fix(#280): align e/p tag report types with nip 56 standard values

[DSanich]

Summary

• Align report event e/p tag third values with NIP-56 standard report types.
• Preserve app-level report categorization via existing NIP-32 L/l tags (social.nos.ontology, NS-*).
• Add targeted tests for mapping correctness and report tag contract on kind 1984 events.

Problem

useReportContent() was writing app-level reasons (for example csam, sexual-content, harassment) directly into the e/p tag third element.

NIP-56 requires these values to be one of:
nudity, malware, profanity, illegal, spam, impersonation, other.

What Changed

• Added an explicit mapping from ContentFilterReason to NIP-56 report types.
• Updated useReportContent() to use mapped NIP-56 values in:
  • ["e", "<eventId>", "<nip56-type>"]
  • ["p", "<pubkey>", "<nip56-type>"]
• Kept NIP-32 label tags unchanged so app-level semantics remain queryable:
  • ["L", "social.nos.ontology"]
  • ["l", "NS-<app-reason>", "social.nos.ontology"]
• Added unit tests covering:
  • full mapping table
  • published report tags for kind 1984
  • regression checks to ensure raw app-level reasons are not used in e/p

Related

• fix(reports): align e/p tag report types with NIP-56 standard values #280
• feat(reports): add NIP-32 l/L tags to restore app-level report categorization divine-mobile#3492
• feat(reports): add NS-prefixed CATEGORY_LABELS for mobile report reasons divine-relay-manager#50

Closes #280

[github-actions]

🚀 Preview Deployment

Property	Value
Preview URL	https://3087de7d.divine-web-fm8.pages.dev
Commit	9cd12a5
Branch	fix/280-align-e/p-tag-report-types-with-NIP-56-standard-values

[mbradley]

Clean fix -- mapping is exhaustive, tests cover the full contract, and the NS-* NIP-32 labels are preserved unchanged so relay-manager's report UI won't lose any categorization fidelity.

One minor observation: HARASSMENT to profanity is the best available NIP-56 match, but it's worth noting that this is a lossy mapping by design. The NS-harassment label in the l tag is what carries the actual intent. Relay operators or other clients consuming only the e/p tag will see "profanity" where the user meant "harassment."

Approving -- this is ready to merge.

[DSanich]

mbradley
Clean fix -- mapping is exhaustive, tests cover the full contract, and the NS-* NIP-32 labels are preserved unchanged so relay-manager's report UI won't lose any categorization fidelity.

One minor observation: HARASSMENT to profanity is the best available NIP-56 match, but it's worth noting that this is a lossy mapping by design. The NS-harassment label in the l tag is what carries the actual intent. Relay operators or other clients consuming only the e/p tag will see "profanity" where the user meant "harassment."

Approving -- this is ready to merge.

@mbradley TYSM for your review! I will wait new assignments from you, @jalcine and @NotThatKindOfDrLiz

[jalcine]

These changes are already handled now upstream!