AuTA can parse the output generated by Docker containers. As soon as execution finished and the logs have been extracted, an archive of the `/var/auta/out` directory is downloaded for analysis. This folder may contain sub-folders that contain files to be processed. Below are the formats AuTA supports.
### :warning: Prevent fraud
If student code is executed as part of analysis, make sure that the output directory is clean before files are moved there. Otherwise, students could place malicious files in the directory and pass insufficient submissions, or worse.
## JUnit
Files in `/var/auta/out/junit` are treated as standard JUnit XML files. Each file represents one or more test suites with test cases. For every error'd or failed test case, a TESTS string metric containing the failure message is added to the entity describing the test case. Its default scripts transform these messages into warnings or failures, depending on the active script.
## 🛠️ Implementation details
The DockerRunner class has a collection of OutputProcessor instances. These processors are very similar to analyzers, except they do not have a list of supported languages. Instead, they always run even if their directory does not exist. In fact, the runner has no notion of sub-directories, so such a processor is free to determine its own analysis method.
Also like analyzers, processors are automatically detected at compile-time and stored in a manifest.
