Indexes for embedded documents

Discuss:
Currently MongoEngine design requires explicitly indexes creation at main document. In the example bellow in order to index each Post comments type an index must be created as 'comments.type', the meta part from Comment Class is ignore.

Upgrade the existing logic to support an additional feature that creates indexes for each field based on the Embedded document meta:

class Comment(db.EmbeddedDocument):
    comment = db.StringField()
    type = db.IntField()
    author = db.ReferenceField(User,dbref=False) 
    meta = {'indexes': ['+author','type']}

class Post(db.Document):
    post = db.StringFiled()
    comments = db.ListField(db.EmbeddedDocumentField(Comment))

No files, tests, or entry points are named. Start by tracing the existing index creation logic for embedded documents and how EmbeddedDocument meta indexes are read. Done means an embedded document's meta indexes are applied to the containing document's embedded fields, including the example's comments.type and comments.author indexes.