eNMS 3.21

• When entering a subworkflow, the selected runtime is now preserved.
• When running a workflow, the runtime is added to the runtime list in workflow builder and selected.
• Workflow Refresh button now updates the list of runtimes in the workflow builder dropdown of runtimes.
• Duplicating a shared service from the workflow builder now creates a NON SHARED deep copy in the current workflow only.
• Created dedicated category for shared services in "Add services to workflow" tree.
• Implemented "Clear all filters" mechanism for all tables
• When displaying workflow services in service table, all search input resetted (otherwise nothing was displayed)
• Add download buttons for configuration and operational data
• Add button in tables to export search result as CSV file.
• When duplicating top-level workflow, display edit panel
• Fix progress display for service in run once mode in workflow builder
• Multiline field for skip / device query
• Add "Maximum number of retries" property to prevent infinite loop (hardcoded before)
• Add "All" option in relationship filtering (filter object with relation to All)
• Rename "never_update" with "manually_defined"
• Set focus on name field when creating a new instance
• New property in service panel (targets section): Update pools before running.
• Extend the custom properties to all classes including services (displayed in an accordion in first tab).
• Add new search mechanism in the "Add services to workflow" panel
• Add new "Trigger" property for runs to know if they were started from the UI or REST API
• Add time-stamp of when the configuration / oper data displayed was collected
• Ability to display config older config from GIT
• Ability to compare currently displayed config/data to any point in time in the past.
• Syntax highlight option: ability to highlight certain keywords based on regular expression match,
  defined in eNMS/static/lib/codemirror/logsMode. Can be customized.
• New logging property to configure log level for a service or disable logging.
• Fix bug when typing invalid regex for table search (eg "(" )
• Dont display Start / End services in service table
• Make configuration search case-insensitive for inclusion ("Search" REST endpoint + UI)
• Use log level of top-level workflow for all services.
• Add context sensitive help mechanism
• Add keyword so that the "log" function in a service can log to the application log (+ create log object)
• Add timestamp for session logs
• Add device result counter in result tree window
• Move to optional_requirements file and catch import error of all optional libraries:
  ansible, hvac, ldap3, pyats, pynetbox, slackclient>=1.3,<2, tacacs_plus
• Fix Napalm BGP example service
• Fix 404 custom passwords logs from Vault
• Encrypt and decrypt all data going in and out of the vault (b64 / Fernet)
• No longer store user password when external authentication is used (LDAP/TACACS+)
• No longer create / import duplicated edges of the same subtype.
• Add preprocessing code area for all services
• all post processing mode: "run on success" / "run on failure" / "run all the time" selector
• Support functions and classes with set_var / get_var
• Fix front end bug when displaying the results if they contain a python SET (invalid JSON):
  all non-JSON compliant types are now automatically converted to a string when saving the results in the
  database, and a warning is issue in the service logs.
• Add superworkflow mechanism
• Add jump on connect support
• Add log deletion support from CLI interface
• Forbid import of "os", "subprocess" and "sys" in a python code area in service panel
  (snippet, pre/postprocessing, etc)
• Refactor logging configuration: all the logging are now configured from a file in setup: logging.json
  Besides, the log function in a workflow takes a new parameter "logger" where you can specify a logger name.
  This means you can first add your own loggers in logging.json, then log to them from a workflow.
• Remove CLI fetch, update and delete endpoint (curl to be used instead if needed)
• Improve workflow stop mechanism: now hitting stop will try to stop ASAP, not just after the on-going
  service but also after the on-going device, or after the on-going retry (e.g many retries...).
  Besides stop should now work from subworkflows too.