from util.decode import decode, decode_to_output, exclude_list, get_exclude_list, truncate_sents
from model.postprocess import PostProcess
from nltk.translate.bleu_score import sentence_bleu
from util.mteval_bleu import MtEval_BLEU
import tensorflow as tf
from os.path import exists
from os import remove, listdir, makedirs
import math
import numpy as np
import time
from util.arguments import get_args
from copy import deepcopy
from model.model_config import get_path
args = get_args()
def get_graph_val_data(objs, model_config, it, data):
input_feed = {}
# Reserved section of vocabuary are same.
voc = data.vocab_simple
if model_config.subword_vocab_size > 0:
pad_id = voc.encode(constant.SYMBOL_PAD)
else:
pad_id = [voc.encode(constant.SYMBOL_PAD)]
output_tmp_sentence_simple, output_tmp_sentence_complex, \
output_tmp_sentence_complex_raw, output_tmp_sentence_complex_raw_lines, \
output_tmp_mapper, output_tmp_ref_raw_lines = [], [], [], [], [], []
--output_directory 'processed/bam' \
--walltime '06:00:00' \
--nodes 2 \
--cores 12 \
--humanonly
The specific pipeline using this script is given in `wes_pipeline.sh`
"""
import os
import yaml
import util.arguments as arguments
# Load command arguments
args = arguments.get_args()
command = args.which
genome = args.genome
input_dir = args.input_directory
output_dir = args.output_directory
config = args.config_yaml
walltime = args.walltime
nodes = str(args.nodes)
cores = str(args.cores)
# Load configuration
with open(config, 'r') as stream:
config = yaml.load(stream)
# Load constants
python = config['python']