# process_pose_data
Tools for fetching, processing, visualizing, and analyzing Wildflower human pose data
## Task list
* Switch parallel overlay code back to `imap_unordered()` (for less chunky progress bars) but sort output before concatenating
* Ensure that all visual specs (colors, line widths, etc.) propagate to video overlay
* Add drawing primitive to `wf-cv-utils` for text with background
* Use new text-with-background drawing primitive for pose labels
* Add timestamp to video overlays
* Consider getting rid of `geom_render` module
* Fix up progress bars so we can have task-level and segment-level progress bars without logging in notebook
* Fix up progress bars so they work properly outside of notebook with multiprocessing
* Add indentation/labeling on segment-level progress bars
* Clean up argument ordering in `reconstruct_poses_3d_alphapose_local_time_segment`
* Add `poses_2d_json_format` option to `reconstruct_poses_3d` entry point
* Add endpoints for all processing functions
* Add ability to write locally generated object IDs to Honeycomb
* Create separate workers for uploading to Honeycomb 3D poses, 3D pose tracks, interpolated 3D pose tracks, 3D pose identification, 3D pose track identification
* Dockerize pipeline
* Set up pipeline for Airflow
* Add additional machinery for checking UWB data integrity (e.g., duplicates)
* Retool `generate_inference_metadata_reconstruct_3d_poses_alphapose_local` to exclude cameras without calibration data
* Make function to delete Honeycomb inference executions
* Make function to delete local inference metadata
* Make function to delete local 3D pose files
* Make function for deleting local 3D pose data
* Rewrite all log messages so formatting isn't called if log isn't printed
* Make functions handle empty poses (all keypoints `NaN`) more gracefully (e.g., `score_pose_pairs()`, `draw_pose_2d()`)
* Make visualization functions handle missing fields (e.g., `pose_quality`) more gracefully
* Figure out inconsistent behavior of `groupby(...).apply(...)` (under what conditions does it add grouping variables to index?)
* For functions that act on dataframes, make it optional to check dataframe structure (e.g., only one timestamp and camera pair)
* For functions than iterate over previous functions, making naming and approach consistent (e.g., always use apply?)
* Add `keypoint_categories` info to pose models in Honeycomb?
* Be consistent about whether to convert track labels to integers (where possible)
* Remove dependence on OpenCV by adding necessary functionality to `cv_utils`
* Consider refactoring split between `video_io` and `cv_utils`
* Fix up `cv_utils` Matplotlib drawing functions so that they accept an axis (or figure, as appropriate)
* Fix up handling of background image alpha (shouldn't assume white background)
* Fix up _y_ axis inversion for images (go back to `cv_utils`?)
* Add option of specifying Honeycomb client info for visualization functions that require Honeycomb
* Reinstate `sns.set()` for Seaborn plots without making it spill over into non-Seaborn plots (see [here](https://stackoverflow.com/questions/26899310/python-seaborn-to-reset-back-to-the-matplotlib))
* Refactor code in `visualize` to make it less repetitive (same pattern over and over for `[verb]_by_camera`)
* Fix up legend on pose track timelines
* Add visualization for number of poses per camera per timestamp
* Replace `cv.triangulatePoints()` to increase speed (and hopefully accuracy)
* Get pose video overlays working again (for data with track labels)
* Rewrite geom rendering functions to handle the possibility of no track labels
* Rewrite function which overlays geoms on videos so that user can specify a time span that it is a subset of the geoms and/or the video
* Make all time inputs more permissive (in terms of type/format) and make all time outputs more consistent
* Be consistent about accepting timestamp arguments in any format parseable by `pd.to_datetime()`
Raw data
{
"_id": null,
"home_page": "https://github.com/WildflowerSchools/wf-process-pose-data",
"name": "wf-process-pose-data",
"maintainer": "",
"docs_url": null,
"requires_python": "",
"maintainer_email": "",
"keywords": "pose reconstruction,pose tracking",
"author": "Theodore Quinn",
"author_email": "ted.quinn@wildflowerschools.org",
"download_url": "https://files.pythonhosted.org/packages/f3/2e/b32beb54385f23333151b4a2e9fa5c742b450bba50cfdc56ecd0f32452d0/wf-process-pose-data-6.3.0.tar.gz",
"platform": null,
