Media Manager - check_in_media() Limitations

[kobo220]

I am currently working on an artifact plugin for a chat app. I am trying to avoid having too permissive of a regex path search (in my case: /private/var/mobile/Containers/Shared/AppGroup/{GUID}/Attachments/{FILES}). I figured I would use Context.seeker.search() to get my attachments so I could avoid bringing in every app that has an Attachments folder. The problem is check_in_media() cannot be used with files that are not found as part of the initial regex path search in the artifact header dict. Any files found manually by calling seeker.search() don't get added to the Context's _files_found or _filename_lookup_map. check_in_media() and _check_in_media() both use: Context.get_source_file_path({path}).

I created a new static method in Context.py that I can pass the results from seeker.search() into for testing, which allowed me to make use of check_in_media() and my artifact results appear as expected in the table view and conversation view (LAVA).

@staticmethod
    def add_files_to_context(new_files):
        if not isinstance(new_files, list):
            new_files = [new_files]
        
        Context._files_found.extend(new_files)
        Context._filename_lookup_map = {}

I don't know the best way to tackle this problem. I'm not sure if adding something to the Context is best, some change to the Media Manager would be better, or if I'm missing some other existing way to handle this.

[JamesHabben]

Hi @kobo220, I ran into this a little while back realizing the same limitation. In that case I was able to work around it by including a pattern in the header, but it makes sense that not all modules have a unique enough path to be reasonable. I haven't gone through all of this to have an opinion yet on the best way forward but I see it needs to be addressed. Thanks for sending this in to start the discussion. I'll come back with some thoughts.

[JamesHabben]

Also, tagging @Johann-PLW as he has done the file mapping work and some media manager work. We got some really great speed boosts with the work in those areas, but it's causing issues with this scenario. Whatever solution we go towards, we need to keep this in mind to avoid regression on that speed.

[abrignoni]

One possible solution @kobo220 is to use check_in_embedded_media to get the media data displayed without having to iterate over files_found for anything. I use this to convert images in one format to another without writing anything to the filesystem and then get the media displayed. It would be trivial to use it in conjunction with the seeker function i think.

Just a thought...

Check in embedded media.
Args:
    source_file: The source file path of the embedded media data.
    data: The bytes of the embedded media data.
    name: The name of the media (optional).
    force_type: The MIME type of the media (optional).
    force_extension: The extension of the media (optional).
    force_creation_date: The creation date of the media (optional).
    force_modification_date: The modification date of the media (optional).
Returns:
    The media reference ID or None.