Ensure pipeline description does not omit anything from main full pipeline doc (#110)

* Incorporate image mistakenly ignored

* Clean up pipeline docs against reference single document

* Clarify section names under pipeline tab

* Fix outdated CVAT reference links

* Correct ethogram link

---------

Co-authored-by: Elizabeth Campolongo <38985481+egrace479@users.noreply.github.com>

Author: thompsonmj

.gitignore

@@ -199,8 +199,8 @@ _hugging_face
 _hugging_face
 _hugging_face/*
 
-# Allow CSV files intended to be in docs, ignore one not to include
-!docs/**/*.csv
+# Allow tracking of intended files in docs, ignore one not to include
+!docs/**
 docs/methodology_comparison/kabr_dataset/data/consolidated_metadata.csv
 
 helper_scripts/mini-scenes

docs/images/im_mini-scenes.jpg

@@ -15,16 +15,16 @@ Time budget analysis shows how animals allocate their time across different beha
   <img src="../images/grevys.png" alt="grevys" style="width: 48%;">
 </div>
 
-**Figure:** Example flight path and video clip from KABR dataset: 2 male Grevy's zebras observed for 10 minutes on 01/18/23.
+**Figure 1:** Example flight path and video clip from KABR dataset: two male Grevy's zebras observed for 10 minutes on 01/18/23.
 
 ![Time Budget](../images/timebudget.png)
 
-**Figure:** Overall time budget for duration of 10 minute observation
+**Figure 2:** Overall time budget for duration of 10 minute observation.
 
 ![Timeline 1](../images/timeline0.png)
 ![Timeline 2](../images/timeline1.png)
 
-**Figure:** Gantt chart for each zebra (3 minute duration)
+**Figure 3:** Gantt chart for each zebra (3 minute duration).
 
 ### Implementation
 
@@ -50,29 +50,29 @@ See the [behavior transitions notebook](../case_studies/2_zebra_transition/behav
 
 Explore real-world applications in the case studies directory:
 
-- **[Grevy's Landscape Analysis](../case_studies/1_grevys_landscape/grevys_landscape.ipynb)** - Landscape-scale behavioral analysis
-- **[Mixed Species Social](../case_studies/3_mixed_species_social/mixed_species_overlap.ipynb)** - Multi-species interaction analysis  
-- **[Zebra Behavior Transitions](../case_studies/2_zebra_transition/behaviortransitionsheatmap.ipynb)** - Behavioral transition patterns
+- **[Grevy's Landscape Analysis](../case_studies/1_grevys_landscape/grevys_landscape.ipynb)** - Landscape-scale behavioral analysis.
+- **[Mixed Species Social](../case_studies/3_mixed_species_social/mixed_species_overlap.ipynb)** - Multi-species interaction analysis.
+- **[Zebra Behavior Transitions](../case_studies/2_zebra_transition/behaviortransitionsheatmap.ipynb)** - Behavioral transition patterns.
 
 ## Key Metrics Generated
 
 The analysis pipeline can generate several key ecological metrics:
 
-1. **Time Budgets** - Proportion of time spent in each behavioral category
-2. **Behavioral Transitions** - Probability matrices of behavior changes
-3. **Social Interactions** - Proximity and interaction frequency between individuals
-4. **Habitat Associations** - Relationship between behaviors and spatial locations
-5. **Group Composition Dynamics** - Changes in group structure over time
+1. **Time Budgets** - Proportion of time spent in each behavioral category.
+2. **Behavioral Transitions** - Probability matrices of behavior changes.
+3. **Social Interactions** - Proximity and interaction frequency between individuals.
+4. **Habitat Associations** - Relationship between behaviors and spatial locations.
+5. **Group Composition Dynamics** - Changes in group structure over time.
 
 ## Visualization Tools
 
 The framework includes tools for creating publication-ready visualizations:
 
-- Gantt charts for individual behavioral timelines
-- Heat maps for behavioral transition probabilities
-- Spatial plots for habitat use patterns
-- Social network diagrams for interaction patterns
-- Time series plots for behavioral trends
+- Gantt charts for individual behavioral timelines.
+- Heat maps for behavioral transition probabilities.
+- Spatial plots for habitat use patterns.
+- Social network diagrams for interaction patterns.
+- Time series plots for behavioral trends.
 
 ## Next Steps
 

docs/pipeline.md

@@ -45,9 +45,9 @@ miniscene2behavior --config config.yml --checkpoint checkpoint_epoch_00075.pyth
 
 ## Resources
 
-- [Pre-trained KABR model](https://huggingface.co/imageomics/x3d-kabr-kinetics) on Hugging Face
-- [Ethogram definitions](https://github.com/Imageomics/kabr-tools/tree/main/CVAT/ethogram) - Behavior classification system used for zebra videos
-- [Example annotated outputs](https://huggingface.co/imageomics/x3d-kabr-kinetics/tree/main/data/mini_scene_behavior_annotations) on Hugging Face
+- [Pre-trained KABR model](https://huggingface.co/imageomics/x3d-kabr-kinetics) on Hugging Face.
+- [Ethogram definitions](https://github.com/Imageomics/kabr-tools/tree/main/ethogram) - Behavior classification system used for zebra videos.
+- [Example annotated outputs](https://huggingface.co/imageomics/x3d-kabr-kinetics/tree/main/data/mini_scene_behavior_annotations) on Hugging Face.
 
 ## Tool Reference
 
@@ -58,12 +58,12 @@ Source: [src/kabr_tools/miniscene2behavior.py](https://github.com/Imageomics/kab
 Apply machine learning models to classify animal behaviors from mini-scene videos.
 
 **Parameters:**
-- `--hub`: Hugging Face hub repository containing model files
-- `--config`: Path to configuration file (local or from hub)
-- `--checkpoint`: Path to model checkpoint file
-- `--gpu_num`: Number of GPUs to use (0 for CPU)
-- `--miniscene`: Path to mini-scene videos directory
-- `--output`: Path for output CSV file (optional)
+- `--hub`: Hugging Face hub repository containing model files.
+- `--config`: Path to configuration file (local or from hub).
+- `--checkpoint`: Path to model checkpoint file.
+- `--gpu_num`: Number of GPUs to use (0 for CPU).
+- `--miniscene`: Path to mini-scene videos directory.
+- `--output`: Path for output CSV file (optional).
 
 ## Next Steps
 

docs/pipeline/analysis.md

@@ -12,28 +12,28 @@ The drone videos for the [KABR dataset](https://huggingface.co/datasets/imageomi
 
 We collaborated with expert ecologists to ensure minimal disturbance to the animals. Our approach included:
 
-- **Launch Position**: Approximately 200 meters horizontally from the animals
-- **Initial Altitude**: 30-40 meters above ground
-- **Approach Strategy**: Gradually approach the herd from the side by reducing altitude and horizontal distance
-- **Monitoring**: Continuously monitor animals for signs of vigilance
+- **Launch Position**: Approximately 200 meters horizontally from the animals.
+- **Initial Altitude**: 30-40 meters above ground.
+- **Approach Strategy**: Gradually approach the herd from the side by reducing altitude and horizontal distance.
+- **Monitoring**: Continuously monitor animals for signs of vigilance.
 
 !!! warning "Species-Specific Considerations"
     The vigilance exhibited by wildlife varies widely by species, habitat, sex, and the level to which animals may be habituated to anthropogenic noise. We recommend that you tailor your approach to your particular species and setting.
 
 ## Equipment Used
 
-- **Drone**: DJI Air 2S
-- **Flight Mode**: Manual control
-- **Recording Settings**: Standard video recording at highest available resolution
+- **Drone**: DJI Air 2S.
+- **Flight Mode**: Manual control.
+- **Recording Settings**: Standard video recording at highest available resolution.
 
 ## Data Collection Workflow
 
-1. **Site Assessment**: Scout the area and identify animal locations
-2. **Equipment Setup**: Prepare drone and check battery levels
-3. **Launch**: Deploy drone at safe distance from animals  
-4. **Approach**: Gradually move closer while monitoring animal behavior
-5. **Recording**: Capture video footage of natural behaviors
-6. **Data Management**: Safely store and backup all collected footage
+1. **Site Assessment**: Scout the area and identify animal locations.
+2. **Equipment Setup**: Prepare drone and check battery levels.
+3. **Launch**: Deploy drone at safe distance from animals.
+4. **Approach**: Gradually move closer while monitoring animal behavior.
+5. **Recording**: Capture video footage of natural behaviors.
+6. **Data Management**: Safely store and backup all collected footage.
 
 ## Publications
 

docs/pipeline/behavior-labeling.md

@@ -51,56 +51,56 @@ Several utility scripts are available in the `helper_scripts` directory:
 
 Scripts for batch processing and annotation of mini-scenes:
 
-- **[launch_job.py](../../helper_scripts/annotate_mini_scenes/launch_job.py)** - Job launcher for batch processing
-- **[run.sh](../../helper_scripts/annotate_mini_scenes/run.sh)** - Shell script for running annotation tasks
+- **[launch_job.py](../../helper_scripts/annotate_mini_scenes/launch_job.py)** - Job launcher for batch processing.
+- **[run.sh](../../helper_scripts/annotate_mini_scenes/run.sh)** - Shell script for running annotation tasks.
 
 See [README](../../helper_scripts/annotate_mini_scenes/README.md) for detailed usage instructions.
 
 ### Video Processing Utilities
 
-- **[downgrade.sh](../../helper_scripts/downgrade.sh)** - Reduce video file sizes for CVAT compatibility
-- **[rename.sh](../../helper_scripts/rename.sh)** - Batch rename files with consistent naming conventions
+- **[downgrade.sh](../../helper_scripts/downgrade.sh)** - Reduce video file sizes for CVAT compatibility.
+- **[rename.sh](../../helper_scripts/rename.sh)** - Batch rename files with consistent naming conventions.
 
 ## Advanced Configuration
 
 ### Custom Model Training
 
 For training custom behavior recognition models, you'll need:
 
-1. **Annotated mini-scenes** from your specific species/environment
-2. **Behavior ethogram** defining your classification categories
-3. **Training configuration** adapted to your dataset size and complexity
+1. **Annotated mini-scenes** from your specific species/environment.
+2. **Behavior ethogram** defining your classification categories.
+3. **Training configuration** adapted to your dataset size and complexity.
 
 ### GPU Requirements
 
 Most tools support both CPU and GPU processing:
 
-- **CPU mode**: Slower but works on any system
-- **GPU mode**: Significantly faster, recommended for large datasets
-- **Multi-GPU**: Supported for distributed processing of large video collections
+- **CPU mode**: Slower but works on any system.
+- **GPU mode**: Significantly faster, recommended for large datasets.
+- **Multi-GPU**: Supported for distributed processing of large video collections.
 
 ### Batch Processing
 
 For processing large video collections:
 
-1. Use the provided shell scripts for batch operations
-2. Consider using job scheduling systems (SLURM, PBS) for cluster environments
-3. Monitor disk space requirements for intermediate files
+1. Use the provided shell scripts for batch operations.
+2. Consider using job scheduling systems (SLURM, PBS) for cluster environments.
+3. Monitor disk space requirements for intermediate files.
 
 ## Customization Options
 
 The modular design allows for customization at multiple levels:
 
-- **Detection models**: Swap YOLO versions or use custom trained models
-- **Tracking algorithms**: Modify tracking parameters for different scenarios
-- **Behavior models**: Train species-specific or behavior-specific classifiers
-- **Analysis pipelines**: Adapt visualization and metrics to research questions
+- **Detection models**: Swap YOLO versions or use custom trained models.
+- **Tracking algorithms**: Modify tracking parameters for different scenarios.
+- **Behavior models**: Train species-specific or behavior-specific classifiers.
+- **Analysis pipelines**: Adapt visualization and metrics to research questions.
 
 ## Performance Optimization
 
 Tips for optimizing performance:
 
-- Use appropriate video resolution for your analysis needs
-- Batch process multiple videos simultaneously when possible
-- Utilize GPU acceleration for model inference
-- Consider video compression for storage efficiency
+- Use appropriate video resolution for your analysis needs.
+- Batch process multiple videos simultaneously when possible.
+- Utilize GPU acceleration for model inference.
+- Consider video compression for storage efficiency.

docs/pipeline/data-collection.md

@@ -22,10 +22,10 @@ Generation of ecological insights including time budgets, behavioral transitions
 
 ## Key Features
 
-- **Modular Design**: Each step can be adapted and customized for different species and study systems
-- **Multiple Detection Options**: Support for both manual annotation and automated YOLO-based detection
-- **Machine Learning Integration**: Pre-trained models available with support for custom model training
-- **Comprehensive Analytics**: Generate multiple types of ecological metrics and visualizations
+- **Modular Design**: Each step can be adapted and customized for different species and study systems.
+- **Multiple Detection Options**: Support for both manual annotation and automated YOLO-based detection.
+- **Machine Learning Integration**: Pre-trained models available with support for custom model training.
+- **Comprehensive Analytics**: Generate multiple types of ecological metrics and visualizations.
 
 ## Getting Started