fix: test if definition for all referenced glossary terms exist

[jeeyyy]

Added a test to check if content/ definition actually exists for a used glossary term in rules. In other words to check if a glossary referred actually exists as a file in /pages/glossary/*.md

@annethyme @Jym77 - could you help fix up the missing glossary items?

Failed tests report:

    glossary missing for #implicit-role, usages below:
    ┌─────────┬───────────────────────────────────────────┬────────────────┐
    │ (index) │                   name                    │      slug      │
    ├─────────┼───────────────────────────────────────────┼────────────────┤
    │    0    │ 'Role has required states and properties' │ 'rules/4e8ab6' │
    └─────────┴───────────────────────────────────────────┴────────────────┘

    glossary missing for #non-empty, usages below:
    ┌─────────┬───────────────────────────────────┬────────────────┐
    │ (index) │               name                │      slug      │
    ├─────────┼───────────────────────────────────┼────────────────┤
    │    0    │ 'Validity of HTML Lang attribute' │ 'rules/bf051a' │
    └─────────┴───────────────────────────────────┴────────────────┘

Closes issue:

• Partially closes Additional validations for rules #553

The changes requested does not aim to solve the actual problem of missing glossary definitions.

[WilcoFiers]

To fix the media ones: @JKODU

• #non-streaming -> #non-streaming-media-element
• #audio -> #audio-output
• #caption -> https://www.w3.org/TR/WCAG21/#dfn-captions

[Jym77]

Given that (i) our definition of "semantic role" already mention implicit ones; and (ii) that in the very same rule, "explicit semantic role" is already pointing to "semantic role":
I went for the quick fix, pointing "implicit semantic role" to "semantic role".

[jeeyyy]

Note: All definition references are fixed, waiting for #604 before this is merged.

[WilcoFiers]

This definitely isn't the way to solve it. I think we need a definition for this.

[Jym77]

Then, do we also need a definition for "explicit semantic role" which is used in the very same rule and simply links to "semantic role"?

The definition of semantic role has mention of both implicit and explicit roles, with links to further explanations. I assumed that if it was OK to link "explicit semantic role" to it, then it would also be OK for "implicit semantic role".

[Jym77]

@WilcoFiers I do not understand why it is fine to have [explicit semantic role](#semantic-role) (as was here before), but it is not fine to have [implicit semantic role](#semantic-role) Especially given that the glossary entry for semantic role mention both implicit and explicit roles, with links to more details.

Could you explain why one is OK and the other is not?
Should we rather add definitions for both explicit and implicit roles? (they are going to essentially be the same links that already exist in the definition of "Semantic Role")

[WilcoFiers]

You're right. We shouldn't have linked explicit semantic role to semantic role either. I'll open an issue for it, as to not block this PR any further.

[jeeyyy]

@WilcoFiers can you respond on this - #634 (comment)