Added web API to clone a document from a remote git repository

src/api/document.py

@@ -1,6 +1,8 @@
 from flask import Blueprint, request
 from markupsafe import escape
 from web_utils.get_arg import get_arg
+from web_utils.run import run
+import os
 
 from data.document import Document
 
@@ -20,3 +22,14 @@ def build():
 	output += doc.build()
 	
 	return output.replace('\n', '<br/>')
+
+@bp.route('/clone')
+def clone():
+	repo = get_arg('repo')
+	doc_name = get_arg('doc', os.path.splitext(os.path.basename(repo))[0])
+	branch = get_arg('branch', 'master')
+	source_dir = get_arg('source', 'source')
+	
+	output = Document.clone(repo, branch, doc_name, source_dir)
+	
+	return output.replace('\n', '<br/>')

src/data/document.py

@@ -1,17 +1,38 @@
 import os
 from web_utils.run import run
+import shutil
+from unicodedata import normalize
+import string
 
 document_root = None
 
+def os_path_separators():
+	seps = ['/','\\']
+	for sep in os.path.sep, os.path.altsep:
+		if sep:
+			seps.append(sep)
+	return seps
+
+def sanitize_name(initial_name):
+	# Sort out unicode characters
+	name = normalize('NFKD', initial_name).encode('ascii', 'ignore').decode('ascii')
+	# Replace path separators with underscores
+	name = name.replace('/', '_slash_')
+	for sep in os_path_separators():
+		name = name.replace(sep, '_')
+	# Ensure only valid characters
+	valid_chars = "-_.(){0}{1}".format(string.ascii_letters, string.digits)
+	name = "".join(ch for ch in name if ch in valid_chars)
+	if len(name) == 0 or '..' in name:
+		raise Exception("Invalid name: " + initial_name)
+	return name
+
 class Document:
 	def __init__(self, doc_name, branch = 'master'):
 		self.doc_name = doc_name
 		self.branch = branch
 		
-		doc_path = os.path.realpath(get_document_root()+'/'+doc_name+'/'+branch)
+		doc_path = Document.make_doc_path(doc_name, branch)
-		if not doc_path.startswith(get_document_root()):
-			raise Exception("Invalid document path for "+doc_name+"@"+branch)
-			
 		if not os.path.isdir(doc_path + "/repo/.git"):
 			raise Exception("This document does not exist: "+doc_name+"@"+branch)
 		
@@ -24,6 +45,46 @@ class Document:
 	
 	def pull(self):
 		return run("cd \"" + self.doc_path + "/repo\" && git pull")
+	
+	@staticmethod
+	def make_doc_path(doc_name, branch):
+		doc_path = os.path.realpath(get_document_root()+'/'+sanitize_name(doc_name)+'/'+sanitize_name(branch))
+		if not doc_path.startswith(get_document_root()):
+			raise Exception("Invalid document path for "+doc_name+"@"+branch)
+		return doc_path
+	
+	@staticmethod
+	def clone(repo, branch, doc_name, source_dir):
+		# check the document does not already exist
+		doc_path = Document.make_doc_path(doc_name, branch)
+		if os.path.isdir(doc_path):
+			raise Exception("This document already exists: "+doc_name+"@"+branch)
+			
+		if source_dir != sanitize_name(source_dir):
+			raise Exception("Invalid source directory name: " + source_dir)
+			
+		# we have potentially serious security issues related to cloning anything. For example cloning from SSH may use a pre-configured server identity, etc.
+		if not repo.startswith("https://"):
+			raise Exception("Only HTTPS repositories are allowed in current implementation")
+		
+		target_dir = doc_path + "/repo"
+		
+		cmd = ""
+		cmd += "mkdir -p \"" + target_dir + "\"\n"
+		cmd += "cd \"" + target_dir + "\"\n"
+		cmd += "git init \"--initial-branch=" + branch + "\"\n"
+		cmd += "git remote add -f origin \"" + repo + "\"\n"
+		cmd += "git sparse-checkout init\n"
+		cmd += "git sparse-checkout set \"" + source_dir + "\"\n"
+		cmd += "git pull origin \"" + branch + "\"\n"
+		cmd += "git branch \"--set-upstream-to=origin/" + branch + "\" \"" +branch + "\""
+		
+		try:
+			return run(cmd)
+		except Exception as e:
+			# cloning failed, clean up and raise the same exception again
+			shutil.rmtree(doc_path)
+			raise e
 
 def set_document_root(dir):
 	global document_root