Execute button calls endpoint to get order, and synchronously hits the execute endpoint for each node.
Status is included in the JS node model, and updated after configuration and execution. The stoplight color reflects the status (unconfigured, configured, complete).
Include a download_result flag on nodes to indicate whether their output should be downloaded after execution. This applies to WriteCsvNodes for now.
We can't just load the "intermediate" data (in JSON format), because it's not the CSV created from the users options. So I added a vulnerable endpoint that will download an arbitrary filename. There's no node- or workflow-ID included in the filenames, so it's also open to name clashes in /tmp.
I know it's ugly, but I used the Django FileStorage object in the execute method of ReadCsvNode, so that the download endpoint could find the file.
Hit download endpoint after execution of nodes with that flag.
Removed the use of options-as-kwargs in the pandas calls in Read/Write CSV node execution methods.
Use display names instead of class names for the node menu categories (e.g. "I/O" instead of "IONode"
download_resultflag on nodes to indicate whether their output should be downloaded after execution. This applies to WriteCsvNodes for now.executemethod ofReadCsvNode, so that the download endpoint could find the file.