Fix ids from tokenizing
During the creation of the ZIP file used during the collation each word receives an ID. This ID can differ from the ID in the temporary files. To avoid this, the ZIP creation now relies on the temporary files as well.
During the creation of the ZIP file used during the collation each word receives an ID. This ID can differ from the ID in the temporary files. To avoid this, the ZIP creation now relies on the temporary files as well.