support for arbitrary URI schemes in `clams source`

[keighrim]

New Feature Summary

As a user, I would like to a convenient way to generate a "source" MMIF with document location URI with some arbitrary scheme other than file://, using clams source command.

Current clams source command is only able to generate documents with file:// locations, so this will be upgrading the feature so that it can generate any location value, with associated document type, but keeping the default behavior as using file:// schema. Here's an example of current behavior

$ clams source video:/data/video.mp4
{
 "metadata": {
 "mmif": "http://mmif.clams.ai/1.0.0"
 },
 "documents": [
 {
 "@type": "http://mmif.clams.ai/vocabulary/VideoDocument/v1",
 "properties": {
 "mime": "video",
 "id": "d1",
 "location": "file:///data/video.mp4"
 }
 }
 ],
 "views": []

And I guess we can add one more CLI argument to make it work like this

$ clams source video:something.something.video --scheme baapb
{
 "metadata": {
 "mmif": "http://mmif.clams.ai/1.0.0"
 },
 "documents": [
 {
 "@type": "http://mmif.clams.ai/vocabulary/VideoDocument/v1",
 "properties": {
 "mime": "video",
 "id": "d1",
 "location": "baapb://something.something.video"
 }
 }
 ],
 "views": []

The source code for the clams source cmd is at

https://github.com/clamsproject/clams-python/blob/a7ce71bae59d2e359cb6d2a14e558638778dd51f/clams/source/__init__.py

Related

clamsproject/mmif-python#222
also, https://github.com/clamsproject/mmif-docloc-baapb

Alternatives

No response

Additional context

No response

[keighrim]

The proposed plan and its initiali implementation (b0a5e23) do not supported mixed-scheme scenarios. Namely, we can't generate a source MMIF like following with a single clams source command from the current implementation in #163 .

{
 "metadata": {
 "mmif": "http://mmif.clams.ai/1.0.0"
 },
 "documents": [
 {
 "@type": "http://mmif.clams.ai/vocabulary/AudioDocument/v1",
 "properties": {
 "mime": "audio",
 "id": "d1",
 "location": "file:///data/audio.mp3"
 }
 }, 
 {
 "@type": "http://mmif.clams.ai/vocabulary/VideoDocument/v1",
 "properties": {
 "mime": "video",
 "id": "d2",
 "location": "baapb://something.something.video"
 }
 }
 ],
 "views": []

Here's an idea to address this issue.

• --scheme flag will set the default scheme to use. When it's absent, default scheme is set to its default of file:// (already implemented)
• users can pass a full URI after the <media_type>: prefix. When the input is already in a URI format with a scheme encoded, the scheme set by --scheme flag is ignored and the user input is directly recorded in the output MMIF.
• with this implementation, a user can generate a mixed MMIF like the above by

clams source video:file:///data/video.mp4 video:something.something.video --scheme baapb 
# OR 
clams source video:/data/video.mp4 video:baapb://something.something.video

• this would requires a check if (a part of )the input string is valid URI format.

@leorabaumgarten what do you think?