[TOC]

This library uses yaml as a config file, to generate shell command auto. It is not a tool to excute functions, just combine them together

• It forces user to follow the rule defined in yaml config, which performs high modularity and low cohension between modules, so more readable and cheaper to re-use code to other works.

• It provides a more humanable way to process params transimission to other scripts user-defined

• It provides a more simple way to process input files, by given a csv file with format user define, but not dirs or paths which is not robust.

• More complex than pure shell scripts, but may be overcome by configs defined by author

This workflow consists of three subcommands, "generate", "cmd", "stat"

generate a config file to the path you set with -o, and current workdir if without -o params:

-h print help info
-n choose the config you want to generate, without suffix.
-o the outdir

blank: a blank templete for create your own configs
chipseq: a config file to process chipseq pipeline, by bowtie as mapping tools
rnaseq: a config file to process chipseq pipeline, by STAR as mapping tools wgbs_pe: a config file to process paired WGBS, and under testing

part to generate shell command files with config file input.
Two params are request: -c and -bf.
-c means config file in yaml format, but json5 may also supported, though not recommended
-b means input seq info by hand, which should followd by -f and filename

-a means search files auto, by given path, but not recommend, because little filename is supported and contradict with mine pupose$

part to run second part of config file, with -c points out the path of config file

this part may fused into cmd part in future, it's an unstable version

For the first, it's a yaml file, a convient format for write config by hand. it has the items below:

seq_info_format: define the input seqpair file in a list, with the first item is the delim char, and others are columns names, and a column called project should be included. by the way, config_id will be set as DEFAULT if absent in here
seq_order: order in the seq pair file you input by -bf, just use numbers as blank config
config_ids: the config ids will be process in this config file. Multi config id support is under development

DEFAULT: the part you define how to process, you may create other names

cmd_fusion_order: how to fusion the order together, by the cmd_order set in wrokflow part cmd_name: the output cmd txt name log_name: useless, but keep it by history reason outdir: all outputs will appear in this dir dataname: name to store run infomation, a yaml format file order: order to process workflow, each step should be contained here, otherwise they will be skipped order_stat: order to process stat part, similar as order item.
workflow: define the detail parts you want to process liki below:

partname:
outdir: hello #output of this part, default is the partname path: path of the program you want to excute in this part cmd_order: the symbol defined in cmd_fusion_order
father: dict, define as [name in this part:[from which part, how it called in the source part outputs]]
variables: define variables, use {} to quote variables get from father variables_eval: complex ones, can process strings by functions defined in basic_functions.py. functions: functions should be excute when the script run, but not the command file run params: a key part, define the real body in the outputfile, dict format, with a speical key blank, whose arrributes is a list and will be placed last in the command. NOTICE: \n is needed if you want to use multi lines. outparams: params you want to transfer to other parts

Any problems, contact me.