# nkondapa/visualknowledgetracing architecture

> Architecture diagram of the GitHub repository nkondapa/visualknowledgetracing, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-11-16). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/nkondapa/visualknowledgetracing
- Repository: https://github.com/nkondapa/visualknowledgetracing

## Overview

Visual Knowledge Tracing recreates experiments on human responses to image sequences. The documented workflow prepares dataset partitions, runs experiment scripts that train tracing models, computes probability or precision–recall analyses, and generates figures. The graph preserves the model, feature-extraction, and analysis capabilities named in the repository; detailed links among unsampled runners and models are left unverified. Relationships reflect README statements or shown source/index references.

Main entry points: [`tracing_models/lstm/fe_builder.py`](https://github.com/nkondapa/visualknowledgetracing/blob/main/tracing_models/lstm/fe_builder.py) (Feature selection), [`experiments/vkt_experiment/analysis_scripts/experiment1_average_prob.py`](https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/analysis_scripts/experiment1_average_prob.py) (Experiment analysis), [`experiments/vkt_experiment/experiment1_script.py`](https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/experiment1_script.py) (Experiment scripts), [`experiments/vkt_experiment/runner_classifier_model1v2_wpca.py`](https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/runner_classifier_model1v2_wpca.py) (Model runners)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_data["Dataset preparation"]
  node_prep["Dataset preparation"]
end

subgraph group_models["Experiments and models"]
  node_experiment["Experiment scripts"]
  node_supplemental["Supplemental experiments"]
  node_runners["Model runners"]
  node_lstm["LSTM tracers"]
  node_baselines["Tracing baselines<br/>[exemplar.py]"]
  node_dkt["DKT translation"]
  node_transformer["Transformer tracer"]
  node_template["Model template<br/>[template.py]"]
  node_featurebuilder["Feature selection<br/>[fe_builder.py]"]
  node_cnn128["CNN 128<br/>[cnn128_backbone.py]"]
  node_cnn144["CNN 144<br/>[cnn144_backbone.py]"]
  node_cnn144gray["CNN 144 1ch"]
  node_resnet["ResNet extractor<br/>[resnet.py]"]
  node_config["Runtime configuration<br/>[config.py]"]
  node_experimentutil["Experiment utilities"]
end

subgraph group_analysis["Analysis"]
  node_analysisrun["Experiment analysis"]
  node_classprob["Class probabilities"]
  node_probutil["Probability utilities"]
  node_prutil["Precision recall"]
  node_trainingstats["Training statistics"]
end

subgraph group_outputs["Figures and utilities"]
  node_figures["Figure scripts"]
  node_hyperplanes["Hyperplane utilities<br/>[hyperplanes.py]"]
end

node_researcher(("Researcher"))
node_rawdata["VKT datasets"]

node_researcher -->|"prepares data"| node_prep
node_rawdata -->|"provides input"| node_prep
node_researcher -->|"runs experiment"| node_experiment
node_experiment -.->|"trains models"| node_runners
node_experiment -->|"runs variants"| node_supplemental
node_runners -.->|"selects tracer"| node_lstm
node_lstm -->|"selects features"| node_featurebuilder
node_featurebuilder -->|"constructs"| node_cnn128
node_featurebuilder -->|"constructs"| node_cnn144
node_featurebuilder -->|"constructs"| node_cnn144gray
node_featurebuilder -->|"constructs"| node_resnet
node_researcher -->|"runs analysis"| node_analysisrun
node_analysisrun -->|"loads data"| node_experimentutil
node_analysisrun -->|"computes probabilities"| node_classprob
node_classprob -->|"generates labels"| node_probutil
node_analysisrun -->|"plots training"| node_trainingstats
node_analysisrun -->|"computes metrics"| node_prutil
node_researcher -->|"generates figures"| node_figures
node_figures -->|"formats results"| node_prutil
node_experimentutil -->|"reads paths"| node_config

click node_prep "https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/prep_butterflies_data.py"
click node_experiment "https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/experiment1_script.py"
click node_supplemental "https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/experiment2_script.py"
click node_runners "https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/runner_classifier_model1v2_wpca.py"
click node_lstm "https://github.com/nkondapa/visualknowledgetracing/blob/main/tracing_models/lstm/classifier_model1v2_wpca.py"
click node_baselines "https://github.com/nkondapa/visualknowledgetracing/blob/main/tracing_models/baseline/exemplar.py"
click node_dkt "https://github.com/nkondapa/visualknowledgetracing/blob/main/tracing_models/dkt_translation/dkt_translation_model.py"
click node_transformer "https://github.com/nkondapa/visualknowledgetracing/blob/main/tracing_models/transformer/direct_response_transformer.py"
click node_template "https://github.com/nkondapa/visualknowledgetracing/blob/main/tracing_models/template/template.py"
click node_featurebuilder "https://github.com/nkondapa/visualknowledgetracing/blob/main/tracing_models/lstm/fe_builder.py"
click node_cnn128 "https://github.com/nkondapa/visualknowledgetracing/blob/main/feature_extractors/cnn128_backbone.py"
click node_cnn144 "https://github.com/nkondapa/visualknowledgetracing/blob/main/feature_extractors/cnn144_backbone.py"
click node_cnn144gray "https://github.com/nkondapa/visualknowledgetracing/blob/main/feature_extractors/cnn144_1ch_backbone.py"
click node_resnet "https://github.com/nkondapa/visualknowledgetracing/blob/main/feature_extractors/resnet.py"
click node_config "https://github.com/nkondapa/visualknowledgetracing/blob/main/config.py"
click node_experimentutil "https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/experiment_utilities.py"
click node_analysisrun "https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/analysis_scripts/experiment1_average_prob.py"
click node_classprob "https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/analysis_utilities/class_conditioned_average_probability_functions.py"
click node_probutil "https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/analysis_utilities/probability_utilities.py"
click node_prutil "https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/analysis_utilities/precision_recall_utilities.py"
click node_trainingstats "https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/analysis_utilities/training_statistics.py"
click node_figures "https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/figures_scripts/experiment1_average_precision_table_main.py"
click node_hyperplanes "https://github.com/nkondapa/visualknowledgetracing/blob/main/utilities/hyperplanes.py"

classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_prep toneBlue
class node_experiment,node_supplemental,node_runners,node_lstm,node_baselines,node_dkt,node_transformer,node_template,node_featurebuilder,node_cnn128,node_cnn144,node_cnn144gray,node_resnet,node_config,node_experimentutil toneAmber
class node_analysisrun,node_classprob,node_probutil,node_prutil,node_trainingstats toneMint
class node_figures,node_hyperplanes toneRose
class node_researcher,node_rawdata toneIndigo
```

## Components

### Dataset preparation

- [Dataset preparation](https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/prep_butterflies_data.py): `experiments/vkt_experiment/prep_butterflies_data.py`

### Experiments and models

- [Experiment scripts](https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/experiment1_script.py): `experiments/vkt_experiment/experiment1_script.py`
- [Supplemental experiments](https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/experiment2_script.py): `experiments/vkt_experiment/experiment2_script.py`
- [Model runners](https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/runner_classifier_model1v2_wpca.py): `experiments/vkt_experiment/runner_classifier_model1v2_wpca.py`
- [LSTM tracers](https://github.com/nkondapa/visualknowledgetracing/blob/main/tracing_models/lstm/classifier_model1v2_wpca.py): `tracing_models/lstm/classifier_model1v2_wpca.py`
- [Tracing baselines](https://github.com/nkondapa/visualknowledgetracing/blob/main/tracing_models/baseline/exemplar.py): `tracing_models/baseline/exemplar.py`
- [DKT translation](https://github.com/nkondapa/visualknowledgetracing/blob/main/tracing_models/dkt_translation/dkt_translation_model.py): `tracing_models/dkt_translation/dkt_translation_model.py`
- [Transformer tracer](https://github.com/nkondapa/visualknowledgetracing/blob/main/tracing_models/transformer/direct_response_transformer.py): `tracing_models/transformer/direct_response_transformer.py`
- [Model template](https://github.com/nkondapa/visualknowledgetracing/blob/main/tracing_models/template/template.py): `tracing_models/template/template.py`
- [Feature selection](https://github.com/nkondapa/visualknowledgetracing/blob/main/tracing_models/lstm/fe_builder.py): `tracing_models/lstm/fe_builder.py`
- [CNN 128](https://github.com/nkondapa/visualknowledgetracing/blob/main/feature_extractors/cnn128_backbone.py): `feature_extractors/cnn128_backbone.py`
- [CNN 144](https://github.com/nkondapa/visualknowledgetracing/blob/main/feature_extractors/cnn144_backbone.py): `feature_extractors/cnn144_backbone.py`
- [CNN 144 1ch](https://github.com/nkondapa/visualknowledgetracing/blob/main/feature_extractors/cnn144_1ch_backbone.py): `feature_extractors/cnn144_1ch_backbone.py`
- [ResNet extractor](https://github.com/nkondapa/visualknowledgetracing/blob/main/feature_extractors/resnet.py): `feature_extractors/resnet.py`
- [Runtime configuration](https://github.com/nkondapa/visualknowledgetracing/blob/main/config.py): `config.py`
- [Experiment utilities](https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/experiment_utilities.py): `experiments/vkt_experiment/experiment_utilities.py`

### Analysis

- [Experiment analysis](https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/analysis_scripts/experiment1_average_prob.py): `experiments/vkt_experiment/analysis_scripts/experiment1_average_prob.py`
- [Class probabilities](https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/analysis_utilities/class_conditioned_average_probability_functions.py): `experiments/vkt_experiment/analysis_utilities/class_conditioned_average_probability_functions.py`
- [Probability utilities](https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/analysis_utilities/probability_utilities.py): `experiments/vkt_experiment/analysis_utilities/probability_utilities.py`
- [Precision recall](https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/analysis_utilities/precision_recall_utilities.py): `experiments/vkt_experiment/analysis_utilities/precision_recall_utilities.py`
- [Training statistics](https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/analysis_utilities/training_statistics.py): `experiments/vkt_experiment/analysis_utilities/training_statistics.py`

### Figures and utilities

- [Figure scripts](https://github.com/nkondapa/visualknowledgetracing/blob/main/experiments/vkt_experiment/figures_scripts/experiment1_average_precision_table_main.py): `experiments/vkt_experiment/figures_scripts/experiment1_average_precision_table_main.py`
- [Hyperplane utilities](https://github.com/nkondapa/visualknowledgetracing/blob/main/utilities/hyperplanes.py): `utilities/hyperplanes.py`

### Other components

- **Researcher**
- **VKT datasets**

## Connections

- Researcher → Dataset preparation: prepares data
- VKT datasets → Dataset preparation: provides input
- Researcher → Experiment scripts: runs experiment
- Experiment scripts → Model runners: trains models
- Experiment scripts → Supplemental experiments: runs variants
- Model runners → LSTM tracers: selects tracer
- LSTM tracers → Feature selection: selects features (evidence: `tracing_models/lstm/classifier_model1v2_wpca.py`)
- Feature selection → CNN 128: constructs (evidence: `tracing_models/lstm/fe_builder.py`)
- Feature selection → CNN 144: constructs (evidence: `tracing_models/lstm/fe_builder.py`)
- Feature selection → CNN 144 1ch: constructs (evidence: `tracing_models/lstm/fe_builder.py`)
- Feature selection → ResNet extractor: constructs (evidence: `tracing_models/lstm/fe_builder.py`)
- Researcher → Experiment analysis: runs analysis
- Experiment analysis → Experiment utilities: loads data (evidence: `experiments/vkt_experiment/analysis_scripts/experiment1_average_prob.py`)
- Experiment analysis → Class probabilities: computes probabilities (evidence: `experiments/vkt_experiment/analysis_scripts/experiment1_average_prob.py`)
- Class probabilities → Probability utilities: generates labels (evidence: `experiments/vkt_experiment/analysis_utilities/class_conditioned_average_probability_functions.py`)
- Experiment analysis → Training statistics: plots training (evidence: `experiments/vkt_experiment/analysis_scripts/experiment1_average_prob.py`)
- Experiment analysis → Precision recall: computes metrics (evidence: `experiments/vkt_experiment/analysis_scripts/experiment1_average_prob.py`)
- Researcher → Figure scripts: generates figures
- Figure scripts → Precision recall: formats results
- Experiment utilities → Runtime configuration: reads paths (evidence: `experiments/vkt_experiment/experiment_utilities.py`)
