X-Git-Url: https://git.arvados.org/arvados.git/blobdiff_plain/bd86112ffba029e0f2c29c7d12ae6192e4097e18..441ef97e93a951b349356df96d8a6ef604c6cab7:/sdk/python/arvados/util.py diff --git a/sdk/python/arvados/util.py b/sdk/python/arvados/util.py index 0889ffcddd..3f50553c4b 100644 --- a/sdk/python/arvados/util.py +++ b/sdk/python/arvados/util.py @@ -1,20 +1,33 @@ import fcntl import hashlib +import httplib2 import os import re import subprocess import errno import sys -from arvados.collection import * + +import arvados +from arvados.collection import CollectionReader HEX_RE = re.compile(r'^[0-9a-fA-F]+$') +keep_locator_pattern = re.compile(r'[0-9a-f]{32}\+\d+(\+\S+)*') +signed_locator_pattern = re.compile(r'[0-9a-f]{32}\+\d+(\+\S+)*\+A\S+(\+\S+)*') +portable_data_hash_pattern = re.compile(r'[0-9a-f]{32}\+\d+') +uuid_pattern = re.compile(r'[a-z0-9]{5}-[a-z0-9]{5}-[a-z0-9]{15}') +collection_uuid_pattern = re.compile(r'[a-z0-9]{5}-4zz18-[a-z0-9]{15}') +group_uuid_pattern = re.compile(r'[a-z0-9]{5}-j7d0g-[a-z0-9]{15}') +user_uuid_pattern = re.compile(r'[a-z0-9]{5}-tpzed-[a-z0-9]{15}') +link_uuid_pattern = re.compile(r'[a-z0-9]{5}-o0j2j-[a-z0-9]{15}') +manifest_pattern = re.compile(r'((\S+)( +[a-f0-9]{32}(\+\d+)(\+\S+)*)+( +\d+:\d+:\S+)+$)+', flags=re.MULTILINE) + def clear_tmpdir(path=None): """ Ensure the given directory (or TASK_TMPDIR if none given) exists and is empty. """ - if path == None: + if path is None: path = arvados.current_task().tmpdir if os.path.exists(path): p = subprocess.Popen(['rm', '-rf', path]) @@ -32,7 +45,7 @@ def run_command(execargs, **kwargs): p = subprocess.Popen(execargs, **kwargs) stdoutdata, stderrdata = p.communicate(None) if p.returncode != 0: - raise errors.CommandFailedError( + raise arvados.errors.CommandFailedError( "run_command %s exit %d:\n%s" % (execargs, p.returncode, stderrdata)) return stdoutdata, stderrdata @@ -97,7 +110,7 @@ def tarball_extract(tarball, path): elif re.search('\.tar$', f.name()): p = tar_extractor(path, '') else: - raise errors.AssertionError( + raise arvados.errors.AssertionError( "tarball_extract cannot handle filename %s" % f.name()) while True: buf = f.read(2**20) @@ -108,10 +121,10 @@ def tarball_extract(tarball, path): p.wait() if p.returncode != 0: lockfile.close() - raise errors.CommandFailedError( + raise arvados.errors.CommandFailedError( "tar exited %d" % p.returncode) os.symlink(tarball, os.path.join(path, '.locator')) - tld_extracts = filter(lambda f: f != '.locator', os.listdir(path)) + tld_extracts = [f for f in os.listdir(path) if f != '.locator'] lockfile.close() if len(tld_extracts) == 1: return os.path.join(path, tld_extracts[0]) @@ -152,7 +165,7 @@ def zipball_extract(zipball, path): for f in CollectionReader(zipball).all_files(): if not re.search('\.zip$', f.name()): - raise errors.NotImplementedError( + raise arvados.errors.NotImplementedError( "zipball_extract cannot handle filename %s" % f.name()) zip_filename = os.path.join(path, os.path.basename(f.name())) zip_file = open(zip_filename, 'wb') @@ -173,11 +186,11 @@ def zipball_extract(zipball, path): p.wait() if p.returncode != 0: lockfile.close() - raise errors.CommandFailedError( + raise arvados.errors.CommandFailedError( "unzip exited %d" % p.returncode) os.unlink(zip_filename) os.symlink(zipball, os.path.join(path, '.locator')) - tld_extracts = filter(lambda f: f != '.locator', os.listdir(path)) + tld_extracts = [f for f in os.listdir(path) if f != '.locator'] lockfile.close() if len(tld_extracts) == 