Victor KuznetsovandClaude Opus 5.5 1dc70df289 Upgrade to OpenCV 5 with frozen Hershey templates and tighten video marks ...

- Lock OpenCV 5 and cap NumPy below 2.4 through uv constraint-dependencies so
  PaddleX 3.7 keeps resolving; Dependabot ignores mirror the cap.
- OpenCV 5 draws Hershey putText glyphs differently, so the Dola and Kling
  video templates ship as committed OpenCV 4 renders with a recipe script, and
  the OpenArt alpha recipe now refuses any OpenCV but 4.x.
- Break Sora/Veo cross-matches in favor of Veo under valid Google AI-video
  provenance, and require the full Sora mascot shape for mascot-only matches.
- Let the Doubao engine detect marks tinted by saturated backgrounds.
- Share C2PA claim-generator, credential-level and vendor-platform helpers,
  JPEG suffixes and the DCT basis instead of per-module copies.
- Log model download failures, document the optional GITHUB_TOKEN, mark the
  photo and source classifier model cards archived, and refresh the lock.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

2026-09-23 19:19:00 -07:00

..

__init__.py

Add project files, tests, and documentation for GitHub release

2026-03-25 11:15:05 -07:00

conftest.py

Validate C2PA credentials before attribution

2026-08-15 11:31:59 -07:00

test_agent_skill.py

Add concrete examples to agent skill

2026-09-09 20:11:46 -07:00

test_api.py

Add source pipeline classifier runtime

2026-09-15 23:17:48 -07:00

test_audioseal_experiment.py

Add audio and video watermark benchmark oracles, cohorts, and research studies

2026-09-07 18:23:18 -07:00

test_audioseal_oracle.py

Add audio and video watermark benchmark oracles, cohorts, and research studies

2026-09-07 18:23:18 -07:00

test_audit_image_contracts.py

Calibrate provider-specific invisible removal profiles

2026-09-15 16:38:17 -07:00

test_baidu_engine.py

Separate TC260 marks by manufacturer

2026-09-10 17:11:26 -07:00

test_c2pa_registry_sync.py

Authenticate C2PA registry-sync GitHub API calls to fix CI rate-limiting

2026-09-22 08:11:48 -07:00

test_classify.py

Pin published ONNX classifier snapshot

2026-09-10 20:07:33 -07:00

test_cli_robustness.py

Add the classify extra for the 2026-08-31 photo freeze. (#90)

2026-08-31 22:45:18 -07:00

test_cli.py

Separate runtime actions from development provenance checks

2026-09-06 20:18:47 -07:00

test_clip_l_ft.py

Add FP32 ONNX photo classifier backend

2026-09-10 19:31:24 -07:00

test_contentseal_corpus.py

Scrub private infrastructure detail and move the GPU harnesses out

2026-09-08 09:29:12 -07:00

test_cpu_offload.py

Calibrate provider-specific invisible removal profiles

2026-09-15 16:38:17 -07:00

test_dependabot.py

Upgrade to OpenCV 5 with frozen Hershey templates and tighten video marks

2026-09-23 19:19:00 -07:00

test_designer_synthid_suppression.py

Sync the C2PA soft-binding registry

2026-09-03 09:44:24 -07:00

test_distribution_boundary.py

Keep the boundary rule from naming what it protects

2026-09-08 10:43:20 -07:00

test_docs_cover_the_public_surface.py

Guard the docs seam, and close the eight knobs it had already lost

2026-09-01 15:33:26 -07:00

test_doubao_engine.py

Upgrade to OpenCV 5 with frozen Hershey templates and tighten video marks

2026-09-23 19:19:00 -07:00

test_engine_selection_analysis.py

Add measured auto engine selection

2026-09-01 10:19:59 -07:00

test_evaluation_audit_regressions.py

Pin Web oracle proxy sessions per batch

2026-09-16 16:17:36 -07:00

test_evaluation_data.py

Reject the uncalibrated text-manifest tiling and verify Content Seal transforms

2026-08-27 16:53:34 -07:00

test_export_photo_classify_onnx.py

Add FP32 ONNX photo classifier backend

2026-09-10 19:31:24 -07:00

test_fidelity_matching.py

Fix output preservation, research validation, and documentation drift

2026-09-08 19:07:10 -07:00

test_forensic_124d.py

Add the classify extra for the 2026-08-31 photo freeze. (#90)

