Service description should allow 'expected outputs' to be reported

[foreveremain]

Slivka 0.8.3 services do not report what results might be generated, only what files are required:
https://www.compbio.dundee.ac.uk/slivka/api/services/msaprobs

{
  "@url": "/slivka/api/services/msaprobs",
  "author": "Jorge González-Domı́nguez, Yongchao Liu, Juan Touriño, Bertil Schmidt",
  "classifiers": [
    "Topic :: Computational biology :: Sequence analysis",
    "Operation :: Analysis :: Sequence analysis :: Sequence alignment :: Multiple sequence alignment"
  ],
  "description": "MSAProbs is a well-established state-of-the-art multiple sequence alignment algorithm for protein sequences. The design of MSAProbs is based on a combination of pair hidden Markov models and partition functions to calculate posterior probabilities.",
  "id": "msaprobs",
  "license": "GNU GPL ver. 3",
  "name": "MSAProbs",
  "parameters": [
    {
      "array": false,
      "default": null,
      "description": "",
      "id": "input",
      "mediaType": "application/fasta",
      "mediaTypeParameters": {

      },
      "mimetype": "application/fasta",
      "name": "Input file",
      "required": true,
      "type": "file"
    },
    {
      "array": false,
      "default": 2,
      "description": "Number of passes of consistency transformation.",
      "id": "consistency",
      "max": 5,
      "min": 0,
      "name": "Consistency repetitions",
      "required": false,
      "type": "integer"
    },
    {
      "array": false,
      "default": 10,
      "description": "Number of iterative refinement passes.",
      "id": "iterative-refinement",
      "max": 1000,
      "min": 0,
      "name": "Interative refinaement",
      "required": false,
      "type": "integer"
    }
  ],
  "presets": [],
  "status": {
    "errorMessage": "",
    "status": "OK",
    "timestamp": "2025-07-15T16:36:19"
  },
  "version": "0.9.7"
}

Without some indication of expected outputs, it means the client has to know what outputs to expect itself in order to use the service. Whilst this works, it means a client needs to run a service in order to discover what outputs are supported (and so learn its operational pattern). It would be better to report expected outputs directly:

"outputs": {
  "alignment":
    { "media-type": "application/fasta" }
  "log":
    {"media-type": "text/plain"}
  "error-log":
    {"media-type": "text/plain"}
 }

The outputs hash is simply the named outputs in the service definition without any 'path' or other internal attributes.