autorun support

Author: jimdigriz

README.md

@@ -22,6 +22,10 @@ information.
 pip install nbgitpuller
 ```
 
+### Configuration
+
+Copy `jupyter_git_pull_config.py` to one of your Jupyter configuration paths (as determined from `jupyter --paths`) and edit it to meet your needs.
+
 ## Example
 
 This example shows how to use the [nbgitpuller link generator]

jupyter_git_pull_config.py

@@ -0,0 +1,34 @@
+# May be set to a list of URLs described as Python regular expressions (using re.fullmatch())
+# where it is permitted to autorun scripts from the pulled project as a pre-initialisation
+# step.
+#
+# WARNING: Enable this only if you understand and accept the risks of AUTORUN.INF.
+# ----
+# c.NbGitPuller.autorun_allow = [
+#   r'https://github\.com/org/name\.git',
+#   r'https://github\.com/org-two/name-two\.git'
+# ]
+# ----
+#
+# To allow all sources (*not* recommended) use:
+# ----
+# c.NbGitPuller.autorun_allow = True
+# ----
+#
+# The default is 'False' which means the autorun functionality is completely disabled
+#c.NbGitPuller.autorun_allow = False
+
+# List of scripts to search for when attempting to autorun. The first match will
+# be run with a single argument of 'init' or 'update' depending on what nbgitpuller
+# is doing.
+# ----
+# c.NbGitPuller.autorun_script = [
+#   '.nbgitpuller.script',
+#   '.different.script'
+# ]
+# ----
+#
+# The 'script' must be executable and when checked out on a 'exec' (ie. not a 'noexec') mountpoint
+#
+# The default is the empty list.
+#c.NbGitPuller.autorun_script = []

nbgitpuller/application.py

@@ -1,6 +1,8 @@
 from .version import __version__ # noqa
 from .pull import GitPuller # noqa
 from jupyter_server.extension.application import ExtensionApp
+from traitlets import Bool, CRegExp, List, Unicode, Union
+from traitlets.config import Configurable
 import os
 
 
@@ -12,6 +14,33 @@ class NbGitPuller(ExtensionApp):
         os.path.join(os.path.dirname(__file__), 'static')
     ]
 
+    autorun_allow = Union(
+        [Bool(), List(CRegExp())],
+        default_value=False,
+        config=True,
+        help="""
+        List of URLs described as Python regular expressions (using re.fullmatch()) where
+        it is permitted to autorun scripts from the pulled project as a pre-initialisation
+        step. Enable this only if you understand and accept the risks of AUTORUN.INF.
+
+        When set to boolean True, all URLs are allowed, whilst False (default) autorun
+        is disabled completely.
+        """
+    )
+
+    autorun_script = List(
+        Unicode(),
+        default_value=[],
+        config=True,
+        help="""
+        List of scripts to search for when attempting to autorun. The first match will
+        be run with a single argument of 'init' or 'update' depending on what nbgitpuller
+        is doing.
+
+        Enable this only if you understand and accept the risks of AUTORUN.INF.
+        """
+    )
+
     def initialize_handlers(self):
         from .handlers import (
             SyncHandler,

nbgitpuller/pull.py

@@ -1,4 +1,5 @@
 import os
+import re
 import subprocess
 import logging
 import time
@@ -81,6 +82,9 @@ def __init__(self, git_url, repo_dir, branch, **kwargs):
         elif not self.branch_exists(self.branch_name):
             raise ValueError(f"Branch: {self.branch_name} -- not found in repo: {self.git_url}")
 
+        self.autorun_allow = kwargs.pop('autorun_allow', False)
+        self.autorun_script = kwargs.pop('autorun_script', [])
+
         newargs = {k: v for k, v in kwargs.items() if v is not None}
         super(GitPuller, self).__init__(**newargs)
 
@@ -143,6 +147,30 @@ def pull(self):
         else:
             yield from self.update()
 
+    def autorun(self, operation="method"):
+        """
+        Search for and execute the autorun script.
+        """
+
+        if not self.autorun_allow:
+            return
+        if not any(( re.fullmatch(pattern, self.git_url) for pattern in self.autorun_allow )):
+            logging.info('autorun skipped, URL does not match any rules')
+            return
+
+        script = next(( s for s in self.autorun_script if os.access(os.path.join(self.repo_dir, s), os.X_OK)), None)
+        if not script:
+            logging.info('autorun skipped, no matching (executable) script')
+            return
+
+        try:
+            for line in execute_cmd([ os.path.join(self.repo_dir, script), operation ], cwd=self.repo_dir, close_fds=True):
+                yield line
+        except subprocess.CalledProcessError:
+            m = f"Problem autorunning {script}"
+            logging.exception(m)
+            raise ValueError(m)
+
     def initialize_repo(self):
         """
         Clones repository
@@ -154,6 +182,7 @@ def initialize_repo(self):
         clone_args.extend(['--branch', self.branch_name])
         clone_args.extend(["--", self.git_url, self.repo_dir])
         yield from execute_cmd(clone_args)
+        yield from self.autorun('init')
         logging.info('Repo {} initialized'.format(self.repo_dir))
 
     def reset_deleted_files(self):
@@ -343,6 +372,8 @@ def update(self):
         yield from self.ensure_lock()
         yield from self.merge()
 
+        yield from self.autorun('update')
+
 
 def main():
     """