Migrate Kik parser to the LAVA pipeline (@artifact_processor)

[OneSixForensics]

Migrate the Kik parser to the LAVA pipeline

kik.py (merged earlier) hand-builds ArtifactHtmlReport objects, so its artifacts render in HTML but never populate the LAVA database — they don't appear in the LAVA viewer. This rebuilds the module onto the modern @artifact_processor decorator so all artifacts emit HTML + TSV + timeline + LAVA from a single (data_headers, data_list, source_path) return. No parsing logic changed.

Changes

• Functions renamed to match their __artifacts_v2__ keys (the decorator and plugin loader key off the function name).
• Subscriber PDF split: the subscriber function previously wrote three tables (Info, Profile Pics, Events) but only kik_subscriber was a declared artifact, so Profile Pics and Events would never reach LAVA. Split into three declared artifacts (kik_subscriber / kik_subscriber_pics / kik_subscriber_events) sharing a cached PDF text extract. report_icons.py already contained icon entries for all three.
• Chat media: now a typed ('Media','media') column registered via an exact seeker.file_infos lookup (handles the Base64-encoded medias/ filenames), rendering inline in both HTML and LAVA. Referenced-but-absent media shows the filename with an empty media cell.
• Human-readable timestamp columns typed ('…','datetime') for the LAVA timeline.
• Schema-fallback (raw passthrough on unexpected headers) preserved.

Artifact count goes 13 → 15 (the two new subscriber sub-artifacts).

Validation

End-to-end on a synthetic Kik return (a generated subscriber PDF + every CSV + a real media file): all 15 artifacts register in LAVA with correct row counts, and the media item/reference is created (real file linked, absent file shown by name).

Real-return QA recommended before merge — I validated against a faithful synthetic return; if you have a sanitized Kik return to confirm against, even better.

Companion to the Synchronoss parser PR (same LAVA-migration recipe).