#!/usr/bin/env python

import arvados
import md5
import subst
import subprocess
import os
import hashlib

p = arvados.current_job()['script_parameters']

merged = ""
src = []
for c in p["input"]:
    c = subst.do_substitution(p, c)
    i = c.find('/')
    if i == -1:
        src.append(c)
        merged += arvados.CollectionReader(c).manifest_text()
    else:
        src.append(c[0:i])
        cr = arvados.CollectionReader(c[0:i])
        j = c.rfind('/')
        stream = c[i+1:j]
        if stream == "":
            stream = "."
        fn = c[(j+1):]
        for s in cr.all_streams():
            if s.name() == stream:
                if fn in s.files():
                    merged += s.files()[fn].as_manifest()

crm = arvados.CollectionReader(merged)

combined = crm.manifest_text(strip=True)

m = hashlib.new('md5')
m.update(combined)

uuid = "{}+{}".format(m.hexdigest(), len(combined))

collection = arvados.api().collections().create(
    body={
        'uuid': uuid,
        'manifest_text': crm.manifest_text(),
    }).execute()

for s in src:
    l = arvados.api().links().create(body={
        "link": {
            "tail_uuid": s,
            "head_uuid": uuid,
            "link_class": "provenance",
            "name": "provided"
        }}).execute()

arvados.current_task().set_output(uuid)