Pyppl

- Set job signature to False if any of the item is False (that means expected files not exists); - Do cache by job itself;
- Make it possible to cache and export successful jobs even when some jobs failed
- Host docs in gitbook
- Init job with log func from proc;
- Add docstring for API generation;
- Redefine return code for outfile not generated;
- Error ignore works now;
- Rewrite runner_local so it fits other runners to extend;
- Fix proc depends on mixed list of procs and aggrs

- Fix local runner not waiting (continuiously submitting jobs);
- Add property alias for aggr;
- Output cleared if job not cached
- Fix bugs when export if outfiles are links;
- change default export method to move;
- add id and tag to calculate suffix for proc;
- add timer;
- add isRunning for job so that even if the main thread quit, we can still retrieve the job status;

- Add `pyppl-report` executable
- Use `diot` instead of `python-box`

- Raise informative exception while rendering markdown for a process.
- Adjust styles for default template.
- Remove unused code in template.js

- Add files (array) support for input;
- Recursive update for configuration;
- Add aggregations;
- Move functions to utils;
- Separate run for runners to submit and wait;
- Add use job class for jobs in a proc;
- Use "1,2 3,4" for channel.fromArgs for multi-width channels;
- Add rbind, cbind, slice for channel;
- Add alias for some proc properties;
- Remove callfront for proc;
- Add export cache mode;
- Add gzip export support (1);
- Unify loggers;
- Use job cache instead of proc cache so that a proc can be partly cached;
- Rewrite buildInput and buildOutput;
- Use job to construct runners;

- Add non-standalone mode for html reports
- Fix bug when disabling report generating report for individual process using `proc.report = False`