generated in CI (#1407)
- add ifbar to cleanup the code about progress bar (#1409)
- use ifbar to print progress of mesh_to_fluxin (#1401)
- pep8 formatting applied to all py files (#1436)
New Capabilities
- uses GithubAction to build the docker image used in CI (#1394)
Change
- CI uses GithubAction instead of CIrcleCI (#1393, #1395)
Fix
- Attribution for the atomic mass data in the data theory doc page (#1387)
- specify encoding when reading files in amalgamation (#1386)
- little fix for usage of filesystem::canonical() for windows... (#1390)
Maintenance
- update MOAB version to 5.3.0 in CI docker image (#1391)
- optimize the mesh_to_fluxin function in r2s step1 (#1397)
- use progress bar for mesh_to_fluxin function (#1401, #1402)
Change
- Update pyne::ValueError to return a pointer to a stored string instead of an out-of-scope string. (#1358)
- Setting CMP0074 in CMakeLists.txt to allow *_ROOT variables in the CMake command.
- Change the e_bounds tag unit from eV to MeV (#1353)
- Add functions to tag the decay_time, source_intensity and version to source.h5m (#1352)
- Add the workflow to support PyNE R2S with DAG-OpenMC (#1355)
- Add vector capabilities to the _do_ops() method in mesh.py (#1371)
- Add tests for multiplication of vector-valued tags by scalars and scalar-valued tags (#1374)
- Add multiplication operator for NativeMeshTag and corresponding tests (#1376)
- Incorporate native use of multiplication symbol for NativeMeshTag (#1411)
- Addition, subtraction, and division symbols are now available to manipulate NativeMeshTags (#1412)
Fix
- An e_bounds reading problem when old sampler is used without e_bounds text file (#1353)
- Fix the compatibility problem of transmuters with numpy version v1.19.x
- Fix to MaterialLibrary.__delitem__ to call the underlying instance.
- Few fixes to enable Windows compilation the DAGMC deps.
- Fix documentation to say that the minimum Cython version is 0.29.21 (#1370)
Maintenance
- Auto-update copyright year in docs (#1369)
- Scrub out Trelis mention (#1377)
Change
- Update OVA link in the docs (#1348)
- Bump license date to 2020 (#1348)
- Add the energy boundaries to photon source file (#1341)
- Add a new source sampling sampler without e_bounds (#1341)
- Fix
- Compatibility fix with MCNP6 ptracs. (#1336)
- Add some missing pieces in MaterialLibrary API to match std::unordered_map one (#1350)
Change
- Now attribute a number to material when adding them into a material library (#1334)
Fix
- change the ref address of the materials group in the nuc_data.h5 material_library (to match new format) (#1337)
- clean the remaining calls to the old material_library write_hdf5 API
- Maintenance
- Add dockerhub authentication for pulling images
- New Capabilities
- Adding unordered_map like API to MaterialLibrary pointing directly to the nested unordered_map
Maintenance
- Documentation Changes
- Update conda install instruction. (#1324)
- move join_string declaration in utils header to allow amalgamate PyNE to be compiled with clang
- Fix the incompatibility between MOAB 5.2.0 and test_activation_responses (#1328)
- Bug fixes
- Update version numbering for release v0.7.1 (#1326)
- Improvements in building and testing
- Update moab version 5.1.0 >> 5.2.1 (#1331)
- IMPORTANT: Python 2 is no longer supported as of this release. Some packages may continue to work,
- but there is no longer testing of Python 2 during development.
New Capabilities
- Subvoxel R2S allows activation and photon source sampling by material within a mesh voxel
- SubMode parameter in source_sampling
- Add code for submode: SUBVOXEL
- Test functions to test SUBVOXEL codes in test_source_sampling.py
- Add subvoxel r2s workflow support in pyne/scripts/r2s.py
- Use new sampler for both voxel and sub-voxel R2S.
- Define cell_mats for subvoxel equals False
- Function Sampler::mesh_tag_data to be compatible for both DEFAULT and SUBVOXEL mode
- OpenMC support
- Add CI test for OpenMC as an optional dependency
- Add options to install OpenMC Python API
- Functions to read OpenMC state point file to get meshtally data
- File scripts/activation_responses.py to get activation responses data for ALARA output file for better visualization. Responses include:
- decay heat
- specific activity
- alpha heat
- beta heat
- gamma heat
- wdr
- photon_source
- multiple response in one output.txt file
- first introduction of GT-CADIS workflows
- added gtcadis.py script
- first step for the GT-CADIS workflow, further steps to follow
Enhancements to Previous Capabilities
MCNP Support * Default library reading capability from material card, such as nlib, plib and etc. (#1316)
Turn off QA warnings by default and enable with env variable 'PYNE_QA_WARN' (#1268)
- Enhancements of Material and MaterialLibrary capabilities
A C++ API for the MaterialLibrary class has been created for direct use in compiled software tools
Change the MaterialLibrary Python API to bind to the new c++ one.
