v0.1.0a64

• Drop Python 3.5. #476
• Drop python-coveralls on Wercker. #477
• Switch from memory_info_ex to memory_info. #473
• Drop setuptools install requirement. #479
• Drop the Matplotlib-based viewer. #480
• Only install docstring-coverage on doc deploy. #481

v0.1.0a63

• Support Python 3.7. #475

v0.1.0a62

• Pin breaking dependencies. #470
• Fix tifffile 0.13 breakage. #471
• Fix sphinx 1.7 breakage. #472

v0.1.0a61

• Bump license year to 2018. #467
• Close other temp file in HDF5 tests. #468

v0.1.0a60

• Upgrade to versioneer 0.18. #457
• Simplify mean computation code. #459
• Drop docstring-coverage from Wercker. #462
• Require Travis CI's Ubuntu Precise image. #463
• Fixes for NumPy 1.14. #464

v0.1.0a59

• Run CI with SPAMS 2.6 on Python 2/3. #450
• Drop old viewer config. #449
• Update files for coverage to ignore. #451
• Support coverage 4. #452
• Adjust some tests for NumPy 1.13 compatibility. #455
• Use Python 3.5 environment to submit coverage. #456
• Drop HDF5 pin. #453

v0.1.0a58

• Package license file as data. #443, #447
• Try tweaking registration algorithm. #444, #446
• Require tifffile. #445
• Delete .gitmodules file. #448

v0.1.0a57

• Unpin conda-build. #434
• Support NumPy 1.12. #440
• Support Python 3.6. #441
• Add 2017 to the license copyright. #442

v0.1.0a56

• Drop remaining Sphinx pins. #431
• Drop functools32 as a build requirement. #432
• Use mplview in place of the included viewer. #433
• Update viewer module's deprecation warning. #433

v0.1.0a55

• Use split_blocks from kenjutsu. #426
• Use future for Python 2/3 compatibility. #427
• Use enumerate_masks_max where possible for speed. #428
• Fix documentation and a bug with generate_contour_fast. #429
• Use imgroi for some functionality. #430