Add customisable server email capabilities

[drussell-x]

In order to facilitate sending announcement (and potentially other) emails from the OMERO server instance the sendEmail method or whatever the new method is called will need to accept several parameters:

This would send a message with the defined subject and message. I currently have no need to send email with CCs or recipients, but I think it would make sense to add support for this initially in case someone wishes to use it. Thus it is possible to specify a list of experimenters and groups. It also accepts a list of strings of arbitrary email addresses. This is useful for CCing a departmental mailing list or similar. This is especially important in an LDAP environment as the server is accessible to new users at any time so they'd need to know about a service disruption up-front etc.

// Sends a single message to a collection of experimenters resolved from the Experimenter
// and ExperimenterGroup lists.
boolean sendEmail(
    boolean all,
    List<Experimenter> experimentersTo,
    List<ExperimenterGroup> groupsTo,
    List<String> cc,
    String subject,
    String message
)

// Sends a message many times, individually (standard automated email procedure)
boolean sendMassEmail(
    boolean all,
    List<Experimenter> experimentersTo,
    List<ExperimenterGroup> groupsTo,
    List<String> additionalRecipients,
    String subject,
    String message
)

I'm undecided what the return types should be. All success/fail is one option. What about a partial failure? What about a failure to resolve a specified user as they don't have an email address, should that fail silently?

It might also be nice to have the capability for an admin to check if the server
has the capability to send emails. Obviously it can't really check if the emails
are correctly being sent, but just that the settings are present in the server
to enable it to send email.

boolean isEmailEnabled()

Some of the required functionality can be recoded in Java from ​here. I think group resolution to email lists is probably the missing piece there.

Cheers,

Douglas

[drussell-x]

Just wondering if there is any progress on this so I can then update my emailing utility to use it and then file a PR for that. Thanks.

[jamoore]

As discussed in devteam, there hasn't been any progress on moving this behind the IAdmin (or similar) interface yet.

Douglas opened his implementation as a PR: ​https://github.com/openmicroscopy/openmicroscopy/pull/2127

If it's merged, then the API changes should likely also be moved from "unscheduled" to "5.1.0*".

Will / Ola: room for discussion of webadmin extension points?

[atarkowska]

We already have that functionality hidden on the server side in ​sendEmail. My intention is to evaluate existing sendEmail method and move it to the new interface called IEmail.

[jamoore]

Ola, what all do you see being in IEmail? If the number of methods is fairly small, it'd likely be better to do this via the omero.cmd-style invocation, because that way we also get status updates ("Sending emails: 25% done") for free.

[atarkowska]

Here is a list of methods

 // ~ Email
    // =========================================================================

    /**
     * Send email to every users.
     */
    boolean sendEmailToAllUsers(List<String> cc, String subject, String message);
    
    /**
     * Send email to active users.
     */
    boolean sendEmailToActiveUsers(List<String> cc, String subject, String message);
    
    /**
     * Send email to inactive users.
     */
    boolean sendEmailToInactiveUsers(List<String> cc, String subject, String message);
    
    /**
     * Send email to the given users.
     */
    boolean sendEmailToUsers(List<Experimenter> exps, List<String> cc, String subject, String message);
    
    /**
     * Send email to users in the given groups.
     */
    boolean sendEmailToGroups(List<ExperimenterGroup> groups, List<String> cc, String subject, String message);
    

I am still thinking about mass email

[atarkowska]

Douglas any comments?

[jamoore]

After discussing briefly with Ola the possibility of using a ​omero.cmd-style service, I'd propose something like:

class Message {
  LongList userIds; // or by name?
  LongList groupIds; // ditto?
  StringSet cc;
};

class TextMessage {
  string subject;
  string body;
};

class SendEmail extends Request {
    Message msg;
};

Questions:

• Any need to support mime messages? (i.e. attachments)
• Should templating be done server side?

in which case

class TemplateMessage extends Message {
    string templateName;
    StringRTypeMap properties;
};

[atarkowska]

Josh, or JoshB how the omero.cmd handle mass request?

[jamoore]

This may not be exactly what we want since it's conceptually in a single transaction, but you can wrap any number of omero.cmd.Request objects with a DoAll. There are a couple examples in gateway/__init__.py.

[atarkowska]

Please check the latest draft in ​https://github.com/aleksandra-tarkowska/openmicroscopy/commit/a074a6438d3a1010c3b93e025d0a8d236fcb708a Comments welcome

[atarkowska]

This code can be tested by

exp = omero.model.ExperimenterI()
exp.email = rstring("user@email.com")
req = omero.cmd.SendEmailRequest(subject="test subject", body="test message", users=[exp])
gateway = BlitzGateway(client_obj=c)
handle = c.sf.submit(req)
try:
    gateway._waitOnCmd(handle, failonerror=True)
    rsp = handle.getResponse()
finally:
    handle.close()

I am still getting LockTimeout?:

module body in test_status.py at line 39
gateway._waitOnCmd(handle, failonerror=True)
function _waitOnCmd in __init__.py at line 3382
callback.loop(loops, ms) # Throw LockTimeout
function loop in callbacks.py at line 321
5000L, int(waited))
copy output
Program exited with code #1 after 10.50 seconds.

[atarkowska]

Initial PR ​https://github.com/openmicroscopy/openmicroscopy/pull/2403. Please give comments.

[atarkowska]

resolved in ​https://github.com/openmicroscopy/openmicroscopy/pull/2403

[agilo]

Updated status, related task in progress

[agilo]

Updated status, related task in progress

[atarkowska]

sorted in ​https://github.com/openmicroscopy/openmicroscopy/pull/3434