AutoDesk Vault Connector Transformer

Excellent idea!

At the Belgian railways, we now access Vault with a Custom Transformer, using SQLExecutors and a Python script. With the File Copy Writer, we download the file to a specific location, so we can use it in the next step of the Workspace.

Why a Transformer? I would rather think of it as a Reader. Similar to the Oracle reader, where you can store your connection details for each user. By analogy with the Oracle Reader (where you can access the table list), this reader should access different data folders for the specified user.

The output could be similar to the PATH reader, so you can link it to a FeatureReader, OR it could read the file in Vault directly. Since you are not sure about the format of the data set, I think the first is better.

To specify the right file, we either use the FileIterationID (the unique key of the file in Vault) or the FileName (field in the Vault environment). This could be a parameter in the Reader, so you do not need a TestFilter afterwards.

Who has some other ideas?

1) One of the most time consuming things when reading data from Vault, is finding which file property is stored in which table of the SQL Server database. Once you know which ID's you can use to link tables, retrieving data is just a matter of creating the query. But this is not a very generalized way of working, as the output schema is fixed.

It would be nice if the output of a VaultConnector could be more flexible. Not only use filenames, but also export data based on the creation date, name of the person who created a file, all files within a sub-project, all files that are not approved yet, etc. There are so many file properties available within Vault, so it would be nice to combine all these parameters to tailor your export.

To be more flexible, a new VaultConnector might use the Vault API? I don't have much experience on the API but it might be a good idea?

2) Writing data to Vault might be something to be prudent about. Can FME overwrite files Vault? Change the revision number of a file? How will we make sure that all the essential file properties remain?

Perhaps we should have the option to set up a Vault connection in the FME Web connections, then allow the use of that connection when reading/writing Vault compatible formats.

Another opportunity we have with Vault is to use the Job Handler API to create a connection to allow FME to be used as a Vault Job Handler. This could connect to either command line FME or FME Server, and would let Vault users incorporate FME processing into their Vault workflows.