Re-parse metadata from existing FS files

[jburel]

Assuming all appropriate files have been uploaded, a later version of Bio-Formats could be used to re-parse metadata contained in FS files and either show the discrepancies to the user or directly modify the database. This would allow for super-fast server-side fixes in which a user would not lose their existing work. One caveat would be if the difference in Bio-Formats changed which files were detected during setId, in which case a full re-import would likely be necessary (or hand-fiddling to add/remove the missed files).

[cblackburn]

In the fs-lite import import model I use isMetadataOnly for files that already exist in a repository. The key existing case here is DropBox but it also allows for imports via the repository system (copy then import) or imports from other repositories, whatever they might be! If this flag is set the import should only proceed if all the usedFiles already exist in a repository - this might need to be stricter: a specified area of a repository. But how this happens eventually will depend on the refactoring of the repository service.

[jburel]

Moved from sprint 2012-03-13 (10)

[jburel]

Moved from sprint 2012-03-27 (11)

[jburel]

Moved from sprint 2012-04-10 (12)

[jburel]

Moved from sprint 2012-04-24 (13)

[cblackburn]

Referencing ticket #909 has changed sprint.

[jmoore]

Referencing ticket #909 has changed sprint.

[jmoore]

Colin, is there anything left to do here? Does this include the refactoring of RepositoryImportContainer, or is that a separate ticket?

[cblackburn]

I don't think this is still valid unless Jean-Marie feels there is a use case at this stage for client-side metadata-only imports. Refactoring the RepositoryImportContainer should be a separate ticket as all the elements of that container need to be assessed. I'll let Jean-Marie comment or close as necessary.

[jburel]

we had cases of importing hcs files w/o binary data (from Ruben???). That's a case I had to handle in the UI.
In that situation, the companion file was the only file parsed. We could argue that the companion file will be stored as part of fs.

[jmoore]

Replying to jburel:

we had cases of importing hcs files w/o binary data (from Ruben???). That's a case I had to handle in the UI.

Can you explain more about what should happen specifically? I.e. Ruben wanted to import and skip the binary?

Replying to jburel:

In that situation, the companion file was the only file parsed. We could argue that the companion file will be stored as part of fs.

All companion files should definitely be stored as a part of FS.

[jburel]

I honestly can't remember what was the rational behind it. But having a metadata only flag could be useful if for some reasons a user import a file, there is a bug in BF related to metadata parsing at the time of import and the metadata needs to be read again.

[jamoore]

Closing in favor of ​https://trello.com/c/22XfjucX/159-export-and-re-import. We may want to re-open this when the time comes. (Currently no tasks)