- Overhaul of HDF5 format of Material and MaterialLibrary
Capability to read nuclides from a specific path when loading material from an hdf5 file (not necessarily relying on autodetect the nucpath).
change the structure of material when written in a hdf5 file:
/material/ /material/my_mat/ /material/my_mat/composition /material/my_mat/nuclidelist /material/my_mat/composition_metadata
where,
/materialand/material/my_matare now HDF5 groups
- Material I/O formats
- pyne::Material: capability to form PHITS input material card
- capability to form a uwuw material name
- A new argument
mult_den(bool) that makes Material class's get_density_frac method (hence also mcnp and phits methods) write atom/mass density if true, otherwise write atom/mass fraction. - increased precision of the material density reported in the comment card for an MCNP material to 5 decimal places
- Adds tests to tests/test:materials.py:test_decay_heat(self) to check more isotopes
- Adds from_activity() to create or redefine materials using radionuclide activities
- Bug fixes in serpent support
- in serpent.py, function VOL_serp2_fix() to correct for _VOL variable being an array. as seen in serpent 2
- fix error in serpent parse_dep
- serpent.py function parse_dep. Catches ValueError that occurs when parse_dep attempts to make_mats with a serpent 2 *_dep.m file due to the *_VOL variable not being a float
- Fixes in examples and tutorials
- in half_life example, changed nuc_data call in anum to give int data type
- in Data Sources, rerun example to update output
- in Chainsolve Transmutation, rerun example to update output
- pytaps to pymoab update in intro/diffusion tutorials (#1282, #1304)
- in material_library example, updated import and class usage to reflect MaterialLibrary changes (#1280)
- Removed QA warnings
- in ACE Module example, change urllib for requests
- adds broken yt warning to examples notebooks, updates deprecated urlretrieve to requests (#1295, #1304)
- remove Origen 2.2: H1 Cross Section for PyNE repo not supporting necessary inputs to reproduce the example.
- in mesh basic tutorial, add broken yt warning
- in endf reader example, cross-section plots from ENDF/B-VII.1 and updated module for deprecated urlretrieve to requests (#1297, #1317)
- change to integer division in example in nuclide naming tutorial, fix typo (#1310)
- Changes in source sampling for mesh-based Monte Carlo sources
- Add statistics summary output of find_cell failure in source sampling.
- Add the ability to allow user turn off the void rejection in source sampling.
- Add cell_fracs and cell_number tags for both default and subvoxel r2s modes
- Check for the existence of the e_bounds file. Print error message when it's missing.
- Check for bias_tag data. Report error when bias tag data are all zero
- Check 'cell_fracs' tag in source_sampling.cpp when sub_mode is DEFAULT. Prevent wrong use of source.h5m.
- Fix the problem of reading cell_number_tag with size of 1
- Change mode range of cell rejection from >3 to >2
- Sort cell_fracs according to the order of 'idx' and 'vol_frac'. For faster source sampling.
- Pass cell_list back to Fortran, to speed up source sampling.
- function to write total photon intensities for subvoxel r2s
- Removed variables
`icl_tmp`and`find_cell`which are not longer needed. - MCNP6 version of source.F90
- Changed source.F90 to use "implicit none" instead of "implicit real"
- Addition & updates of unit tests for above improvements
- Improvements in Rigorous-2-Step shutdown dose rate analysis workflow
- Documentation improvements
- Provide example files for variety of problems/problem modes
- Improvements in testing of R2S
- Use example files for automated testing
- Streamline code related to addition of subvoxel mode
- Combine the subvoxel/voxel R2S loops to calculate the total photon source intensities.
- Keep cell_number, cell_fracs, cell_largest_frac_number and cell_largest_frac tag in r2s step1
- Use subvoxel and normal r2s compatible workflow parameters
- Input check of cell_fracs tag under voxel mode. As the cell_fracs tag is there for voxel/sub-voxel mode.
