Metadata file names should be configurable

[konstjar]

The name of files for parsing should be configurable in DataNode configuration, static value that admin can extend.
It may be an array of file names that should be looked in archive like execution-config.json or analysisMetadata.json

The metadata should be loaded from the first available file in archive.

datanode.metadata.fileNames: execution-config.json,analysisMetadata.json

Implementation Notes

To facilitate the server-side implementation for detection of metadata and entry points, the following endpoints are added:

POST /api/v1/analysis/upload/(zip|files)

Both of these accept multipart/form-data with the file part containing the file(s) being uploaded and store them in the same way currently uploaded analyses are persited, then provide response structure as follows:

{name: string, files: [string], metadata: {...}}

name - unique identifier of the upload.
metadata - the contents of detected metadata file. It is validated to be a valid JSON, however otherwise provided "as is", the responsibility to map specific data items to analysis fields remains with frontend.
In the future, we can consider to have a configurable mapping in addition to name, however this is not in scope of the current task.
files - the list of all files in the archive. For the moment, any filtering requirements should be implemented on frontend, however for the future it can be considered to implement configurable filtering on server side (and we might even skip saving undesirable files)

POST /api/v1/analysis/execute/{name}

{executableFileName, datasourceId, title, study type, environmentId, dockerImage}

Request body specification (as well as response contract) is the same as existing POST /api/v1/analysis/zip endpoint however unlike it, the new endpoint accepts "application/json" and doesn't accept files uploaded.
For execution, it uses preuploaded files stored under name identifier.