Chunkflow

Breaking Changes
- make `verbose` a integer rather than boolean number for variation of verbosity.

Features
- add setup-env operator to automatically compute patch number, cloud storage block size and other metadata. ingest tasks into AWS SQS queue. After this operation, you are ready to launch workers!
- support cropped output patch size for inference
- refactored normalize section contrast operator to make it faster and easier to use. We precompute a lookup table to avoid redundent computation of the voxel mapping.
- avoid creating a mask buffer by directly applying the high-mip mask to low-mip chunk

Bug Fixes
- fix a typo of thumbnail_mip

Improved Documentation
- add more complex production inference example

Breaking Changes
- remove c++ compilation module

Features
- neuroglancer operator works with multiple chunks (https://github.com/seung-lab/chunkflow/pull/123)
- add connected components operator
- add iterative mean agglomeration operator, including watershed computation.

Improved Documentation
- tutorial for cleft and cell boundary detection (https://github.com/seung-lab/chunkflow/pull/123)

Features
- a new operator called `save-images` to save chunk as a serials of PNG images.
- add log directory parameter for benchmark script.
Bug Fixes
- queue becomes None
- pznet was not working.
- fix Dockerfile for pznet build (not sure whether breaks PyTorch or not)

Improved Documentation
- add kubernetes documentation to find monitor of operations per second and bandwidth usage.

Breaking Changes
- change the mask command parameter `mask-mip` to `mip`

Bug Fixes
- fix log uploading and cloud watch
- fix the misplaced log in task, which will make the inference log not working
- fix blackout section bug. Previous implementation will have indexing error while blacking out z outside of chunk.
Improved Documentation
- clarify patch, chunk and block.
- add realistic example for ConvNet inference.

Breaking Changes
- merge operator upload-log to save since we normally just put log inside saved volume.

Features
- add batch size option for ConvNet inference
- better memory footprint by deleting the ascontiguousarray
Bug Fixes
- the inference framework was hard coded as identity, fixed this bug.

Improved Documentation
- updated the documentation to focus on chunk operators, not just ConvNet Inference.