Hack for syncing local .shed.yml files with a Galaxy Tool Shed

make_shed_yml.py

#
# A hack, loosly based on Eric Rasche's disgusting.py
# https://gist.github.com/erasche/4ac3448b036f09979e14
#
# Intended as a one-off use script to help with syncing local
# .shed.yml files with a Galaxy Tool Shed. See also:
# https://gist.github.com/peterjc/5ebbf446d799f3aaa639

import yaml
import os
from bioblend import toolshed

shed_url = "https://toolshed.g2.bx.psu.edu/"
#owner = "peterjc"
#github_base = "https://github.com/peterjc/pico_galaxy/tools/"
#local_base = "/mnt/galaxy/repositories/pico_galaxy/tools/"
#owner = "devteam"
github_base = "https://github.com/peterjc/galaxy_blast/"
local_base = "/mnt/galaxy/repositories/galaxy_blast/"

#owner = "iuc"
#repo_type = "dependencies"

#owner = "peterjc"
#repo_type = "workflows"

owner = "peterjc"
repo_type = "tools"

github_base += repo_type + "/"
local_base += repo_type + "/"
assert os.path.isdir(local_base), local_base

tsi = toolshed.ToolShedInstance(shed_url)


repos = tsi.repositories.get_repositories()
repos = [r for r in repos if r["owner"]==owner]
cats = tsi.repositories.get_categories()

#github_base = "https://github.com/peterjc/pico_galaxy/tools/"
#local_base = "/mnt/galaxy/repositories/pico_galaxy/tools/"

keys = ("name", "owner",
        "homepage_url", "remote_repository_url",
        "description", "long_description")

# Ripped from planemo
def shed_repo_config(path):
    shed_yaml_path = os.path.join(path, ".shed.yml")
    if os.path.exists(shed_yaml_path):
        with open(shed_yaml_path, "r") as f:
            return yaml.load(f)
    else:
        return {}
 
 
def serialize_shed_repo_config(shed_yaml_path, data):
    with open(shed_yaml_path, "w") as f:
        for k in keys:
            if k in data:
                v = data[k]
                if "\n" in v:
                    v = yaml.dump(v).rstrip()
                    # Strip redundant quote chars...
                    if v.count("'") == 2 and v[0] == "'" and v[-1] == "'":
                        v = v[1:-1]
                assert "..." not in v, data[k]
                f.write("%s: %s\n" % (k, v))
        #f.write(yaml.dump(data, default_flow_style=False))

def sync_changes(name, shed_path, data):
    conf = shed_repo_config(shed_path)
    for k in ("name", "owner", "type",
              "homepage_url", "remote_repository_url",
              "description", "long_description"):
        if k in data and data[k] is not None:
            conf[k] = str(data[k])  # anti-unicode hack
    if 'long_description' in conf:
        ld = conf['long_description'].rstrip().split('\n')
        good = []
        for x in ld:
            if 'Repository-' in x:
                continue
            good.append(x.strip())
 
        lds = ''
        for x in good:
            if x != '':
                lds += x.strip() + '\n'
        conf['long_description'] = lds.strip()
 
    if 'homepage_url' not in conf:
        conf['homepage_url'] = github_base + name
    if 'remote_repository_url' not in conf:
        conf['remote_repository_url'] = github_base + name
    serialize_shed_repo_config(shed_path, conf)

for r in repos:
    name = r["name"]
    if not os.path.isdir(os.path.join(local_base, name)):
        continue
    shed = os.path.join(local_base, name, ".shed.yml")
    print(shed)

    info = tsi.repositories.show_repository(r["id"])

    if os.path.isfile(shed):
        print("%s exists" % shed)
        os.remove(shed)
    else:
        print("%s missing" % shed)
    sync_changes(name, shed, info)