# xtiwby/metaphor_detector2 architecture

> Architecture diagram of the GitHub repository xtiwby/metaphor_detector2, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-10-04). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/xtiwby/metaphor_detector2
- Repository: https://github.com/xtiwby/metaphor_detector2

## Overview

This repository is a staged multimodal metaphor-detection research pipeline: it prepares Flickr-based inputs to train a Word2Vec semantic disambiguator, applies that model to MET-Meme examples, then trains and evaluates a metaphor classifier. The README also documents a Contrast classifier branch, although its implementation is not shown in the sampled source. Training, evaluation, and experiment artifacts form the main internal layer. Some intermediate file-writing details and the Contrast implementation were not sampled, so those links rely on README statements.

Main entry points: [`run_classifier_train.py`](https://github.com/xtiwby/metaphor_detector2/blob/main/run_classifier_train.py) (Classifier training), [`run_disambiguator_train.py`](https://github.com/xtiwby/metaphor_detector2/blob/main/run_disambiguator_train.py) (Disambiguator training), [`train_eval/trainer.py`](https://github.com/xtiwby/metaphor_detector2/blob/main/train_eval/trainer.py) (Training and evaluation), [`calculate_disambiguator_input.py`](https://github.com/xtiwby/metaphor_detector2/blob/main/calculate_disambiguator_input.py) (Flickr encoding)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_prep["Input preparation"]
  node_flickr_encode["Flickr encoding"]
  node_flickr_data["Flickr datasets<br/>[flickr_dataset.py]"]
end

subgraph group_disamb["Disambiguation training"]
  node_disamb_train["Disambiguator training"]
  node_w2v_model["W2V disambiguator<br/>[clip_w2v_gpt.py]"]
end

subgraph group_classify["Metaphor classification"]
  node_meme_encode["MET-Meme encoding"]
  node_meme_data["Embedded dataset"]
  node_classifier_train["Classifier training"]
  node_gpt_classifier["GPT2 classifier<br/>[clip_w2v_gpt.py]"]
  node_contrast["Contrast branch<br/>[clip_w2v_gpt.py]"]
end

subgraph group_support["Training support"]
  node_trainer["Training and evaluation<br/>[trainer.py]"]
  node_artifacts["Model and metrics<br/>[trainer.py]"]
  node_settings["Seed settings<br/>[setting.py]"]
  node_logger["Progress logging<br/>[logger_handler.py]"]
end

node_researcher(("Researcher"))
node_flickr[("Flickr30k data")]
node_metmeme[("MET-Meme data")]

node_researcher -->|"prepares inputs"| node_flickr_encode
node_flickr -->|"encodes captions"| node_flickr_encode
node_flickr_encode -->|"supplies features"| node_disamb_train
node_flickr -->|"loads examples"| node_flickr_data
node_flickr_data -->|"provides batches"| node_disamb_train
node_researcher -->|"starts training"| node_disamb_train
node_disamb_train -->|"instantiates"| node_w2v_model
node_disamb_train -->|"delegates training"| node_trainer
node_disamb_train -->|"sets seed"| node_settings
node_trainer -->|"saves model"| node_artifacts
node_metmeme -->|"encodes examples"| node_meme_encode
node_w2v_model -->|"disambiguates features"| node_meme_encode
node_meme_encode -->|"writes embedded data"| node_meme_data
node_meme_data -->|"provides batches"| node_classifier_train
node_researcher -->|"starts training"| node_classifier_train
node_classifier_train -->|"instantiates"| node_gpt_classifier
node_classifier_train -->|"delegates evaluation"| node_trainer
node_classifier_train -->|"sets seed"| node_settings
node_classifier_train -->|"configures logging"| node_logger
node_trainer -->|"reports metrics"| node_artifacts
node_contrast -.->|"yields results"| node_artifacts

click node_flickr_encode "https://github.com/xtiwby/metaphor_detector2/blob/main/calculate_disambiguator_input.py"
click node_flickr_data "https://github.com/xtiwby/metaphor_detector2/blob/main/data_processing/flickr_dataset.py"
click node_disamb_train "https://github.com/xtiwby/metaphor_detector2/blob/main/run_disambiguator_train.py"
click node_w2v_model "https://github.com/xtiwby/metaphor_detector2/blob/main/model/clip_w2v_gpt.py"
click node_trainer "https://github.com/xtiwby/metaphor_detector2/blob/main/train_eval/trainer.py"
click node_meme_encode "https://github.com/xtiwby/metaphor_detector2/blob/main/calculate_classifier_input.py"
click node_meme_data "https://github.com/xtiwby/metaphor_detector2/blob/main/data_processing/metaphor_dataset.py"
click node_classifier_train "https://github.com/xtiwby/metaphor_detector2/blob/main/run_classifier_train.py"
click node_gpt_classifier "https://github.com/xtiwby/metaphor_detector2/blob/main/model/clip_w2v_gpt.py"
click node_contrast "https://github.com/xtiwby/metaphor_detector2/blob/main/model/clip_w2v_gpt.py"
click node_artifacts "https://github.com/xtiwby/metaphor_detector2/blob/main/train_eval/trainer.py"
click node_settings "https://github.com/xtiwby/metaphor_detector2/blob/main/utils/setting.py"
click node_logger "https://github.com/xtiwby/metaphor_detector2/blob/main/utils/logger_handler.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_flickr_encode,node_flickr_data toneBlue
class node_disamb_train,node_w2v_model,node_flickr,node_metmeme toneAmber
class node_meme_encode,node_meme_data,node_classifier_train,node_gpt_classifier,node_contrast toneMint
class node_trainer,node_artifacts,node_settings,node_logger toneRose
class node_researcher toneIndigo
```

## Components

### Input preparation

- [Flickr encoding](https://github.com/xtiwby/metaphor_detector2/blob/main/calculate_disambiguator_input.py): `calculate_disambiguator_input.py`
- [Flickr datasets](https://github.com/xtiwby/metaphor_detector2/blob/main/data_processing/flickr_dataset.py): `data_processing/flickr_dataset.py`

### Disambiguation training

- [Disambiguator training](https://github.com/xtiwby/metaphor_detector2/blob/main/run_disambiguator_train.py): `run_disambiguator_train.py`
- [W2V disambiguator](https://github.com/xtiwby/metaphor_detector2/blob/main/model/clip_w2v_gpt.py): `model/clip_w2v_gpt.py`

### Metaphor classification

- [MET-Meme encoding](https://github.com/xtiwby/metaphor_detector2/blob/main/calculate_classifier_input.py): `calculate_classifier_input.py`
- [Embedded dataset](https://github.com/xtiwby/metaphor_detector2/blob/main/data_processing/metaphor_dataset.py): `data_processing/metaphor_dataset.py`
- [Classifier training](https://github.com/xtiwby/metaphor_detector2/blob/main/run_classifier_train.py): `run_classifier_train.py`
- [GPT2 classifier](https://github.com/xtiwby/metaphor_detector2/blob/main/model/clip_w2v_gpt.py): `model/clip_w2v_gpt.py`
- [Contrast branch](https://github.com/xtiwby/metaphor_detector2/blob/main/model/clip_w2v_gpt.py): `model/clip_w2v_gpt.py`

### Training support

- [Training and evaluation](https://github.com/xtiwby/metaphor_detector2/blob/main/train_eval/trainer.py): `train_eval/trainer.py`
- [Model and metrics](https://github.com/xtiwby/metaphor_detector2/blob/main/train_eval/trainer.py): `train_eval/trainer.py`
- [Seed settings](https://github.com/xtiwby/metaphor_detector2/blob/main/utils/setting.py): `utils/setting.py`
- [Progress logging](https://github.com/xtiwby/metaphor_detector2/blob/main/utils/logger_handler.py): `utils/logger_handler.py`

### Other components

- **Researcher**
- **Flickr30k data**
- **MET-Meme data**

## Connections

- Researcher → Flickr encoding: prepares inputs (evidence: `README.md`)
- Flickr30k data → Flickr encoding: encodes captions (evidence: `calculate_disambiguator_input.py`)
- Flickr encoding → Disambiguator training: supplies features (evidence: `README.md`)
- Flickr30k data → Flickr datasets: loads examples (evidence: `data_processing/flickr_dataset.py`)
- Flickr datasets → Disambiguator training: provides batches (evidence: `run_disambiguator_train.py`)
- Researcher → Disambiguator training: starts training (evidence: `README.md`)
- Disambiguator training → W2V disambiguator: instantiates (evidence: `run_disambiguator_train.py`)
- Disambiguator training → Training and evaluation: delegates training (evidence: `run_disambiguator_train.py`)
- Disambiguator training → Seed settings: sets seed (evidence: `run_disambiguator_train.py`)
- Training and evaluation → Model and metrics: saves model (evidence: `train_eval/trainer.py`)
- MET-Meme data → MET-Meme encoding: encodes examples (evidence: `calculate_classifier_input.py`)
- W2V disambiguator → MET-Meme encoding: disambiguates features (evidence: `calculate_classifier_input.py`)
- MET-Meme encoding → Embedded dataset: writes embedded data (evidence: `README.md`)
- Embedded dataset → Classifier training: provides batches (evidence: `run_classifier_train.py`)
- Researcher → Classifier training: starts training (evidence: `README.md`)
- Classifier training → GPT2 classifier: instantiates (evidence: `run_classifier_train.py`)
- Classifier training → Training and evaluation: delegates evaluation (evidence: `run_classifier_train.py`)
- Classifier training → Seed settings: sets seed (evidence: `run_classifier_train.py`)
- Classifier training → Progress logging: configures logging (evidence: `run_classifier_train.py`)
- Training and evaluation → Model and metrics: reports metrics (evidence: `train_eval/trainer.py`)
- Contrast branch → Model and metrics: yields results (evidence: `README.md`)