- Load geom and calculate cell_mats in r2s step2
- Read decay times from r2s config.ini, and then write them into alara_inp.
- In R2S step2, add option to write only 'total' to h5 file, reduce the CPU time
- Error in voxel R2S.
- Changes in processing of ALARA input/output
- Change some default names of alara_inp.
- Decay times in the alara_params.txt.
- Add input units check to the function utils.py/to_sec
- Use function utils.py/to_sec to replace alara.py/_TO_SEC
- Simplify the method to get the list of decay/cooling times
- Nuclear Data Handling and Reporting
- Fixed issue where some gamma x-rays where throwing
NotANuclideerrors because the underlying nuclides were being read & recorded with negative ids. All nuclide ids are now ensured to be positive. - Misidentification of descriptive text in (MF,MT)=(1,451) as contents lines.
- decay_heat() in material.cpp now calls metastable_id to convert zas_id to state_id
- Fix ENDF parsing of TSL files with short collision time approximation for non-principal atoms.
- endf.Library._read_headers() and regular expressions in endf.pyx
- Removed regexps: CONTENTS_R, SPACE66_R, NUMERICAL_DATA_R
- Added regexps: SPACEINT11_R
- Added methods: _isContentLine(parts)
- ENSDF database link to 2019 Oct 4th database
- Update the C012-n.ace file link.
- Missing elements name_to_zz dictionary
- Updated half_life in data.pyx to return nan if isotope not found (#1257)
- Fixed issue where some gamma x-rays where throwing
- Improvements in Mesh capabilities
- added mesh tally definitions to tallies
- store multi particle tally (for Volume and Surface)
- mcnp can write multi-particle tally
- Move check of tag_names to mesh.py
- Fix a problem of creating mesh from reading h5m files in unstructued R2S
- Default initializer pyne.mesh.Mesh() now raises an exception with info on how to properly make a mesh
- Move class MeshTally from mcnp.py to mesh.py
- Change the method of creating meshtally from mcnp meshtal
- pyne.mesh now takes advantage of PyMOAB instead of PyTAPS:
- IMeshTag changed to NativeMeshTag, with according tagetype name change: from 'imesh' to 'nat_mesh'
- write_hdf5(self, filename) -> write_hdf5(self, filename, write_mats)
- new save(self, filename, write_mats) (alias for write hdf5)
- new class MeshSetIterator()
- new get_tag(self, tag_name) and delete_tag(self, tag_name) methods
- when tagging the root set of a mesh, a new syntax is available:
- mymesh.mytag[mesh.mesh.getRootSet()] = val can now be written as mymesh.mytag[mymesh] = val
- direct call to the mesh entities change accordingly for example:
- getEntSets() -> get_entities_by_type( , )
- getTagHandle('XXX') -> tag_get_handle(types.XXXXX)
- iterate() -> mesh_iterate()
- getAllTags(xx) -> tag_get_tags_on_entity(xx)
- mesh.destroyTag(self, boolean) -> mesh.delete_tag(self)
- ... (see PyTAPS and PyMOAB respective documentation)
- those changes have been propagated in mcnp.py, alara.py, ccc.py, dagmc.pyx, r2s.py, variancereduction.py, expand_tags.py, and their respective tests...
Maintenance
- Documentation Changes
- Credit Rochman for allowing redistribute TENDL file
- Fix various typos
- automatic deployment of a updated version of the website on tags
- automatic creation of a new version of the website (not deployed) for
verification purposes in
pyne.github.com/website_preview - New developers guide: The update adds information about creating an environment, updates formatting for more consistency, details considerations and methods to check the version of dependencies, and adds additional links to coding resources.
- In website index, change C++ API link to "C++ API Documentation" instead of "C++ & Fortran API Documentation"
- Added publications to bibliography (PR #1256)
- Adding contributing guide and code of conduct (#1258)
- Changed Doc and Tutorial mentions of iPython notebooks to Jupyter notebooks (PR #1262)
- Improvements in documentation for release (#1290 #1285 #1299)
- Fix broken links (#1311)
- Update install instructions (#1238 & #1315)
- Update New Developers Guide
- Improvements in building and testing
- require contributor to change CHANGELOG
- stopped all testing in Python 2 (#1321)
- now get the base branch name from github and check change against it
- (inspired by https://github.com/NarrativeScience/circleci-orb-ghpr/blob/master/src/commands/get-pr-info.yml)
- Expand testing matrix to include:
- python 2 vs 3
- with vs without PyMOAB
- with vs without DAGMC
- Added FindDAGMC.cmake file
- turn off BLAS/LAPACK & FORTRAN in MOAB build
- Dockerfile to build many variations of PyNE docker image, with python script CLI
- Add hdf5-tools as dependency for docker images used in CircleCI, for better nose test comparing h5 files
- Add future as dependency for docker images used in CircleCI, for python2 and python3 compatibility
- "--dagmc" flag added to
setup.pyin order to build PyNE against DAGMC - new check won't now be triggered after a merge only on PRs
- utils.py: updated the download timeout time to 120sec (from 30sec)
- updated CI to use CircleCI 2.1 workflows: now build separately from tests with state saved between runs
- test_fluka:
- added test to check the data tag name of the different tally part and error.
