# orange-opensource/trustfuse architecture

> Architecture diagram of the GitHub repository orange-opensource/trustfuse, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-11-19). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/orange-opensource/trustfuse
- Repository: https://github.com/orange-opensource/trustfuse

## Overview

TrustFuse is a knowledge-graph data-fusion experimentation playground, accompanied by the WikiConflict dataset. Its main workflow loads conflicting claims, preprocesses them, dispatches a fusion model, restores fused results, and evaluates them; the UI also exposes dataset visualizations. WikiConflict adds a dataset-construction branch based on Wikidata revision history. The sampled excerpts establish the command-line fusion and evaluation calls, but not the UI’s complete runtime wiring or the internal links among dataset-generation stages, so those connections are omitted or marked unverified.

Main entry points: [`settings.py`](https://github.com/orange-opensource/trustfuse/blob/main/settings.py) (Model configuration), [`trustfuse/models/model.py`](https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/models/model.py) (Model input/output), [`fusion_pipeline.py`](https://github.com/orange-opensource/trustfuse/blob/main/fusion_pipeline.py) (Fusion runner), [`app.py`](https://github.com/orange-opensource/trustfuse/blob/main/app.py) (Fusion UI)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_workflow["Fusion workflow"]
  node_pipeline["Fusion runner<br/>[fusion_pipeline.py]"]
  node_dataset["Dataset handling<br/>[dataset.py]"]
  node_preprocess["Metadata preprocessing<br/>[preprocessing.py]"]
  node_modelbase["Model input/output<br/>[model.py]"]
  node_evaluation["Metrics evaluation<br/>[evaluation.py]"]
  node_settings["Model configuration<br/>[settings.py]"]
end

subgraph group_models["Fusion algorithms"]
  node_crh["CRH<br/>[crh.py]"]
  node_catd["CATD<br/>[catd.py]"]
  node_kdem["KDEm<br/>[kdem.py]"]
  node_gtm["GTM<br/>[gtm.py]"]
  node_truthfinder["TruthFinder<br/>[truthfinder.py]"]
  node_ltm["LTM<br/>[ltm.py]"]
  node_slimfast["SLiMFast<br/>[slimfast.py]"]
  node_accu["ACCU<br/>[accu.py]"]
end

subgraph group_ui["User interface"]
  node_webapp["Fusion UI<br/>[app.py]"]
  node_visualization["Graph visualization<br/>[visualization.py]"]
end

subgraph group_wiki["WikiConflict data"]
  node_wiki_generation["Dataset construction"]
  node_wiki_data[("WikiConflict data")]
end

node_operator(("Experimenter"))

node_operator -->|"runs"| node_pipeline
node_pipeline -->|"loads data"| node_dataset
node_pipeline -->|"preprocesses"| node_preprocess
node_pipeline -->|"selects model"| node_settings
node_settings -->|"maps model"| node_crh
node_settings -->|"maps model"| node_catd
node_settings -->|"maps model"| node_kdem
node_settings -->|"maps model"| node_gtm
node_settings -->|"maps model"| node_truthfinder
node_settings -->|"maps model"| node_ltm
node_settings -->|"maps model"| node_slimfast
node_settings -->|"maps model"| node_accu
node_pipeline -.->|"fuses claims"| node_modelbase
node_pipeline -->|"records results"| node_dataset
node_pipeline -->|"computes metrics"| node_evaluation
node_dataset -->|"uses functions"| node_preprocess
node_crh -->|"extends"| node_modelbase
node_catd -->|"extends"| node_modelbase
node_kdem -->|"extends"| node_modelbase
node_gtm -->|"extends"| node_modelbase
node_truthfinder -->|"extends"| node_modelbase
node_ltm -->|"extends"| node_modelbase
node_slimfast -->|"extends"| node_modelbase
node_accu -->|"extends"| node_modelbase
node_operator -.->|"uses"| node_webapp
node_webapp -->|"loads data"| node_dataset
node_webapp -->|"requests metrics"| node_evaluation
node_webapp -->|"renders graphs"| node_visualization
node_wiki_generation -.->|"constructs dataset"| node_wiki_data

click node_pipeline "https://github.com/orange-opensource/trustfuse/blob/main/fusion_pipeline.py"
click node_dataset "https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/conflicting_dataset/dataset.py"
click node_preprocess "https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/conflicting_dataset/preprocessing.py"
click node_modelbase "https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/models/model.py"
click node_evaluation "https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/evaluation/evaluation.py"
click node_settings "https://github.com/orange-opensource/trustfuse/blob/main/settings.py"
click node_crh "https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/models/crh.py"
click node_catd "https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/models/catd.py"
click node_kdem "https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/models/kdem.py"
click node_gtm "https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/models/gtm.py"
click node_truthfinder "https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/models/truthfinder.py"
click node_ltm "https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/models/ltm.py"
click node_slimfast "https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/models/slimfast.py"
click node_accu "https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/models/accu.py"
click node_webapp "https://github.com/orange-opensource/trustfuse/blob/main/app.py"
click node_visualization "https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/visualization/visualization.py"
click node_wiki_generation "https://github.com/orange-opensource/trustfuse/tree/main/dataset_generation"
click node_wiki_data "https://github.com/orange-opensource/trustfuse/tree/main/data/input_trustfuse/wikiconflict/monuments_in_paris"

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_pipeline,node_dataset,node_preprocess,node_modelbase,node_evaluation,node_settings toneBlue
class node_crh,node_catd,node_kdem,node_gtm,node_truthfinder,node_ltm,node_slimfast,node_accu toneAmber
class node_webapp,node_visualization toneMint
class node_wiki_generation,node_wiki_data toneRose
class node_operator toneIndigo
```

## Components

### Fusion workflow

- [Fusion runner](https://github.com/orange-opensource/trustfuse/blob/main/fusion_pipeline.py): `fusion_pipeline.py`
- [Dataset handling](https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/conflicting_dataset/dataset.py): `trustfuse/conflicting_dataset/dataset.py`
- [Metadata preprocessing](https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/conflicting_dataset/preprocessing.py): `trustfuse/conflicting_dataset/preprocessing.py`
- [Model input/output](https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/models/model.py): `trustfuse/models/model.py`
- [Metrics evaluation](https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/evaluation/evaluation.py): `trustfuse/evaluation/evaluation.py`
- [Model configuration](https://github.com/orange-opensource/trustfuse/blob/main/settings.py): `settings.py`

### Fusion algorithms

- [CRH](https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/models/crh.py): `trustfuse/models/crh.py`
- [CATD](https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/models/catd.py): `trustfuse/models/catd.py`
- [KDEm](https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/models/kdem.py): `trustfuse/models/kdem.py`
- [GTM](https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/models/gtm.py): `trustfuse/models/gtm.py`
- [TruthFinder](https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/models/truthfinder.py): `trustfuse/models/truthfinder.py`
- [LTM](https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/models/ltm.py): `trustfuse/models/ltm.py`
- [SLiMFast](https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/models/slimfast.py): `trustfuse/models/slimfast.py`
- [ACCU](https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/models/accu.py): `trustfuse/models/accu.py`

### User interface

- [Fusion UI](https://github.com/orange-opensource/trustfuse/blob/main/app.py): `app.py`
- [Graph visualization](https://github.com/orange-opensource/trustfuse/blob/main/trustfuse/visualization/visualization.py): `trustfuse/visualization/visualization.py`

### WikiConflict data

- [Dataset construction](https://github.com/orange-opensource/trustfuse/tree/main/dataset_generation): `dataset_generation`
- [WikiConflict data](https://github.com/orange-opensource/trustfuse/tree/main/data/input_trustfuse/wikiconflict/monuments_in_paris): `data/input_trustfuse/wikiconflict/monuments_in_paris`

### Other components

- **Experimenter**

## Connections

- Experimenter → Fusion runner: runs (evidence: `fusion_pipeline.py`)
- Fusion runner → Dataset handling: loads data (evidence: `fusion_pipeline.py`)
- Fusion runner → Metadata preprocessing: preprocesses (evidence: `fusion_pipeline.py`)
- Fusion runner → Model configuration: selects model (evidence: `fusion_pipeline.py`)
- Model configuration → CRH: maps model (evidence: `settings.py`)
- Model configuration → CATD: maps model (evidence: `settings.py`)
- Model configuration → KDEm: maps model (evidence: `settings.py`)
- Model configuration → GTM: maps model (evidence: `settings.py`)
- Model configuration → TruthFinder: maps model (evidence: `settings.py`)
- Model configuration → LTM: maps model (evidence: `settings.py`)
- Model configuration → SLiMFast: maps model (evidence: `settings.py`)
- Model configuration → ACCU: maps model (evidence: `settings.py`)
- Fusion runner → Model input/output: fuses claims
- Fusion runner → Dataset handling: records results (evidence: `fusion_pipeline.py`)
- Fusion runner → Metrics evaluation: computes metrics (evidence: `fusion_pipeline.py`)
- Dataset handling → Metadata preprocessing: uses functions (evidence: `trustfuse/conflicting_dataset/dataset.py`)
- CRH → Model input/output: extends (evidence: `trustfuse/models/crh.py`)
- CATD → Model input/output: extends (evidence: `trustfuse/models/catd.py`)
- KDEm → Model input/output: extends (evidence: `trustfuse/models/kdem.py`)
- GTM → Model input/output: extends (evidence: `trustfuse/models/gtm.py`)
- TruthFinder → Model input/output: extends (evidence: `trustfuse/models/truthfinder.py`)
- LTM → Model input/output: extends (evidence: `trustfuse/models/ltm.py`)
- SLiMFast → Model input/output: extends (evidence: `trustfuse/models/slimfast.py`)
- ACCU → Model input/output: extends (evidence: `trustfuse/models/accu.py`)
- Experimenter → Fusion UI: uses
- Fusion UI → Dataset handling: loads data (evidence: `app.py`)
- Fusion UI → Metrics evaluation: requests metrics (evidence: `app.py`)
- Fusion UI → Graph visualization: renders graphs (evidence: `app.py`)
- Dataset construction → WikiConflict data: constructs dataset (evidence: `README.md`)
