Pyppl

- Change the default path of flowchart and log from script directory to current directory.
- Rewrite the way of attribute setting for aggregations.
- Introduce modules for aggregations.
- Allow setting attributes from Proc constructor.
- Implement 33, `Ctrl-c` now also halts the pipeline, and hides the exception details.

- Fix runner name issue 31.
- Use mkdocs to generate documentations and host them on GitHub pages.
- Keep stdout and stderr when a job is cached: 30.
- Allow command line arguments to overwrite Parameter's type.
- Host the testing procedures with Travis.
- Fix other bugs.

- Auto-delegate common proc config names to aggr
- Add proc.origin to save the original proc id for copied procs
- Remove brings, add proc.infile to swith '{{in.(infile)}}' to job.indir path, original path or realpath of the input file
- Add process lock to avoid the same processes run simultaneously
- Use built-in Box instead of box.Box
- Merge template function Rvec and Rlist into R, add repr
- Fix 29 and 31, and fix other bugs

- Add proc.dirsig to disable/enable calculating signatures from deep directories
- Add Jobmgr class to handle job distribution
- Allow channel.rowAt and colAt to return multiple rows and columns, respectively
- Allow empty channel as input (process will be skipped)
- Refine all tests
- Rewrite thread-safe file system helper functions
- Add specific exception classes
- Report line when error happens in template
- Add progress bar for jobs
- Allow stdout and stderr file as output

- Add yaml support for config file (26).
- Allow empty list for input files.
- Merge continuous job ids in log (Make the list shorter).
- More details when internal template failed to render (25)
- Ignore .yaml config files if yaml module is not installed.
- sleep before testing isRunning to avoid all jobs running it at the same time.
- Use repr to output p.args and p.props.
- Merge Proc attributes profile and runner. Profile is now an alias of runner, and will be removed finally.

- Beautify parameters help page.
- Enable multithreading for job construction and cache checking (set by proc.nthread).
- Uniform multiprocessing/threading.
- Fix Aggr delegate problems.
- Add ProcTree to manage process relations.
- Report processes will not run due to prior processes not ran.
- Add cclean option for enable/disable cleanup (output check/export) if a job is cached.
- Add tooltip for flowchart svg.
- Fix job id not saved for runner_sge.
- Fix resume assignment issue.
- Rewrite proc.log function so that logs from jobs do not mess up when they are multithreaded.
- Fix params loaded from file get overwriten.
- Add coverage report.