2026-08-31 22:45:18 -07:00

test_forensic_metadata.py

Upgrade to OpenCV 5 with frozen Hershey templates and tighten video marks

2026-09-23 19:19:00 -07:00

test_gemini_engine.py

Document Gemini sparkle decorative-glyph false positive, add observability

2026-09-21 21:41:59 -07:00

test_generic_ai_label_engine.py

Add generic bare "AI生成" text-mark detector (brand-less TC260 fallback)

2026-09-21 22:01:10 -07:00

test_humanizer.py

Fix sharpening quantization and cross-platform regression tests

2026-09-08 19:25:44 -07:00

test_identify.py

Upgrade to OpenCV 5 with frozen Hershey templates and tighten video marks

2026-09-23 19:19:00 -07:00

test_image_io.py

Upgrade to OpenCV 5 with frozen Hershey templates and tighten video marks

2026-09-23 19:19:00 -07:00

test_image_math.py

Upgrade to OpenCV 5 with frozen Hershey templates and tighten video marks

2026-09-23 19:19:00 -07:00

test_infer_text_lines.py

Evaluate selective text restoration

2026-08-15 12:20:25 -07:00

test_inpaint_fallback.py

feat(visible): auto fill prefers LaMa > MI-GAN > cv2, warn on cv2 fallback

2026-07-09 18:24:20 +03:00

test_invisible_engine.py

Add SDXL verified-text restoration

2026-09-15 19:20:37 -07:00

test_invisible_watermark.py

Add real watermark fixtures and expand provider coverage

2026-09-05 13:12:58 -07:00

test_jimeng_engine.py

feat(visible): localize->fill rewrite, sensitivity/backend + api, HEIC + lossless IO

2026-07-09 14:20:52 +03:00

test_kling_engine.py

Separate TC260 marks by manufacturer

2026-09-10 17:11:26 -07:00

test_liblib_engine.py

Separate TC260 marks by manufacturer

2026-09-10 17:11:26 -07:00

test_maintenance.py

Fix output preservation, research validation, and documentation drift

2026-09-08 19:07:10 -07:00

test_metadata_audit_regressions.py

Fix output preservation, research validation, and documentation drift

2026-09-08 19:07:10 -07:00

test_metadata_internals.py

Attribute discovered C2PA signers

2026-09-14 11:51:19 -07:00

test_metadata_record.py

Complete corpus signal discovery fixes

2026-09-13 16:54:58 -07:00

test_metadata.py

Upgrade to OpenCV 5 with frozen Hershey templates and tighten video marks

2026-09-23 19:19:00 -07:00

test_microsoft_engine.py

Enable measured provenance relaxation for the Microsoft badge

2026-08-28 10:06:13 -07:00

test_microsoft_invismark_local.py

Add local Microsoft InvisMark payload disruption

2026-09-14 12:07:00 -07:00

test_openai_provenance.py

Add unified provider oracle tooling

2026-09-08 13:26:43 -07:00

test_openart_engine.py

Add OpenArt visible watermark detection and removal

2026-09-21 21:51:46 -07:00

test_optional_deps.py

Fix agent skill dependency checks and metadata guidance

2026-09-05 11:32:22 -07:00

test_packaging.py

Add FP32 ONNX photo classifier backend

2026-09-10 19:31:24 -07:00

test_pill_engine.py

Add real watermark fixtures and expand provider coverage

2026-09-05 13:12:58 -07:00

test_pixel_evidence.py

Add versioned forensic metadata transports

2026-08-05 21:10:55 -07:00

test_platform.py

Raise OpenAI Qwen strength floor

2026-09-21 23:02:49 -07:00

test_provider_oracle_microsoft.py

Add unified provider oracle tooling

2026-09-08 13:26:43 -07:00

test_provider_oracle_web.py

Pin Web oracle proxy sessions per batch

2026-09-16 16:17:36 -07:00

test_provider_oracles.py

Fix output preservation, research validation, and documentation drift

2026-09-08 19:07:10 -07:00

test_publish_photo_classify_hf.py

Upgrade to OpenCV 5 with frozen Hershey templates and tighten video marks

2026-09-23 19:19:00 -07:00

test_publish_source_classify_hf.py

Upgrade to OpenCV 5 with frozen Hershey templates and tighten video marks

2026-09-23 19:19:00 -07:00

test_qwen_engine.py

Release 0.40.1 with Qwen symbol false-positive fix

2026-09-11 10:55:28 -07:00

test_qwen_zimage_pipeline.py

Upgrade to OpenCV 5 with frozen Hershey templates and tighten video marks

2026-09-23 19:19:00 -07:00

test_region_eraser.py

Preserve 16-bit images through fill and metadata stripping (#101)

