# Backlog

## Iteration +1
- Ship version 0.0.0: PyPI and OCI

## Iteration +2
- Operate on folders including multiple files
- IO: Read and write ROS Bags files for compatibility with SARCAM environment
  - https://wiki.ros.org/Bags
  - https://pypi.org/project/rosbag-tools/
  - https://pypi.org/project/rosbags/
  - https://pypi.org/project/rosbags-image/
  - https://pypi.org/project/rosbags-dataframe/
  - https://gitlab.com/ternaris/marv-robotics/-/blob/master/code/marv-robotics/marv_robotics/bag.py
- IO: RTSP/File in, HLS out. For easy dissemination and archival without too much
  overhead on implementing infrastructure and frontend.
- Error handling: If processing fails, e.g. because of bad input file,
  the program currently does not exit with a corresponding exit code.
- Docker: Suppress:
  - `[ONNXRuntimeError] : 1 : FAIL : Failed to load library libonnxruntime_providers_tensorrt.so with error: libcudnn.so.8: cannot open shared object file: No such file or directory`
  - `[ONNXRuntimeError] : 1 : FAIL : Failed to load library libonnxruntime_providers_cuda.so with error: libcublasLt.so.11: cannot open shared object file: No such file or directory`
- https://docs.gitlab.com/ee/ci/testing/unit_test_reports.html
- https://medium.com/@yian806884/comprehensive-guide-to-gitlab-ci-cd-081305b06a9a
- CI: Stable release builds when tagging
- CI: Nightly builds per schedule
- CI: Use `CI_COMMIT_TAG`: https://github.com/woodpecker-ci/woodpecker/pull/2934
- CI: More Codeberg / Woodpecker CI
  - https://woodpecker-ci.org/plugins/Docker%20Tags
  - https://github.com/woodpecker-ci/woodpecker/discussions/4246
  - https://woodpecker-ci.org/plugins/Codecov
  - https://woodpecker-ci.org/plugins/Twine

## Iteration +3
- IO: Make output codec configurable. Currently `DIVX`.
- Core: Adjust default value for `--fps=5.0` back to `14.6` or `20`?
- Util: Generic media converter (image sequence vs. video)
- Unlock OpenCV-based ROI detector again?
- Feature: Web frontend based on HLS player
- Feature: Audit log / Report output
- Feature: Save frames of interest
  https://stackoverflow.com/questions/30136257/how-to-get-image-from-video-using-opencv-python
- CI: Upload code coverage data to Codecov.io
- CI: Dependabot
  - https://dependabot-gitlab.gitlab.io/dependabot/
  - https://dependabot-gitlab.gitlab.io/dependabot/guide/standalone-mode.html
  - https://gitlab.com/dependabot-gitlab/dependabot-standalone

## Iteration +4
- Event telemetry: MQTT and NMEA
- Notifications
- Docs: Philosophy: Bazaar, DWIM, KISS, RORO
- Docs: Add info about ffmpeg and GStreamer
- Docs: 
  - Provides end-to-end integration test cases together with canonical `datasets` repository
  - Taking the role as a package maintainer
- Docs: https://sarcam.flova.de/
- Docs: https://events.ccc.de/congress/2024/hub/de/assembly/sarcam/
- Are there other models than `yolov10m.onnx`?

## Done
- Review `ARG002 Unused method argument:` mask markers
- Small-scale in-a-box image processing pipeline: Yes, make it work.
- Read directory of images and/or single image
- Add software tests
- Caching for datasets to avoid redundant downloading
- Test more IO variants of processing (video vs. bitmap sequence)
- Record frame number and possibly other information within `DetectionEvent`
- Rename from `nauscopy` to `nauscopie`
- Last minute fixes and improvements
- Rename from `nauscopie` to `nauscopia`
- Documentation @ RTD: https://nauscopia.readthedocs.io/
- README: Badges
- Migrate to Codeberg / Woodpecker CI
  - https://woodpecker-ci.org/plugins/Git%20Clone
  - https://woodpecker-ci.org/plugins/Docker%20Buildx