eukrhythmic-assembly

configfile: "config.yaml"

import io
import os
from os import listdir
from os.path import isfile, join
import sys
import pandas as pd
import numpy as np
import pathlib
from snakemake.exceptions import print_exception, WorkflowError  
sys.path.insert(1, 'scripts')

# Import relevant variable from config; exit if not supplied
import importworkspace
from importworkspace import *
os.system("python " + os.path.join("scripts", "importworkspace.py"))

# Contains function to check that variables are present and formatted correctly beyond provided values in config file
import checkrequirements
from checkrequirements import *

## CHECK THAT ALL REQUIREMENTS ARE SATISFIED BY EXECUTING checkrequirements() from `scripts/checkrequirements.py` ##
checkrequirementsfct()

include: "rules/trinity-snake.smk"
include: "rules/velvet-snake.smk"
include: "rules/megahit-snake.smk"
include: "rules/spades-snake.smk"
include: "rules/transabyss-snake.smk"
include: "rules/transabyss-merge.smk"
include: "rules/quast-snake.smk"

ruleorder: trinity > trinity_SE
ruleorder: megahit > megahit_SE
ruleorder: velvet > velvet_SE

rule all:
    input:
        # TRIMMOMATIC OUTPUTS
        trimmed = expand([os.path.join("{base}", "firsttrim", "{sample}_1.trimmed.fastq.gz"), 
                          os.path.join("{base}", "firsttrim", "{sample}_2.trimmed.fastq.gz")], zip, 
                         base = OUTPUTDIR, sample = filenames),
        # ASSEMBLER OUTPUTS
        assemblersout = expand(os.path.join("{base}", "{assembly}_{assembler}.fasta"), 
                               base = ASSEMBLEDDIR, assembly = assemblygroups, assembler = ASSEMBLERS), 
        # QUAST OUTPUTS
        quast = expand(os.path.join("{base}", "quast", "{assembly}"), base = OUTPUTDIR, assembly = assemblygroups),
        # COMBINE QUAST OUTPUTS
        quastcombine = expand(os.path.join("{base}", "quast", "fullresults", "allresults.tsv"), base = OUTPUTDIR),