2026-09-05 13:11:16 -07:00

test_registered_mark_calibrate.py

Align watermark content with detector coverage

2026-08-27 20:05:03 -07:00

test_release_workflows.py

Fix Space pin waiter indent so distribute can find the run

2026-09-11 11:01:04 -07:00

test_review_detection_contracts.py

Fix output preservation, research validation, and documentation drift

2026-09-08 19:07:10 -07:00

test_runninghub_engine.py

Separate TC260 marks by manufacturer

2026-09-10 17:11:26 -07:00

test_samsung_engine.py

Improve visible-mark evaluation and Samsung masking

2026-09-05 11:51:02 -07:00

test_script_help.py

Add stratified watermark benchmark and resource profiling

2026-09-06 20:18:48 -07:00

test_security_clamp.py

Read WebP metadata past the scan window and surface C2PA reader failures

2026-08-05 21:10:38 -07:00

test_selective_text_restoration.py

Add verified text restoration

2026-08-15 12:25:35 -07:00

test_signal_discovery_scripts.py

Attribute discovered C2PA signers

2026-09-14 11:51:19 -07:00

test_skill_evals.py

Keep the skill-eval harness tests portable on Windows

2026-09-05 11:27:17 -07:00

test_source_classify.py

Add source pipeline classifier runtime

2026-09-15 23:17:48 -07:00

test_source_pipeline_classifier.py

Publish abstaining source pipeline classifier

2026-09-15 19:18:35 -07:00

test_spectral_pipeline_probe.py

Add phase-free spectral pipeline probe

2026-09-15 16:20:31 -07:00

test_synthid_adaptive_carrier_suppress.py

Merge the docs/arxiv-paper-review research line into main (#95)

2026-09-01 11:37:32 -07:00

test_synthid_affine_lattice_probe.py

Fix output preservation, research validation, and documentation drift

2026-09-08 19:07:10 -07:00

test_synthid_audit_regressions.py

Fix output preservation, research validation, and documentation drift

2026-09-08 19:07:10 -07:00

test_synthid_color_space_probe.py

Merge the docs/arxiv-paper-review research line into main (#95)

2026-09-01 11:37:32 -07:00

test_synthid_confirmation.py

Merge the docs/arxiv-paper-review research line into main (#95)

2026-09-01 11:37:32 -07:00

test_synthid_conformal_cascade.py

Merge the docs/arxiv-paper-review research line into main (#95)

2026-09-01 11:37:32 -07:00

test_synthid_confound_probe.py

Merge the docs/arxiv-paper-review research line into main (#95)

2026-09-01 11:37:32 -07:00

test_synthid_consensus_probe.py

Merge the docs/arxiv-paper-review research line into main (#95)

2026-09-01 11:37:32 -07:00

test_synthid_corpus.py

Fix SynthID provenance evidence and release 0.26.1

2026-08-06 18:17:40 -07:00

test_synthid_cyclostationary_probe.py

Merge the docs/arxiv-paper-review research line into main (#95)

2026-09-01 11:37:32 -07:00

test_synthid_detector.py

Merge the docs/arxiv-paper-review research line into main (#95)

2026-09-01 11:37:32 -07:00

test_synthid_ensemble_attack.py

Merge the docs/arxiv-paper-review research line into main (#95)

2026-09-01 11:37:32 -07:00

test_synthid_ensemble_detector.py

Merge the docs/arxiv-paper-review research line into main (#95)

2026-09-01 11:37:32 -07:00

test_synthid_fragment_attack.py

Merge the docs/arxiv-paper-review research line into main (#95)

2026-09-01 11:37:32 -07:00

test_synthid_hybrid_attack.py

Merge the docs/arxiv-paper-review research line into main (#95)

2026-09-01 11:37:32 -07:00

test_synthid_oracle_batch.py

Merge the docs/arxiv-paper-review research line into main (#95)

2026-09-01 11:37:32 -07:00

test_synthid_periodic_tile_ablation.py

Merge the docs/arxiv-paper-review research line into main (#95)

2026-09-01 11:37:32 -07:00

test_synthid_periodic_tile_probe.py

Merge the docs/arxiv-paper-review research line into main (#95)

2026-09-01 11:37:32 -07:00

test_synthid_phase_carrier.py

Merge the docs/arxiv-paper-review research line into main (#95)

2026-09-01 11:37:32 -07:00

test_synthid_pixel_attack.py

Merge the docs/arxiv-paper-review research line into main (#95)

