Allow builds to be started with an external archive URL

Fixes #114

endpoints/api/build.py

@@ -3,8 +3,10 @@
 import logging
 import json
 import datetime
+import hashlib
 
 from flask import request
+from rfc3987 import parse as uri_parse
 
 from app import app, userfiles as user_files, build_logs, log_archive, dockerfile_build_queue
 from endpoints.api import (RepositoryParamResource, parse_args, query_param, nickname, resource,
@@ -134,8 +136,11 @@ def build_status_view(build_obj):
     }
   }
 
-  if can_write and build_obj.resource_key is not None:
-    resp['archive_url'] = user_files.get_file_url(build_obj.resource_key, requires_cors=True)
+  if can_write:
+    if build_obj.resource_key is not None:
+      resp['archive_url'] = user_files.get_file_url(build_obj.resource_key, requires_cors=True)
+    elif job_config.get('archive_url', None):
+      resp['archive_url'] = job_config['archive_url']
 
   return resp
 
@@ -148,14 +153,15 @@ class RepositoryBuildList(RepositoryParamResource):
     'RepositoryBuildRequest': {
       'type': 'object',
       'description': 'Description of a new repository build.',
-      'required': [
-        'file_id',
-      ],
       'properties': {
         'file_id': {
           'type': 'string',
           'description': 'The file id that was generated when the build spec was uploaded',
         },
+        'archive_url': {
+          'type': 'string',
+          'description': 'The URL of the .tar.gz to build. Must start with "http" or "https".',
+        },
         'subdirectory': {
           'type': 'string',
           'description': 'Subdirectory in which the Dockerfile can be found',
@@ -204,7 +210,26 @@ class RepositoryBuildList(RepositoryParamResource):
     logger.debug('User requested repository initialization.')
     request_json = request.get_json()
 
-    dockerfile_id = request_json['file_id']
+    dockerfile_id = request_json.get('file_id', None)
+    archive_url = request_json.get('archive_url', None)
+
+    if not dockerfile_id and not archive_url:
+      raise InvalidRequest('file_id or archive_url required')
+
+    if archive_url:
+      archive_match = None
+      try:
+        archive_match = uri_parse(archive_url, 'URI')
+      except ValueError:
+        pass
+
+      if not archive_match:
+        raise InvalidRequest('Invalid Archive URL: Must be a valid URI')
+
+      scheme = archive_match.get('scheme', None)
+      if scheme != 'http' and scheme != 'https':
+        raise InvalidRequest('Invalid Archive URL: Must be http or https')
+
     subdir = request_json['subdirectory'] if 'subdirectory' in request_json else ''
     tags = request_json.get('docker_tags', ['latest'])
     pull_robot_name = request_json.get('pull_robot', None)
@@ -228,18 +253,24 @@ class RepositoryBuildList(RepositoryParamResource):
     # Check if the dockerfile resource has already been used. If so, then it
     # can only be reused if the user has access to the repository in which the
     # dockerfile was previously built.
-    associated_repository = model.build.get_repository_for_resource(dockerfile_id)
-    if associated_repository:
-      if not ModifyRepositoryPermission(associated_repository.namespace_user.username,
-                                        associated_repository.name):
-        raise Unauthorized()
+    if dockerfile_id:
+      associated_repository = model.build.get_repository_for_resource(dockerfile_id)
+      if associated_repository:
+        if not ModifyRepositoryPermission(associated_repository.namespace_user.username,
+                                          associated_repository.name):
+          raise Unauthorized()
 
     # Start the build.
     repo = model.repository.get_repository(namespace, repository)
 
+    build_name =  (user_files.get_file_checksum(dockerfile_id)
+                    if dockerfile_id
+                    else hashlib.sha224(archive_url).hexdigest()[0:7])
+
     prepared = PreparedBuild()
-    prepared.build_name = user_files.get_file_checksum(dockerfile_id)
+    prepared.build_name = build_name
     prepared.dockerfile_id = dockerfile_id
+    prepared.archive_url = archive_url
     prepared.tags = tags
     prepared.subdirectory = subdir
     prepared.is_manual = True