ammaraziz

My thoughts on workflow execution services (WES) from previous discussion. Note: I'm not all that crash hot at setting up web servers/infrastructure.

    Nextflow tower/Seqera platform/whatever it's called now

    Cost: ~70-100k per year from what I hear - do not quote me. edit: if you are in australia you can access this service for free/reduce price.
    Documentation: Good docs, easy to follow albeit the install instructions seem to be scattered (could be just the service provider I was using)
    Setup: Smooth, has a few dependencies but they're relatively easy to install on a VM machine for testing purposes (docker, slurm). Java gave me some headaches due to the weird VM I had setup. But once compiled the tower agent, it worked fine.
    Workflow engines supported: Nextflow only.
    Comments: Supports on-prem hardware which is really cool (needs SLURM), can specify cloud storage or data stored locally. Web interface relies on the nextflow.config being properly configured (nf-core guidelines work ~95% 

ammaraziz

import re
from pathlib import Path
import argparse

parser = argparse.ArgumentParser(
    description='Convert BioEdit Project binary file to fasta',
    prog='bio2fasta',
    formatter_class=argparse.ArgumentDefaultsHelpFormatter
    )

ammaraziz

"""
    This code pulls data from the WHO's influenza surveillance database:

        https://apps.who.int/flumart/Default?ReportNo=12

    This website is pretty tricky to parse; you must pass realistic headers to the POST requests, and you must also
    issue 3 total requests: 1) a GET request, 2) a POST request, and 3) another POST request. All 3 of these requests,
    in order, are required to actually collect the underlying data that's displayed in the table. See `get_table_data`
    for more documentation on this process.

ammaraziz

import argparse
import textwrap
import os
import sys
from datetime import timedelta, datetime


# function for reading a multifasta file
# returns a dictionary with sequence headers and nucleotide sequences
def get_seqs_from_fasta(filepath):

ammaraziz

mle <- function(data,start=NULL,eps=10**(-9),logit=FALSE,cov=NULL){
  # INPUT:
  # data: data matrix with
  # 1) observed outcome D (0: no; 1: yes)
  # 2) observed types at risk for (0: no; 1: yes)
  # start: starting vector for the algorithm
  # eps: accuracy parameter to stop the algorithm
  # logit: logical to indicate whether the logit model is used
  # (i.e. correct for a covariate)
  # cov: covariate vector to correct for (only if logit=TRUE)