Merge branch 'devel'

[qcg-portal.git] / filex / ftp.py

diff --git a/filex/ftp.py b/filex/ftp.py
index c0b3e25..85161a4 100644 (file)
--- a/filex/ftp.py
+++ b/filex/ftp.py
@@ -1,18 +1,26 @@
 from datetime import datetime
 from Queue import Queue, Empty
+from itertools import chain
+import os
 import re
 from threading import Event
+from urlparse import urlparse, urlunparse
 
+from django.utils.http import urlquote
 from django.utils.timezone import localtime, UTC
-from gridftp import FTPClient, Buffer, HandleAttr, OperationAttr
 
 
-class FTPException(Exception):
-    pass
+class FTPError(Exception):
+    def __init__(self, message, verbose=None, *args, **kwargs):
+        super(FTPError, self).__init__(message, *args, **kwargs)
+
+        self.verbose = verbose
 
 
 class FTPOperation:
     def __init__(self, proxy=None, buffer_size=4096):
+        from gridftp import FTPClient, Buffer, HandleAttr, OperationAttr
+
         self._end = Event()
         self._error = None
         self._buffer = Buffer(buffer_size)
@@ -68,7 +76,7 @@ class FTPOperation:
 
             msg = match.groups()[0] if match else "Unknown error"
 
-            raise FTPException(msg)
+            raise FTPError(msg, self._error)
 
     def listing(self, url):
         self.cli.verbose_list(url, self._done, None, self.op_attr)
@@ -84,25 +92,26 @@ class FTPOperation:
 
     @staticmethod
     def _parse_mlst(listing):
-        data = []
-
         for item in listing.strip().splitlines():
             # we may receive empty string when there are multiple consecutive newlines in listing
             if item:
                 attrs, name = item.split(' ', 1)
 
-                attrs = dict((attr.split('=') for attr in attrs.split(';') if attr))
+                attrs_dict = {}
+                for attr in attrs.split(';'):
+                    try:
+                        key, value = attr.split('=', 1)
+                    except ValueError:
+                        key, value = attr, ''
 
-                date = localtime(datetime.strptime(attrs['Modify'], "%Y%m%d%H%M%S").replace(tzinfo=UTC()))
+                    attrs_dict[key] = value
 
-                data.append({
+                yield {
                     'name': name,
-                    'type': 'file' if attrs['Type'] == 'file' else 'directory',
-                    'size': int(attrs['Size']),
-                    'date': date,
-                })
-
-        return data
+                    'type': 'directory' if attrs_dict['Type'].endswith('dir') else 'file',
+                    'size': int(attrs_dict['Size']),
+                    'date': localtime(datetime.strptime(attrs_dict['Modify'], "%Y%m%d%H%M%S").replace(tzinfo=UTC())),
+                }
 
     def get(self, url):
         self.cli.get(url, self._done, None, self.op_attr)
@@ -125,3 +134,100 @@ class FTPOperation:
         self.cli.move(src, dst, self._done, None, self.op_attr)
 
         self.wait()
+
+    def info(self, url):
+        data = self.listing(url).next()
+
+        if data['name'] == '.':
+            data['name'] = os.path.basename(os.path.normpath(url))
+
+        return data
+
+    def exists(self, url):
+        self.cli.exists(url, self._done, None, self.op_attr)
+
+        try:
+            self.wait()
+        except FTPError as e:
+            if 'No such file or directory' in e.message:
+                return False
+            raise
+        else:
+            return True
+
+    def delete(self, url):
+        self.cli.delete(url, self._done, None, self.op_attr)
+
+        self.wait()
+
+    def rmdir(self, url):
+        self.cli.rmdir(url, self._done, None, self.op_attr)
+
+        self.wait()
+
+    def mkdir(self, url, parents=False):
+        if parents:
+            if self.exists(url):
+                return
+
+            u = urlparse(url)
+            parent_url = urlunparse((u.scheme, u.netloc, os.path.dirname(os.path.normpath(u.path)), '', '', ''))
+
+            self.mkdir(parent_url, parents=True)
+
+        self.cli.mkdir(url, self._done, None, self.op_attr)
+
+        self.wait()
+
+    @staticmethod
+    def match_ext(archive, *extensions):
+        for ext in extensions:
+            if archive.endswith(ext):
+                return True
+        return False
+
+    def compress(self, server, path, files, archive):
+        self._check_disk_stack_args(*([path, archive] + files))
+
+        if self.match_ext(archive, '.tar.gz', '.tgz'):
+            cmd, args = 'tar', ['czf', archive, '-C', path] + files
+        elif self.match_ext(archive, '.tar.bz2', '.tbz'):
+            cmd, args = 'tar', ['cjf', archive, '-C', path] + files
+        elif self.match_ext(archive, '.zip'):
+            # zip doesn't support unicode file names
+            for arg in files:
+                try:
+                    arg.encode('ascii')
+                except UnicodeEncodeError as e:
+                    raise ValueError(u'Unsupported character `{}` in `{}`!'.format(arg[e.start:e.start + 1], arg))
+
+            cmd, args = 'jar', (['cMf', archive] + list(chain.from_iterable(('-C', path, f) for f in files)))
+        else:
+            raise ValueError('Unknown archive type: {}'.format(archive))
+
+        self.op_attr.set_disk_stack('#'.join(["popen:argv=", cmd] + map(urlquote, args)))
+
+        return self.get(server)
+
+    def extract(self, server, archive, dst):
+        self._check_disk_stack_args(*[archive, dst])
+
+        if self.match_ext(archive, '.tar.gz', '.tgz'):
+            cmd, args = 'tar', ('xzf', archive, '-C', dst)
+        elif self.match_ext(archive, '.tar.bz2', '.tbz'):
+            cmd, args = 'tar', ('xjf', archive, '-C', dst)
+        elif self.match_ext(archive, '.zip'):
+            cmd, args = 'unzip', ('-qo', archive, '-d', dst)
+        else:
+            raise ValueError('Unknown archive type: {}'.format(archive))
+
+        self.op_attr.set_disk_stack('#'.join(["popen:argv=", cmd] + map(urlquote, args)))
+
+        return self.get(server)
+
+    @staticmethod
+    def _check_disk_stack_args(*args):
+        for char in ['#', ';']:
+            for arg in args:
+                if char in arg:
+                    raise ValueError(u'Unsupported character `{}` in `{}`!'.format(char, arg))