refactor: reduce reliance on regex in path handling

This change tries to modernize the way we handle paths in nbgrader
by using pathlib.Path and centralizing logic for searching and
extracting information from paths into CourseDirectory.

By using `Path(root).glob(pat)` instead of `glob.glob(root / pat)`, we
can avoid some issues of needing to escape CourseDirectory.root because
the glob will be scoped underneath a fixed path.

Introduces a new `CourseDirectory.find_assignments` method to help
locate assignment directories across steps and students.

Fixes #1965

Author: shreve

nbgrader/apps/api.py

@@ -4,6 +4,8 @@
 import os
 import logging
 import warnings
+import typing
+from pathlib import Path
 
 from traitlets.config import LoggingConfigurable, Config, get_config
 from traitlets import Instance, Enum, Unicode, observe
@@ -123,7 +125,7 @@ def gradebook(self):
         """
         return Gradebook(self.coursedir.db_url, self.course_id)
 
-    def get_source_assignments(self):
+    def get_source_assignments(self) -> typing.Set[str]:
         """Get the names of all assignments in the `source` directory.
 
         Returns
@@ -132,29 +134,14 @@ def get_source_assignments(self):
             A set of assignment names
 
         """
-        filenames = glob.glob(self.coursedir.format_path(
-            self.coursedir.source_directory,
-            student_id='.',
-            assignment_id='*'))
-
-        assignments = set([])
-        for filename in filenames:
-            # skip files that aren't directories
-            if not os.path.isdir(filename):
-                continue
 
-            # parse out the assignment name
-            regex = self.coursedir.format_path(
-                self.coursedir.source_directory,
-                student_id='.',
-                assignment_id='(?P<assignment_id>.*)',
-                escape=True)
-
-            matches = re.match(regex, filename)
-            if matches:
-                assignments.add(matches.groupdict()['assignment_id'])
-
-        return assignments
+        return {
+            entry['assignment_id'] for entry in
+            self.coursedir.find_assignments(
+                nbgrader_step=self.coursedir.source_directory,
+                student_id=".",
+            )
+        }
 
     def get_released_assignments(self):
         """Get the names of all assignments that have been released to the
@@ -194,32 +181,14 @@ def get_submitted_students(self, assignment_id):
             A set of student ids
 
         """
-        # get the names of all student submissions in the `submitted` directory
-        filenames = glob.glob(self.coursedir.format_path(
-            self.coursedir.submitted_directory,
-            student_id='*',
-            assignment_id=assignment_id))
 
-        students = set([])
-        for filename in filenames:
-            # skip files that aren't directories
-            if not os.path.isdir(filename):
-                continue
-
-            # parse out the student id
-            if assignment_id == "*":
-                assignment_id = ".*"
-            regex = self.coursedir.format_path(
-                self.coursedir.submitted_directory,
-                student_id='(?P<student_id>.*)',
-                assignment_id=assignment_id,
-                escape=True)
-
-            matches = re.match(regex, filename)
-            if matches:
-                students.add(matches.groupdict()['student_id'])
-
-        return students
+        return {
+            entry['student_id'] for entry in
+            self.coursedir.find_assignments(
+                nbgrader_step=self.coursedir.submitted_directory,
+                assignment_id=assignment_id
+            )
+        }
 
     def get_submitted_timestamp(self, assignment_id, student_id):
         """Gets the timestamp of a submitted assignment.
@@ -243,10 +212,7 @@ def get_submitted_timestamp(self, assignment_id, student_id):
             student_id,
             assignment_id))
 
-        timestamp_pth = os.path.join(assignment_dir, 'timestamp.txt')
-        if os.path.exists(timestamp_pth):
-            with open(timestamp_pth, 'r') as fh:
-                return parse_utc(fh.read().strip())
+        return self.coursedir.get_existing_timestamp(assignment_dir)
 
     def get_autograded_students(self, assignment_id):
         """Get the ids of students whose submission for a given assignment
@@ -439,21 +405,14 @@ def get_notebooks(self, assignment_id):
 
             # if it doesn't exist in the database
             else:
-                sourcedir = self.coursedir.format_path(
-                    self.coursedir.source_directory,
-                    student_id='.',
-                    assignment_id=assignment_id)
-                escaped_sourcedir = self.coursedir.format_path(
+                sourcedir = Path(self.coursedir.format_path(
                     self.coursedir.source_directory,
                     student_id='.',
-                    assignment_id=assignment_id,
-                    escape=True)
+                    assignment_id=assignment_id))
 
                 notebooks = []
-                for filename in glob.glob(os.path.join(sourcedir, "*.ipynb")):
-                    regex = re.escape(os.path.sep).join([escaped_sourcedir, "(?P<notebook_id>.*).ipynb"])
-                    matches = re.match(regex, filename)
-                    notebook_id = matches.groupdict()['notebook_id']
+                for filename in sourcedir.glob("*.ipynb"):
+                    notebook_id = filename.stem
                     notebooks.append({
                         "name": notebook_id,
                         "id": None,

nbgrader/converters/autograde.py

@@ -181,18 +181,18 @@ def _init_preprocessors(self) -> None:
         for pp in preprocessors:
             self.exporter.register_preprocessor(pp)
 
-    def convert_single_notebook(self, notebook_filename: str) -> None:
+    def convert_single_notebook(self, notebook_filename: str, assignment_id: str, student_id: str) -> None:
         self.log.info("Sanitizing %s", notebook_filename)
         self._sanitizing = True
         self._init_preprocessors()
-        super(Autograde, self).convert_single_notebook(notebook_filename)
+        super(Autograde, self).convert_single_notebook(notebook_filename, assignment_id, student_id)
 
         notebook_filename = os.path.join(self.writer.build_directory, os.path.basename(notebook_filename))
         self.log.info("Autograding %s", notebook_filename)
         self._sanitizing = False
         self._init_preprocessors()
         try:
             with utils.setenv(NBGRADER_EXECUTION='autograde'):
-                super(Autograde, self).convert_single_notebook(notebook_filename)
+                super(Autograde, self).convert_single_notebook(notebook_filename, assignment_id, student_id)
         finally:
             self._sanitizing = True