Description

Update the archive status, video metadata completeness, transcript loop report, and archive manager transcript auditing callers so they use a forthcoming shared Generators::ArchiveState module when it is available at src/generators/archive_state.rb. The callers should treat structured transcript turns as valid transcript content, normalize text consistently for reports, and surface YAML parse failures instead of silently counting invalid transcripts as complete. Do not create the shared module in this task.

Acceptance Criteria

• #1 Archive status counts transcript assets that have either scalar content or structured turns.
• #2 Video metadata completeness marks transcript completeness true for structured turns with reasonable joined text length and false for invalid YAML.
• #3 Transcript loop reporting analyzes normalized text from either scalar content or joined turns.
• #4 Archive manager transcript auditing delegates transcript-content validity to Generators::ArchiveState without creating that module.
• #5 Invalid YAML parse failures are visible in status/report output where feasible and are not counted as complete.
• #6 Only the named generator/report files are edited unless a tiny targeted spec addition is necessary.
• #7 Targeted scripts or specs are run when feasible and results are recorded.

Implementation Plan

1. Read the named scripts, archive manager, current generated YAML outputs, nearby transcript examples, and the collaborator-provided src/generators/archive_state.rb API.
2. Patch only the named callers to require src/generators/archive_state.rb and use the state-object API: Generators::ArchiveState.for_path/for_id, state.has_transcript?, state.text, state.invalid?, and state.load_error.
3. Preserve existing report shapes while adding parse-error counts/details where feasible and ensuring invalid transcript YAML is not counted as complete.
4. Avoid regenerating _data or tmp reports; verify with syntax checks, read-only count probes, the archive auditor path, and the collaborator’s focused ArchiveState spec.

Implementation Notes

Implemented the caller-side ArchiveState integration in the four requested files only. src/generators/archive_state.rb appeared during the turn as collaborator work; I read its public API and did not edit it. Verification included Ruby syntax checks for all four modified files, read-only state/count probes, a read-only TranscriptAuditor run, and bundle exec rspec spec/src/generators/archive_state_spec.rb.

Final Summary

Updated archive/report callers to consume the shared archive state API instead of directly inspecting transcript['content'].

Changed files:

• bin/generate_archive_status.rb: counts assets whose transcript state has text from either legacy content or structured turns; records transcript parse errors in the generated status payload.
• bin/generate_video_metadata_completeness.rb: rates transcript completeness from ArchiveState text, so structured turns can be complete when long enough; invalid transcript YAML becomes a partial transcript rating with an error field and summary count.
• bin/report_transcript_loops.rb: analyzes ArchiveState text for either content or joined turns and includes parse-error counts/details in JSON/Markdown reports.
• src/generators/archive_manager.rb: delegates missing-content detection to ArchiveState and uses invalid state errors for invalid transcript files.

No generated _data or tmp reports were regenerated. Verification: four ruby -c syntax checks passed; read-only probes found assets_with_transcript_content=195, with_transcript_complete_probe=193, short_transcripts_probe=2, transcript_parse_errors=1; archive auditor read-only run reported missing_files=0 missing_content=0 invalid_files=1; bundle exec rspec spec/src/generators/archive_state_spec.rb passed with 8 examples.

Definition of Done

• #1 AC criteria is completed and the change has been verified