vdbwrair / bactpipeline

Bacteria Pipeline modified from MRSN Bacteria Pipeline

bactpipeline

Bacteria Pipeline that utilizes flash, btrim and Newbler(gsAssembler)

Copyright and License

This project is licensed under the GPLv2.

You can view the license here and the copyright here

bactpipeline's Issues

Master summary

Need a way to concat all individual project summary csv files as well as contig files

Make project public

• Make newbler external dependency
• Add license(flash is GNU, btrim has no license?)
• add sphinx RTD
• add travis

more descriptive args in docstrings

Just a note that a contracts-compatible alternative to the :type tag is the :param tag which can incude a doc of the arg as well.

@contract
def write_summary(data, outfile, delim='\t'): # data is 2d list
    ''' write data as a csv/tsv file to outfile.
    :type data:    Iterable
  etc."""

could be written as:

@contract
def write_summary(data, outfile, delim='\t'): # data is 2d list
    ''' write data as a csv/tsv file to outfile.
    :param   Iterable data: some docs
    :param str,!exists outfile: more docs ..
    etc '''

remove newbler from previous commits

before going public we will need to remove newbler from dependencies from all existing commits

Whatever that crazy git command is

Malformed CSV should cause error

csv.DictReader will happily take a malformed csv like
ASampledir,AName,APrimer
which is missing its header and simply return an empty iterable . . . dubious design choice.

bactpipeline/bactpipeline/runsample.py

Lines 61 to 62 in c4fb29d

	sheet = csv.DictReader(open(f))
	results = map(_run_sample, sheet)

runsample -s samplesheet docs

Right now analysis.rst is broken down by Running single sample but the running sample sheet docs need to be a heading on their own

Make Contract messages more user-friendly

e.g.

contracts.interface.ContractNotRespected: Breach for argument 'outdir' to run_sample_sheet().
Shouldn't have satisfied the clause exists.
checking: !exists for value: Instance of <type 'str'>: './'
checking: str,!exists for value: Instance of <type 'str'>: './'
Variables bound in inner context:

Could become

contracts.interface.ContractNotRespected: Breach for argument 'outdir' to run_sample_sheet().
=====================
ERROR:  blah blah blah
======================
checking: !exists for value: Instance of <type 'str'>: './'
checking: str,!exists for value: Instance of <type 'str'>: './'
Variables bound in inner context:

Do this by raising a ValueError rather than returning false. Example:

def fails(x):
    raise ValueError("="*9 + "\nERROR: blah blah blah, do xyz\n" + "="*9)
contracts.new_contract("failit", fails)
contracts.decorate(lambda x: x, x="failit")(1)

I would just wrap our current boolean-returning functions

def wrap_validation(func, message):
   def ret_func(arg):
      if not func(arg): 
         raise ValueError(message)
   return ret_func

contracts.new_contract("is_truthy", wrap_validation(lambda x: bool(x), "Not truthy, use truthy input"))

ouptut directory should be allowed to exist

ouptut directory should not be required to not exist, as it's making things harder for users

related: https://vdbpm.org/issues/10846

offender

change setup?

My changes require BioPython so the setup file could change and the modules file at VDBWRAIR should change

Analysis should not start unless newAssembly and runAssembly are in path

Don't want to have analysis start and get through flash and btrim and then die because they don't exist.

Should tell user to make sure Roche Analysis apps are installed and then try again

Summary report

comma or tab separated

sample_id,contig_num,num_reads,%total_reads,N50

primer trimming

Supply primer fasta file that will trim primers

I believe newbler supports supplying primer fasta file to be used to trim

contig fasta file

Example

>sample_id__contig_num numreads=X