Closed hguan-dev closed 1 month ago
Here is the code in the python script I am using to put together the command: command = [AUDIVERIS_PATH, '-batch', '-output', output_dir, input_filepath]
And here is the Audiveris output in the terminal: stderr: DEBUG:root:Export output: INFO [] CLI 281 | CLI args: [-export, /home/ubuntu/omr-api/output/e90d3bb9-ccc5-4eb4-a391-de9630d4b1e9/93c83f614bc0b1cf1d7a30d942e83237.omr, -output, /home/ubuntu/omr-api/output/e90d3bb9-ccc5-4eb4-a391-de9630d4b1e9/93c83f614bc0b1cf1d7a30d942e83237.mxl]
Not yet answering your problem with a headless configuration, but just a remark about the CLI arguments.
The CLI arguments, as printed by your Audiveris run, look like (I'm putting one argument per line for readability): -export /path/to/some.omr -output /path/to/some.mxl
1/ These CLI arguments don't seem to correspond to the python command (the -batch
argument is gone?)
2/ The -output
argument should introduce a folder, not an .mxl file
Perhaps a stupid question.
If you don't use the -batch
CLI argument, Audiveris is by default launched in the interactive mode.
Which will be non-compatible with your headless environment.
Perhaps your problem is as simple as that?
Fixed the issue, had separate endpoints of upload and transcribe and found the error. Thanks for your help!
After fixing the previous issue of accessing a working build, I'm having difficulties running Audiveris in headless mode through the CLI. As is, my local version is also using -batch as part of the CLI command, but a GUI still ends up popping up (although the processing works fine). After porting this to a Ubuntu AWS EC2 instance, this problem results in some Java Exceptions as it seems the -batch command still does not work properly for me.
Really appreciate any help!
Here is the code in the python script I am using to put together the command:
And here is the Audiveris output in the terminal: