vidhimudaliar commited on
Commit
7939242
·
verified ·
1 Parent(s): c13840d

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +91 -1
README.md CHANGED
@@ -7,4 +7,94 @@ sdk: static
7
  pinned: false
8
  ---
9
 
10
- Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7
  pinned: false
8
  ---
9
 
10
+ Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
11
+
12
+ # Human-Robot Collaboration Failure Browser
13
+
14
+ - browse all failure categories
15
+ - open a category detail page with its definition and representative example
16
+ - inspect every coded segment attached to that category
17
+ - search all coded segments across the dataset
18
+
19
+ ## Project Structure
20
+
21
+ ```text
22
+ .
23
+ |-- index.html # Homepage with failure category cards
24
+ |-- category.html # Detail page for one failure category
25
+ |-- all_videos.html # Searchable table of all coded segments
26
+ |-- assets/
27
+ | |-- css/ # Shared and page-specific styles
28
+ | `-- js/
29
+ | |-- failures.js # Failure definitions and metadata
30
+ | |-- homepage.js # Homepage search + category stats
31
+ | |-- category.js # Category page rendering
32
+ | |-- all_videos.js # All-videos table rendering
33
+ | `-- video_segment.js # Video loop/playback behavior
34
+ |-- raw_data.csv # Original Label Studio export
35
+ |-- cleaned_data.csv # Flattened CSV used by the web app
36
+ |-- clean.py # Converts raw_data.csv -> cleaned_data.csv
37
+ `-- videos/ # Source MP4 files referenced by the app
38
+ ```
39
+
40
+
41
+ ## Data Files
42
+
43
+ ### `raw_data.csv`
44
+
45
+ This is the original annotation export. It contains:
46
+
47
+ - `id`
48
+ - `participant`
49
+ - `condition`
50
+ - `run`
51
+ - `filename`
52
+ - `videoLabels`
53
+
54
+ The `videoLabels` column stores Label Studio timeline annotations as JSON.
55
+
56
+ ### `cleaned_data.csv`
57
+
58
+ This is the flattened file the app actually uses. It contains one row per coded label occurrence:
59
+
60
+ - `filename`
61
+ - `video_id`
62
+ - `video_path`
63
+ - `label`
64
+ - `start`
65
+ - `end`
66
+
67
+
68
+ ## Video Playback Rules
69
+
70
+ Segment playback behavior is implemented in `assets/js/video_segment.js`.
71
+
72
+ Each embedded video:
73
+
74
+ - lazy-loads when it gets near the viewport
75
+ - starts from a computed frame-based window
76
+ - loops only within the relevant segment
77
+ - respects manual pause and scrubbing
78
+
79
+ Some labels have custom playback behavior. For example:
80
+
81
+ - `Success` and `Failure` play the full video
82
+ - `Safety Avoidance` and `Redundant Retrieval` play from the annotated start to the end of the video
83
+ - `Missed Grab` and `Slippage` expand the playback window around the annotation
84
+
85
+ Default FPS is set to `30`.
86
+
87
+ ## Typical Update Workflow
88
+
89
+ 1. Replace or update `raw_data.csv`.
90
+ 2. Run `python3 clean.py`.
91
+ 3. Confirm `cleaned_data.csv` was regenerated correctly.
92
+ 4. Add or update files in `videos/` if needed.
93
+ 5. Start a local server and verify `index.html`, `category.html`, and `all_videos.html`.
94
+
95
+ ## Current Dataset Snapshot
96
+
97
+ At the time of this README update, the repository contains:
98
+
99
+ - `288` data rows in `raw_data.csv` excluding the header
100
+ - `657` data rows in `cleaned_data.csv` excluding the header