"description": "# process_pose_data\n\nTools for fetching, processing, visualizing, and analyzing Wildflower human pose data\n\n## Task list\n\n* Switch parallel overlay code back to `imap_unordered()` (for less chunky progress bars) but sort output before concatenating\n* Ensure that all visual specs (colors, line widths, etc.) propagate to video overlay\n* Add drawing primitive to `wf-cv-utils` for text with background\n* Use new text-with-background drawing primitive for pose labels\n* Add timestamp to video overlays\n* Consider getting rid of `geom_render` module\n* Fix up progress bars so we can have task-level and segment-level progress bars without logging in notebook\n* Fix up progress bars so they work properly outside of notebook with multiprocessing\n* Add indentation/labeling on segment-level progress bars\n* Clean up argument ordering in `reconstruct_poses_3d_alphapose_local_time_segment`\n* Add `poses_2d_json_format` option to `reconstruct_poses_3d` entry point\n* Add endpoints for all processing functions\n* Add ability to write locally generated object IDs to Honeycomb\n* Create separate workers for uploading to Honeycomb 3D poses, 3D pose tracks, interpolated 3D pose tracks, 3D pose identification, 3D pose track identification\n* Dockerize pipeline\n* Set up pipeline for Airflow\n* Add additional machinery for checking UWB data integrity (e.g., duplicates)\n* Retool `generate_inference_metadata_reconstruct_3d_poses_alphapose_local` to exclude cameras without calibration data\n* Make function to delete Honeycomb inference executions\n* Make function to delete local inference metadata\n* Make function to delete local 3D pose files\n* Make function for deleting local 3D pose data\n* Rewrite all log messages so formatting isn't called if log isn't printed\n* Make functions handle empty poses (all keypoints `NaN`) more gracefully (e.g., `score_pose_pairs()`, `draw_pose_2d()`)\n* Make visualization functions handle missing fields (e.g., `pose_quality`) more gracefully\n* Figure out inconsistent behavior of `groupby(...).apply(...)` (under what conditions does it add grouping variables to index?)\n* For functions that act on dataframes, make it optional to check dataframe structure (e.g., only one timestamp and camera pair)\n* For functions than iterate over previous functions, making naming and approach consistent (e.g., always use apply?)\n* Add `keypoint_categories` info to pose models in Honeycomb?\n* Be consistent about whether to convert track labels to integers (where possible)\n* Remove dependence on OpenCV by adding necessary functionality to `cv_utils`\n* Consider refactoring split between `video_io` and `cv_utils`\n* Fix up `cv_utils` Matplotlib drawing functions so that they accept an axis (or figure, as appropriate)\n* Fix up handling of background image alpha (shouldn't assume white background)\n* Fix up _y_ axis inversion for images (go back to `cv_utils`?)\n* Add option of specifying Honeycomb client info for visualization functions that require Honeycomb\n* Reinstate `sns.set()` for Seaborn plots without making it spill over into non-Seaborn plots (see [here](https://stackoverflow.com/questions/26899310/python-seaborn-to-reset-back-to-the-matplotlib))\n* Refactor code in `visualize` to make it less repetitive (same pattern over and over for `[verb]_by_camera`)\n* Fix up legend on pose track timelines\n* Add visualization for number of poses per camera per timestamp\n* Replace `cv.triangulatePoints()` to increase speed (and hopefully accuracy)\n* Get pose video overlays working again (for data with track labels)\n* Rewrite geom rendering functions to handle the possibility of no track labels\n* Rewrite function which overlays geoms on videos so that user can specify a time span that it is a subset of the geoms and/or the video\n* Make all time inputs more permissive (in terms of type/format) and make all time outputs more consistent\n* Be consistent about accepting timestamp arguments in any format parseable by `pd.to_datetime()`\n\n\n",
"bugtrack_url": null,
"license": "",
"summary": "Tools for fetching, processing, visualizing, and analyzing Wildflower human pose data",
"version": "6.3.0",
"project_urls": {
"Homepage": "https://github.com/WildflowerSchools/wf-process-pose-data"
},
"split_keywords": [
"pose reconstruction",
"pose tracking"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "8b55101d741428acd90437f25db5888235a7c24d18723a08ed4586c400bafd48",
"md5": "aed7d3ae1e417c6a78f9f71442c2c8e3",
"sha256": "453eeeb8c559b229cb30aedbc5409838fb7580de2c008fc77eab73c31da9f51a"
},
"downloads": -1,
"filename": "wf_process_pose_data-6.3.0-py3-none-any.whl",
"has_sig": false,
"md5_digest": "aed7d3ae1e417c6a78f9f71442c2c8e3",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": null,
"size": 49814,
"upload_time": "2023-12-12T18:20:35",
"upload_time_iso_8601": "2023-12-12T18:20:35.638660Z",
"url": "https://files.pythonhosted.org/packages/8b/55/101d741428acd90437f25db5888235a7c24d18723a08ed4586c400bafd48/wf_process_pose_data-6.3.0-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "f32eb32beb54385f23333151b4a2e9fa5c742b450bba50cfdc56ecd0f32452d0",
"md5": "ad1888e7dae193f87f2b0a25f746b2ff",
"sha256": "86a64b05298bcd643c86c11437e11af9328af747eabcc8b1c6ec35cf6bd6d5bb"
},
"downloads": -1,
"filename": "wf-process-pose-data-6.3.0.tar.gz",
"has_sig": false,
"md5_digest": "ad1888e7dae193f87f2b0a25f746b2ff",
"packagetype": "sdist",
"python_version": "source",
"requires_python": null,
"size": 49393,
"upload_time": "2023-12-12T18:20:40",
"upload_time_iso_8601": "2023-12-12T18:20:40.687867Z",
"url": "https://files.pythonhosted.org/packages/f3/2e/b32beb54385f23333151b4a2e9fa5c742b450bba50cfdc56ecd0f32452d0/wf-process-pose-data-6.3.0.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2023-12-12 18:20:40",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "WildflowerSchools",
"github_project": "wf-process-pose-data",
"travis_ci": false,
"coveralls": false,
"github_actions": false,
"lcname": "wf-process-pose-data"
}
JSON
