import os
import sys
import tempfile

from common import get_one  # noqa:I100,I201

from bioblend.galaxy.objects import GalaxyInstance

# This is a "toy" example that should run much faster
# (once the cluster's resource manager allows it to run) than the
# real-world ones.  The workflow, which is imported from local disk,
# removes two header lines from a tabular file, then extracts one of
# the columns.  The input dataset is publicly available on CRS4's
# Orione Galaxy server.

URL = "https://orione.crs4.it"
API_KEY = os.getenv("GALAXY_API_KEY", "YOUR_API_KEY")
if API_KEY == "YOUR_API_KEY":
    sys.exit("API_KEY not set, see the README.txt file")
gi = GalaxyInstance(URL, API_KEY)

# import the workflow from the JSON dump

with open("small.ga") as f:
    wf = gi.workflows.import_new(f.read())

# Select the "Orione SupMat" library

library_name = "Orione SupMat"
l = get_one(gi.libraries.list(name=library_name))

# Select the input dataset

ds_name = "/RNA-Seq - Listeria monocytogenes/Listeria_monocytogenes_EGD_e_uid61583/NC_003210.rnt"
ld = get_one(l.get_datasets(name=ds_name))
input_map = {"input_tsv": ld}

# Run the workflow on a new history with the selected dataset as
# input, overriding the index of the column to remove; wait until the
# computation is complete.

history_name = "get_col output"
params = {"Cut1": {"columnList": "c2"}}
print(f"Running workflow: {wf.name} [{wf.id}]")
outputs, out_hist = wf.run(input_map, history_name, params=params, wait=True)
print("Job has finished")
assert out_hist.name == history_name
print(f"Output history: {out_hist.name} [{out_hist.id}]")

# Save results to local disk
out_ds = get_one([_ for _ in outputs if _.name == "Cut on data 1"])
with tempfile.NamedTemporaryFile(prefix="bioblend_", delete=False) as f:
    out_ds.download(f)
print(f'Output downloaded to "{f.name}"')