- revert internal nuc_data_path to origin value after internal data test
- added DEFINE variable to allow material.cpp amalgamation without decay.cpp
- now skips endf test when website is not reachable to allow completeness of the other tests.
- test file for ENDF was wrong
- Add functions to do file, file block, line, and string almost the same compare functions in pyne/utils.py
- make data available as replacement for data.pyne.io (#1261, #1265)
- Removed iPython check from
setup.pyand added Jupyter to be an optional dependency in documentation (#1273) - Install Python dependencies with Pip instead of APT in Dockerfile
- Remove if block in travis-run-tests.sh (just run nosetests)
- Deprecating most of the python 2 tests. Only testing python2 with pyMOAB and DAGMC deps.
- updates tutorials and examples to python3 syntax to avoid failures (#1271)
- added a welcome bot for first PR, Issue, and merge (#1287)
- allow spatial solvers to be built in conda environments (#1305)
- ensure that ENSDF fortran tools are linked correctly. (#1306)
- allow spatial solvers to be turned on/off at time of setup (#1308)
- Code cleanup
- Formatting improvements
- Compatibility with language updates
- update the way that
collectionsis imported in preparation for deprecated changes in future python versions - removed some imports of
collectionsthat were not necessary - change return type of method to avoid compiler compatibility issue
- Convert some code and tests to enable python2/3 compatibility
- update the way that
- Clean up some hard coded strings in test_source_sampling.py
rxname.child()andrxname.parent()now acceptstrfor thezargument in Python 3.- dagmc_bridge: added a static DagMC instance
- cleanup throws return from
return (const char *)to simplereturn(it was suggested that those complicated return might cause seg fault, on some system -- OsX+conda) - Fixes string formatting typo in
pyne/fortranformat/_parser.py. Entire file is nowstr.format()(#1275) - Remove deprecated file open mode 'U' (#1298)
Fixed:
- Bug fix in xsdir class for parsing correct ZAID XS (#1314)
Added:
- Function to convert unit to s in pyne/alara.py
- Function to do float match for decay times
- Add SourceParticle class in pyne/src/source_sampling.
- Codes to read ALARA output file under subvoxel R2S condition
- A function to build up a subvoxel_array from mesh and cell_mats information
- A test function to test the process of reading ALARA output file
- Test function for subvoxel with (N, 1) condition in test_mesh.py
- Reshape the array when max_num_cells == 1
Changed:
- shape of IMeshTag when input value is a (N, 1) array
- set tag as array rather than number
- decaygen now gets the include dir based on the compiler path.
- Build system now explitily looks for C++11 standard compatability.
- Unit of e_bounds changed from eV to MeV
- Change loop variables to be v for volume elements and e for energy groups (instead of i & j)
- Use bias_mode instead of mode to allow for additional mode types in future
- A parameter in test_alara.py, to test modified match method
- Correct the wrong mode description comment in pyne/src/source_sampling.h
- Change the particle_birth return value from std::vectot<double> to SourceParticle object
- Some code clean up
- Some clean up of white space
Removed:
- Code in mesh.py to reshpe a (N,1) to (N, ) array is no longer needed if PR #971 merged
Fixed:
- decaygen now can properly produce Clang assembly.
- Build system would always download cram sources, even if they already existed. This has been fixed.
- ENDF error bounds bug that was preventing
nuc_data_makefrom working. - NNDC no longer provides the
mednew.datdata set. A fallback has been supplied.
Fixed:
- Made SSL context creation Python 2 & 3 Compatible.
Changed:
- Downloading files now uses null SSL context.
Changed:
- Downloading data now uses HTTP, rather than HTTPS.
Fixed:
- Occassional bug with downloading URL fix.