We need a new table to store the parallel corpora.
The design is at https://docs.google.com/document/d/11fFVBcu190u8J4uVJdyrghuHx5BdWrRHRZvd7fArwE0/edit#
Table Name: cx_corpora
| column | type | description |
| translation_id | int not null | Translation id |
| section_id | varchar not null | Section Id |
| type | varchar not null | Type of text. Can be source, translation |
| content | mediumblob not null | Content |
| origin | varchar not null | Origin of the text. Can be Source, Apertium, Yandex etc, and User |
| sequence_id | int nullable | For linear ordering of sections |
| timestamp | varchar not null | Timestamp of the translation |