2026-09-01 11:37:32 -07:00

test_synthid_pixel_probe.py

Fix output preservation, research validation, and documentation drift

2026-09-08 19:07:10 -07:00

test_synthid_research_inventory.py

Merge the docs/arxiv-paper-review research line into main (#95)

2026-09-01 11:37:32 -07:00

test_synthid_research_manifest.py

Fix output preservation, research validation, and documentation drift

2026-09-08 19:07:10 -07:00

test_synthid_routed_detector.py

Fix output preservation, research validation, and documentation drift

2026-09-08 19:07:10 -07:00

test_synthid_routed_expert_bank.py

Fix output preservation, research validation, and documentation drift

2026-09-08 19:07:10 -07:00

test_synthid_runtime_expert_scores.py

Fix output preservation, research validation, and documentation drift

2026-09-08 19:07:10 -07:00

test_synthid_spectral_probe.py

Merge the docs/arxiv-paper-review research line into main (#95)

2026-09-01 11:37:32 -07:00

test_synthid_tile_attack.py

Merge the docs/arxiv-paper-review research line into main (#95)

2026-09-01 11:37:32 -07:00

test_synthid_v3_carrier_subtract.py

Merge the docs/arxiv-paper-review research line into main (#95)

2026-09-01 11:37:32 -07:00

test_synthid_v3_codebook_probe.py

Merge the docs/arxiv-paper-review research line into main (#95)

2026-09-01 11:37:32 -07:00

test_text_draft.py

Fix output preservation, research validation, and documentation drift

2026-09-08 19:07:10 -07:00

test_text_mark_engine.py

Improve Samsung detector recall on textured images

2026-09-05 11:51:02 -07:00

test_text_mark_faint_mask.py

Align watermark content with detector coverage

2026-08-27 20:05:03 -07:00

test_text_mark_memory.py

feat(visible): localize->fill rewrite, sensitivity/backend + api, HEIC + lossless IO

2026-07-09 14:20:52 +03:00

test_text_restoration.py

Fix adaptive text restoration crop bounds

2026-08-25 12:44:05 -07:00

test_tiling.py

Align watermark content with detector coverage

2026-08-27 20:05:03 -07:00

test_trustmark_detector.py

Add stratified watermark benchmark and resource profiling

2026-09-06 20:18:48 -07:00

test_video_fidelity_probe.py

Align watermark content with detector coverage

2026-08-27 20:05:03 -07:00

test_video_invisible.py

Reject the uncalibrated text-manifest tiling and verify Content Seal transforms

2026-08-27 16:53:34 -07:00

test_video_synthid_sweep.py

Add streaming video SynthID regeneration

2026-07-29 22:03:23 -07:00

test_video_temporal.py

feat: complete product video watermark pipeline

2026-07-31 10:42:39 -07:00

test_video.py

Upgrade to OpenCV 5 with frozen Hershey templates and tighten video marks

2026-09-23 19:19:00 -07:00

test_videoseal_oracle.py

Fix output preservation, research validation, and documentation drift

2026-09-08 19:07:10 -07:00

test_visible_examples.py

Upgrade to OpenCV 5 with frozen Hershey templates and tighten video marks

2026-09-23 19:19:00 -07:00

test_visible_fill_quality.py

Fix missing OpenArt gallery fixture and alternate-size entry

2026-09-21 22:37:21 -07:00

test_watermark_benchmark_audio_cohort.py

Add audio and video watermark benchmark oracles, cohorts, and research studies

2026-09-07 18:23:18 -07:00

test_watermark_benchmark_cohort.py

Add stratified watermark benchmark and resource profiling

2026-09-06 20:18:48 -07:00

test_watermark_benchmark_real_cohort.py

Add stratified watermark benchmark and resource profiling

2026-09-06 20:18:48 -07:00

test_watermark_benchmark_report.py

Add audio and video watermark benchmark oracles, cohorts, and research studies

2026-09-07 18:23:18 -07:00

test_watermark_benchmark_resources.py

Add stratified watermark benchmark and resource profiling

2026-09-06 20:18:48 -07:00

test_watermark_benchmark_video_cohort.py

Add audio and video watermark benchmark oracles, cohorts, and research studies

2026-09-07 18:23:18 -07:00

test_watermark_benchmark.py

Add audio and video watermark benchmark oracles, cohorts, and research studies

2026-09-07 18:23:18 -07:00

test_watermark_registry.py

Add generic bare "AI生成" text-mark detector (brand-less TC260 fallback)

2026-09-21 22:01:10 -07:00

test_yuanbao_engine.py

Add Tencent Yuanbao visible watermark removal

2026-07-25 20:52:15 -07:00