Merge branch '3499-home-project' closes #3499
[arvados.git] / sdk / python / arvados / util.py
index 552fbbe31fad1d106a0ff41f574e0455ec8ae7e8..d5ea18ba49ad749a60aa9faba7141ddb6e3bbf63 100644 (file)
@@ -4,6 +4,10 @@ import os
 import re
 import subprocess
 import errno
+import sys
+from arvados.collection import *
+
+HEX_RE = re.compile(r'^[0-9a-fA-F]+$')
 
 def clear_tmpdir(path=None):
     """
@@ -11,7 +15,7 @@ def clear_tmpdir(path=None):
     exists and is empty.
     """
     if path == None:
-        path = current_task().tmpdir
+        path = arvados.current_task().tmpdir
     if os.path.exists(path):
         p = subprocess.Popen(['rm', '-rf', path])
         stdout, stderr = p.communicate(None)
@@ -26,7 +30,10 @@ def run_command(execargs, **kwargs):
     kwargs.setdefault('close_fds', True)
     kwargs.setdefault('shell', False)
     p = subprocess.Popen(execargs, **kwargs)
-    stdoutdata, stderrdata = p.communicate(None)
+    if kwargs['stdout'] == subprocess.PIPE:
+        stdoutdata, stderrdata = p.communicate(None)
+    else:
+        p.wait()
     if p.returncode != 0:
         raise errors.CommandFailedError(
             "run_command %s exit %d:\n%s" %
@@ -35,12 +42,12 @@ def run_command(execargs, **kwargs):
 
 def git_checkout(url, version, path):
     if not re.search('^/', path):
-        path = os.path.join(current_job().tmpdir, path)
+        path = os.path.join(arvados.current_job().tmpdir, path)
     if not os.path.exists(path):
-        util.run_command(["git", "clone", url, path],
-                         cwd=os.path.dirname(path))
-    util.run_command(["git", "checkout", version],
-                     cwd=path)
+        run_command(["git", "clone", url, path],
+                    cwd=os.path.dirname(path))
+    run_command(["git", "checkout", version],
+                cwd=path)
     return path
 
 def tar_extractor(path, decompress_flag):
@@ -63,7 +70,7 @@ def tarball_extract(tarball, path):
     path -- where to extract the tarball: absolute, or relative to job tmp
     """
     if not re.search('^/', path):
-        path = os.path.join(current_job().tmpdir, path)
+        path = os.path.join(arvados.current_job().tmpdir, path)
     lockfile = open(path + '.lock', 'w')
     fcntl.flock(lockfile, fcntl.LOCK_EX)
     try:
@@ -87,11 +94,11 @@ def tarball_extract(tarball, path):
 
         for f in CollectionReader(tarball).all_files():
             if re.search('\.(tbz|tar.bz2)$', f.name()):
-                p = util.tar_extractor(path, 'j')
+                p = tar_extractor(path, 'j')
             elif re.search('\.(tgz|tar.gz)$', f.name()):
-                p = util.tar_extractor(path, 'z')
+                p = tar_extractor(path, 'z')
             elif re.search('\.tar$', f.name()):
-                p = util.tar_extractor(path, '')
+                p = tar_extractor(path, '')
             else:
                 raise errors.AssertionError(
                     "tarball_extract cannot handle filename %s" % f.name())
@@ -124,7 +131,7 @@ def zipball_extract(zipball, path):
     path -- where to extract the archive: absolute, or relative to job tmp
     """
     if not re.search('^/', path):
-        path = os.path.join(current_job().tmpdir, path)
+        path = os.path.join(arvados.current_job().tmpdir, path)
     lockfile = open(path + '.lock', 'w')
     fcntl.flock(lockfile, fcntl.LOCK_EX)
     try:
@@ -158,7 +165,7 @@ def zipball_extract(zipball, path):
                     break
                 zip_file.write(buf)
             zip_file.close()
-            
+
             p = subprocess.Popen(["unzip",
                                   "-q", "-o",
                                   "-d", path,
@@ -193,7 +200,7 @@ def collection_extract(collection, path, files=[], decompress=True):
     else:
         collection_hash = hashlib.md5(collection).hexdigest()
     if not re.search('^/', path):
-        path = os.path.join(current_job().tmpdir, path)
+        path = os.path.join(arvados.current_job().tmpdir, path)
     lockfile = open(path + '.lock', 'w')
     fcntl.flock(lockfile, fcntl.LOCK_EX)
     try:
@@ -263,7 +270,7 @@ def stream_extract(stream, path, files=[], decompress=True):
     path -- where to extract: absolute, or relative to job tmp
     """
     if not re.search('^/', path):
-        path = os.path.join(current_job().tmpdir, path)
+        path = os.path.join(arvados.current_job().tmpdir, path)
     lockfile = open(path + '.lock', 'w')
     fcntl.flock(lockfile, fcntl.LOCK_EX)
     try:
@@ -281,7 +288,7 @@ def stream_extract(stream, path, files=[], decompress=True):
             files_got += [outname]
             if os.path.exists(os.path.join(path, outname)):
                 os.unlink(os.path.join(path, outname))
-            util.mkdir_dash_p(os.path.dirname(os.path.join(path, outname)))
+            mkdir_dash_p(os.path.dirname(os.path.join(path, outname)))
             outfile = open(os.path.join(path, outname), 'wb')
             for buf in (f.readall_decompressed() if decompress
                         else f.readall()):
@@ -300,7 +307,29 @@ def listdir_recursive(dirname, base=None):
         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 += util.listdir_recursive(ent_path, ent_base)
+            allfiles += listdir_recursive(ent_path, ent_base)
         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))