X-Git-Url: https://git.arvados.org/arvados.git/blobdiff_plain/c59093180fd92f0d7c6607a49458446212ebd058..b40d2110e1fe102ea687252cf6b0f48a6397f11e:/sdk/python/arvados/commands/arv_copy.py diff --git a/sdk/python/arvados/commands/arv_copy.py b/sdk/python/arvados/commands/arv_copy.py index c64b645a38..a10eb2b348 100755 --- a/sdk/python/arvados/commands/arv_copy.py +++ b/sdk/python/arvados/commands/arv_copy.py @@ -32,6 +32,8 @@ import arvados.util import arvados.commands._util as arv_cmd import arvados.commands.keepdocker +from arvados.api import OrderedJsonModel + logger = logging.getLogger('arvados.arv-copy') # local_repo_dir records which git repositories from the Arvados source @@ -69,10 +71,10 @@ def main(): help='Perform copy even if the object appears to exist at the remote destination.') copy_opts.add_argument( '--src', dest='source_arvados', required=True, - help='The name of the source Arvados instance (required). May be either a pathname to a config file, or the basename of a file in $HOME/.config/arvados/instance_name.conf.') + help='The name of the source Arvados instance (required) - points at an Arvados config file. May be either a pathname to a config file, or (for example) "foo" as shorthand for $HOME/.config/arvados/foo.conf.') copy_opts.add_argument( '--dst', dest='destination_arvados', required=True, - help='The name of the destination Arvados instance (required). May be either a pathname to a config file, or the basename of a file in $HOME/.config/arvados/instance_name.conf.') + help='The name of the destination Arvados instance (required) - points at an Arvados config file. May be either a pathname to a config file, or (for example) "foo" as shorthand for $HOME/.config/arvados/foo.conf.') copy_opts.add_argument( '--recursive', dest='recursive', action='store_true', help='Recursively copy any dependencies for this object. (default)') @@ -180,11 +182,19 @@ def api_for_instance(instance_name): client = arvados.api('v1', host=cfg['ARVADOS_API_HOST'], token=cfg['ARVADOS_API_TOKEN'], - insecure=api_is_insecure) + insecure=api_is_insecure, + model=OrderedJsonModel()) else: abort('need ARVADOS_API_HOST and ARVADOS_API_TOKEN for {}'.format(instance_name)) return client +# Check if git is available +def check_git_availability(): + try: + arvados.util.run_command(['git', '--help']) + except Exception: + abort('git command is not available. Please ensure git is installed.') + # copy_pipeline_instance(pi_uuid, src, dst, args) # # Copies a pipeline instance identified by pi_uuid from src to dst. @@ -209,6 +219,8 @@ def copy_pipeline_instance(pi_uuid, src, dst, args): pi = src.pipeline_instances().get(uuid=pi_uuid).execute(num_retries=args.retries) if args.recursive: + check_git_availability() + if not args.dst_git_repo: abort('--dst-git-repo is required when copying a pipeline recursively.') # Copy the pipeline template and save the copied template. @@ -262,6 +274,8 @@ def copy_pipeline_template(pt_uuid, src, dst, args): pt = src.pipeline_templates().get(uuid=pt_uuid).execute(num_retries=args.retries) if args.recursive: + check_git_availability() + if not args.dst_git_repo: abort('--dst-git-repo is required when copying a pipeline recursively.') # Copy input collections, docker images and git repos. @@ -315,10 +329,11 @@ def copy_collections(obj, src, dst, args): obj = arvados.util.portable_data_hash_pattern.sub(copy_collection_fn, obj) obj = arvados.util.collection_uuid_pattern.sub(copy_collection_fn, obj) return obj - elif type(obj) == dict: - return {v: copy_collections(obj[v], src, dst, args) for v in obj} - elif type(obj) == list: - return [copy_collections(v, src, dst, args) for v in obj] + elif isinstance(obj, dict): + return type(obj)((v, copy_collections(obj[v], src, dst, args)) + for v in obj) + elif isinstance(obj, list): + return type(obj)(copy_collections(v, src, dst, args) for v in obj) return obj def migrate_jobspec(jobspec, src, dst, dst_repo, args): @@ -533,31 +548,30 @@ def copy_collection(obj_uuid, src, dst, args): else: progress_writer = None - for line in manifest.splitlines(True): + for line in manifest.splitlines(): words = line.split() - dst_manifest_line = words[0] + dst_manifest += words[0] for word in words[1:]: try: loc = arvados.KeepLocator(word) - blockhash = loc.md5sum - # copy this block if we haven't seen it before - # (otherwise, just reuse the existing dst_locator) - if blockhash not in dst_locators: - logger.debug("Copying block %s (%s bytes)", blockhash, loc.size) - if progress_writer: - progress_writer.report(obj_uuid, bytes_written, bytes_expected) - data = src_keep.get(word) - dst_locator = dst_keep.put(data) - dst_locators[blockhash] = dst_locator - bytes_written += loc.size - dst_manifest_line += ' ' + dst_locators[blockhash] except ValueError: # If 'word' can't be parsed as a locator, # presume it's a filename. - dst_manifest_line += ' ' + word - dst_manifest += dst_manifest_line - if line.endswith("\n"): - dst_manifest += "\n" + dst_manifest += ' ' + word + continue + blockhash = loc.md5sum + # copy this block if we haven't seen it before + # (otherwise, just reuse the existing dst_locator) + if blockhash not in dst_locators: + logger.debug("Copying block %s (%s bytes)", blockhash, loc.size) + if progress_writer: + progress_writer.report(obj_uuid, bytes_written, bytes_expected) + data = src_keep.get(word) + dst_locator = dst_keep.put(data) + dst_locators[blockhash] = dst_locator + bytes_written += loc.size + dst_manifest += ' ' + dst_locators[blockhash] + dst_manifest += "\n" if progress_writer: progress_writer.report(obj_uuid, bytes_written, bytes_expected) @@ -566,7 +580,6 @@ def copy_collection(obj_uuid, src, dst, args): # Copy the manifest and save the collection. logger.debug('saving %s with manifest: <%s>', obj_uuid, dst_manifest) - dst_keep.put(dst_manifest.encode('utf-8')) c['manifest_text'] = dst_manifest return create_collection_from(c, src, dst, args)