3826: Merge branch 'master' into 3826-crunchstat-netstats

[arvados.git] / sdk / python / arvados / util.py

diff --git a/sdk/python/arvados/util.py b/sdk/python/arvados/util.py
index 7148b9295b354d37cc941beb92a570e68720117e..2609f112fe31d050a5a1db094e32ca8087774c5c 100644 (file)
--- a/sdk/python/arvados/util.py
+++ b/sdk/python/arvados/util.py
@@ -7,6 +7,15 @@ import errno
 import sys
 from arvados.collection import *
 
+HEX_RE = re.compile(r'^[0-9a-fA-F]+$')
+
+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}')
+
 def clear_tmpdir(path=None):
     """
     Ensure the given directory (or TASK_TMPDIR if none given)
@@ -160,7 +169,7 @@ def zipball_extract(zipball, path):
                     break
                 zip_file.write(buf)
             zip_file.close()
-            
+
             p = subprocess.Popen(["unzip",
                                   "-q", "-o",
                                   "-d", path,
@@ -306,3 +315,36 @@ def listdir_recursive(dirname, base=None):
         else:
             allfiles += [ent_base]
     return allfiles
+
+def is_hex(s, *length_args):
+    """is_hex(s[, length[, max_length]]) -> boolean
+
+    Return True if s is a string of hexadecimal digits.
+    If one length argument is given, the string must contain exactly
+    that number of digits.
+    If two length arguments are given, the string must contain a number of
+    digits between those two lengths, inclusive.
+    Return False otherwise.
+    """
+    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))
+    elif num_length_args == 2:
+        good_len = (length_args[0] <= len(s) <= length_args[1])
+    elif num_length_args == 1:
+        good_len = (len(s) == length_args[0])
+    else:
+        good_len = True
+    return bool(good_len and HEX_RE.match(s))
+
+def list_all(fn, num_retries=0, **kwargs):
+    items = []
+    offset = 0
+    items_available = sys.maxint
+    while len(items) < items_available:
+        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