1: return os.path.join(path, tld_extracts[0]) @@ -237,7 +250,7 @@ def collection_extract(collection, path, files=[], decompress=True): outfile.write(buf) outfile.close() if len(files_got) < len(files): - raise errors.AssertionError( + raise arvados.errors.AssertionError( "Wanted files %s but only got %s from %s" % (files, files_got, [z.name() for z in CollectionReader(collection).all_files()])) @@ -292,19 +305,37 @@ def stream_extract(stream, path, files=[], decompress=True): outfile.write(buf) outfile.close() if len(files_got) < len(files): - raise errors.AssertionError( + raise arvados.errors.AssertionError( "Wanted files %s but only got %s from %s" % (files, files_got, [z.name() for z in stream.all_files()])) lockfile.close() return path -def listdir_recursive(dirname, base=None): +def listdir_recursive(dirname, base=None, max_depth=None): + """listdir_recursive(dirname, base, max_depth) + + Return a list of file and directory names found under dirname. + + If base is not None, prepend "{base}/" to each returned name. + + If max_depth is None, descend into directories and return only the + names of files found in the directory tree. + + If max_depth is a non-negative integer, stop descending into + directories at the given depth, and at that point return directory + names instead. + + If max_depth==0 (and base is None) this is equivalent to + sorted(os.listdir(dirname)). + """ allfiles = [] for ent in sorted(os.listdir(dirname)): ent_path = os.path.join(dirname, ent) ent_base = os.path.join(base, ent) if base else ent - if os.path.isdir(ent_path): - allfiles += listdir_recursive(ent_path, ent_base) + if os.path.isdir(ent_path) and max_depth != 0: + allfiles += listdir_recursive( + ent_path, base=ent_base, + max_depth=(max_depth-1 if max_depth else None)) else: allfiles += [ent_base] return allfiles @@ -321,8 +352,8 @@ def is_hex(s, *length_args): """ num_length_args = len(length_args) if num_length_args > 2: - raise ArgumentError("is_hex accepts up to 3 arguments ({} given)". - format(1 + num_length_args)) + raise arvados.errors.ArgumentError( + "is_hex accepts up to 3 arguments ({} given)".format(1 + num_length_args)) elif num_length_args == 2: good_len = (length_args[0] <= len(s) <= length_args[1]) elif num_length_args == 1: @@ -331,13 +362,34 @@ def is_hex(s, *length_args): good_len = True return bool(good_len and HEX_RE.match(s)) -def list_all(fn, **kwargs): +def list_all(fn, num_retries=0, **kwargs): + # Default limit to (effectively) api server's MAX_LIMIT + kwargs.setdefault('limit', sys.maxsize) items = [] offset = 0 - items_available = sys.maxint + items_available = sys.maxsize while len(items) < items_available: - c = fn(offset=offset, **kwargs).execute() + c = fn(offset=offset, **kwargs).execute(num_retries=num_retries) items += c['items'] items_available = c['items_available'] offset = c['offset'] + len(c['items']) return items + +def ca_certs_path(fallback=httplib2.CA_CERTS): + """Return the path of the best available CA certs source. + + This function searches for various distribution sources of CA + certificates, and returns the first it finds. If it doesn't find any, + it returns the value of `fallback` (httplib2's CA certs by default). + """ + for ca_certs_path in [ + # Arvados specific: + '/etc/arvados/ca-certificates.crt', + # Debian: + '/etc/ssl/certs/ca-certificates.crt', + # Red Hat: + '/etc/pki/tls/certs/ca-bundle.crt', + ]: + if os.path.exists(ca_certs_path): + return ca_certs_path + return fallback