Add maximum size for entire directory as submission option

The previous maximum size option was applied to individual files only.
This adds a separate option to limit the maximum size of the entire
directory. The option is called max_dir_size and is implemented
similarly to max_file_size. The default is no maximum directory size
limit.

Author: Frans Welin

nbgrader/coursedir.py

@@ -279,6 +279,17 @@ def _validate_root(self, proposal: Bunch) -> str:
         )
     ).tag(config=True)
 
+    max_dir_size = Integer(
+        100000,
+        help=dedent(
+            """
+            Maximum size of directories (in kilobytes; default: 100Mb).
+            Upon copying directories recursively, larger files will be
+            ignored with a warning.
+            """
+        )
+    ).tag(config=True)
+
     def format_path(self, nbgrader_step: str, student_id: str, assignment_id: str, escape: bool = False) -> str:
         kwargs = dict(
             nbgrader_step=nbgrader_step,

nbgrader/exchange/default/exchange.py

@@ -70,13 +70,33 @@ def copy_files(self):
         """Actually do the file transfer."""
         raise NotImplementedError
 
+    def get_size(self, root):
+        """
+        Return total size of directory in bytes.
+        """
+        total_size = 0
+        for dirpath, dirnames, filenames in os.walk(root):
+            for f in filenames:
+                fp = os.path.join(dirpath, f)
+                # skip if it is symbolic link
+                if not os.path.islink(fp):
+                    total_size += os.path.getsize(fp)
+        return total_size
+
+
     def do_copy(self, src, dest, log=None):
         """
         Copy the src dir to the dest dir, omitting excluded
         file/directories, non included files, and too large files, as
         specified by the options coursedir.ignore, coursedir.include
         and coursedir.max_file_size.
         """
+        dir_size = self.get_size(src)
+        max_dir_size = self.coursedir.max_dir_size
+        if dir_size > 1000 * max_dir_size:
+            self.log.error("Directory size is too big")
+            raise RuntimeError(f"Directory size is too big. Size is {dir_size}, maximum size is {1000 * max_dir_size}")
+
         shutil.copytree(src, dest,
                         ignore=ignore_patterns(exclude=self.coursedir.ignore,
                                                include=self.coursedir.include,

nbgrader/tests/apps/test_nbgrader_submit.py

@@ -2,6 +2,7 @@
 import datetime
 import time
 import stat
+import pytest
 
 from os.path import join, isfile, exists
 
@@ -231,7 +232,7 @@ def test_submit_include(self, exchange, cache, course_dir):
         filename, = os.listdir(join(exchange, "abc101", "inbound"))
         assert not exists(join(exchange, "abc101", "inbound", filename, "foo.txt"))
 
-    def test_submit_include(self, exchange, cache, course_dir):
+    def test_submit_max_file_size(self, exchange, cache, course_dir):
         self._release_and_fetch("ps1", exchange, cache, course_dir)
         self._make_file(join("ps1", "small_file"), contents="x" * 2000)
         self._make_file(join("ps1", "large_file"), contents="x" * 2001)
@@ -240,3 +241,11 @@ def test_submit_include(self, exchange, cache, course_dir):
         filename, = os.listdir(join(exchange, "abc101", "inbound"))
         assert exists(join(exchange, "abc101", "inbound", filename, "small_file"))
         assert not exists(join(exchange, "abc101", "inbound", filename, "large_file"))
+
+    def test_submit_max_dir_size(self, exchange, cache, course_dir):
+        self._release_and_fetch("ps1", exchange, cache, course_dir)
+        self._make_file(join("ps1", "small_file"), contents="x" * 2000)
+        self._make_file(join("ps1", "large_file"), contents="x" * 2001)
+        with pytest.raises(RuntimeError):
+            self._submit("ps1", exchange, cache,
+                        flags=['--CourseDirectory.max_dir_size=3'])