How krabberod/dagens-uniform works
Dagens uniform is a research workflow that turns outfit photographs into a scored specimen dataset, measures garment similarity, tests whether photo partners match, and presents the findings. The graph follows the documented pipeline while distinguishing automated scoring, human review, analysis, and figure generation. Some analysis and browser-tool implementation is not in the supplied tree; those subsystems are represented only at the documented directory level, and uncertain links are omitted.
Main entry points: make_demo_data.py Scored specimens · analysis Similarity distances · 27_classify_and_sample_colors_with_local_vlm.py Garment classification · 01_build_photo_inventory.py Photo inventory
Components 21
Ingest and scoring
- Photographs
- Synthetic data generator
demo/make_demo_data.py - Photo inventory
02_scripts/ingest/01_build_photo_inventory.py - Specimen template
02_scripts/ingest/03_build_specimen_template.py - Garment classification
02_scripts/classify/27_classify_and_sample_colors_with_local_vlm.py - Scored specimens
demo/make_demo_data.py
Review and colour
- Scoring corrections
02_scripts/classify/50_hand_corrections.py - QA and consensus
02_scripts/review - Palette refinement
02_scripts/ingest/47_reduce_categories.py - Colour science
02_scripts/colourlib.py
Analysis and tests
- Similarity distances
02_scripts/analysis - Ordination
02_scripts/analysis - Permutation and seasonal tests
02_scripts/analysis - Outfit accumulation
02_scripts/analysis
Figures and outputs
- Analysis figures
02_scripts/analysis - Photo colour grid
02_scripts/analysis/18_photo_color_grid.py - Person outfit stacks
02_scripts/analysis/48_person_outfit_stacks.py - Interactive 3D scene
02_scripts/analysis/53_export_3d_scene.py - Tables and figures
Other components
- Researcher
- Local VLM
Connections 31
- Researcher → Photographs: provides
- Researcher → Synthetic data generator: runs
- Synthetic data generator → Scored specimens: generates
- Photographs → Photo inventory: inventories
- Photo inventory → Specimen template: seeds
- Specimen template → Garment classification: scores
- Garment classification → Local VLM: queries
- Garment classification → Scored specimens: writes scores
- Researcher → Scoring corrections: corrects
- Scoring corrections → Scored specimens: updates
- Researcher → QA and consensus: reviews
- QA and consensus → Scored specimens: checks
- Palette refinement → Colour science: uses colour math
- Researcher → Palette refinement: refines
- Scored specimens → Similarity distances: scores characters
- Colour science → Similarity distances: measures colour
- Similarity distances → Ordination: ordains
- Similarity distances → Permutation and seasonal tests: tests similarity
- Scored specimens → Permutation and seasonal tests: supplies metadata
- Scored specimens → Outfit accumulation: counts combinations
- Ordination → Analysis figures: provides coordinates
- Scored specimens → Analysis figures: provides specimen data
- Scored specimens → Photo colour grid: reads specimen data
- Scored specimens → Person outfit stacks: reads appearances
- Ordination → Interactive 3D scene: provides coordinates
- Interactive 3D scene → Tables and figures: exports scene
- Analysis figures → Tables and figures: writes figures
- Photo colour grid → Tables and figures: writes grid
- Person outfit stacks → Tables and figures: writes stacks
- Permutation and seasonal tests → Tables and figures: writes results
- Outfit accumulation → Tables and figures: writes estimates