Pyppl

- Config file defaults to `~/.pyppl.json` (`~/.pyppl` also works)
- Callfront added
- Empty input allowed
- Same basename name allowed for input files of a job
- Description of a proc added
- Aggr Optimized
- Bug 9 Fixed
- Private key supported for ssh runner
- Feature 7 Implemented

- Optimize isRunning function (using specific job id)
- Support python3 now
- Test on OSX
- More debug information for caching
- Bug fixes

- Move pyppl-cli to bin/pyppl
- channel.collapse now return the most common directory of paths
- Report oringinal file of input and bring files
- Show number of omitted logs
- Bug fixes

- Add colored log
- Put jobs in different directories (files with same basename can be used as input files, otherwise it will be overwritten).
- Add configuration `checkrun` for `pyppl` allow `runner.isRunning` to be disabled (save resources on local machine).
- Add built-in functions for placeholders; lambda functions do not need to call (just define)
- File placeholders (.fn, .bn, .prefix, etc) removed, please use built-in functions instead.
- Add an alias `p.ppldir` for `p.tmpdir` to avoid confusion.
- Update command line tool accordingly
- Split base runner class into two.

- Update docs and fix compilation errors from gitbook
- Change pyppl.dot to pyppl.pyppl.dot;
- Add channel.fromFile method;
- Add aggr.addProc method;
- Fix proc/aggr copy bugs;
- Fix utils.varname bugs;
- Fix bugs: channel._tuplize does not change list to tuple any more.
- Add fold/unfold to channel;
- cache job immediately after it's done;
- remove proc in nexts of its depends when its depends are reset;
- add dir for input files, prefix for output files;
- Fix utilfs.dirmtime if file not exists;
- add pyppl-cli;
- Change rc code, make it consistent with real rc code.

- Overwrite input file if it exists and not the same file;
- fix varname bug when there are dots in the function name;
- Add brings feature;
- Add features to README, and brings to docs