UFAL/Cannot download and preview files after migration

[milanmajchrak]

Problem description

Downloading of the bitstream thrown a RuntimeException because bitstream file has not extensions e.g., zip, pdf,.. what is not a true.
Preview - generating of preview content cannot be processed because the bitstream extensions are empty.

That means both errors are because of the same reason.
The bitstream format is not properly mapped to the bitstream during migration because the bitstream_format_id is different in the CLARIN-DSpace5 and CLARIN-DSpace7.
Solution could be to find bitstream format by mimetype instead of its ID

[milanmajchrak]

Python update:

diff --git a/data_pump/bitstream.py b/data_pump/bitstream.py
index 88e0562..d2d6519 100644
--- a/data_pump/bitstream.py
+++ b/data_pump/bitstream.py
@@ -24,9 +24,15 @@ def import_bitstream(metadata_class,
     """
     bitstream_json_name = 'bitstream.json'
     bundle2bitstream_json_name = 'bundle2bitstream.json'
+    bitsteamformat_json_name = 'bitstreamformatregistry.json'
     bitstream_url = 'clarin/import/core/bitstream'
     imported = 0

+    bitstreamformat_json_list = read_json(bitsteamformat_json_name)
+    if not bitstreamformat_json_list:
+        logging.info("Bitstreamformatregistry JSON is empty.")
+        return
+
     # load bundle2bitstream
     bundle2bitstream_json_list = read_json(bundle2bitstream_json_name)
     if bundle2bitstream_json_list:
@@ -66,8 +72,7 @@ def import_bitstream(metadata_class,
             bitstream['bitstream_format_id'] = unknown_format_id_val
         params = {'internal_id': bitstream['internal_id'],
                   'storeNumber': bitstream['store_number'],
-                  'bitstreamFormat': bitstreamformat_id_dict[
-                      bitstream['bitstream_format_id']],
+                  'bitstreamFormat': bitstreamformat_json_list[bitstream['bitstream_format_id']]['mimetype'],
                   'deleted': bitstream['deleted'],
                   'sequenceId': bitstream['sequence_id'],
                   'bundle_id': None,

[milanmajchrak]

Pythone updated in this commit: dataquest-dev/dspace-import@16db97d