bechdelai

KO | EN

gitlite — search

#typescript #ai-agents #deepseek-harness #dsh-plugin #open-source #ai #claude-code #cli #windows #codex #developer-tools #dsh

★ 37 Open GitHub ↗

Automating the Bechdel test and its variants for feminine representation in movies with AI

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# BechdelAI
Measurement and automation of the Bechdel test, female (under)representation and representation inequalities in cinema and audiovisual media

![](https://raw.githubusercontent.com/dataforgoodfr/bechdelai/main/docs/assets/cover_bechdelai.png)

> This repo is under active development, the documentation may not be up to date

## Structure of the repo

```
- bechdelai         ----- Python code as a library
- docs              ----- Documentation
- data              ----- Reusable datasets
- notebooks         ----- Experiments
    - video         ----- Experiments on video data
    - audio         ----- Experiments on audio data
    - posters       ----- Experiments on posters
    - nlp           ----- Experiments on nlp
```

## Documentation

- Official documentation is at https://dataforgood.fr/bechdelai
- Official website for the project is here (French) - https://dataforgood.fr/projects/bechdelai
- Project documentation is [here](https://dataforgood.slite.com/app/docs/~8IRrgyKYR) (French)

## Developers

```
mkdocs serve        ---- launch documentation locally
poetry build        ---- build package locally
poetry publish      ---- publish package on PyPi
```

### Installing pre-commit

To install `pre-commit`, you can use these command lines at the root of this repo:

```bash
pip install pre-commit
pre-commit install
```

Then when you'll use `git commit` it will automatically run the following checks : [BechdelAI pre-commit hooks](pre-commit-hooks.md).

**Important**: if a file is updated by one hook you need to add it again before reusing `git commit`.

More details in the [`pre-commit` documentation](https://pre-commit.com/)

### Golden rules & best practices
- Use type annotations as much as possible (see tutorial [here](https://towardsdatascience.com/type-annotations-in-python-d90990b172dc))
- Use Google style docstrings (you can use Visual Studio Code extension)
- Documentation of all functions and classes is done using mkdocstrings https://mkdocstrings.github.io/usage/

## Get movies from Allociné / TMDB (may be deprecated)

### Install

```bash
pip install -e .
```

### Run with cli

```
$ scrap-movies --help
Usage: scrap-movies [OPTIONS]

CLI for scraping movies from Allociné and TMDB.

It follows these steps:

1. Extract N movies from Allociné

2. Find corresponding TMDB id for each movie

3. Get data for each movie with TMDB API: metadata, crew and cast

4. Save csv results

Options:
  -n, --movies-number TEXT  Number of movies wanted.  [required]
  -p, --path TEXT           Path where to save csv results.  [required]
  --sort-by TEXT            How to sort movies in Allociné pages. Available
                            choices: ['popularity', 'alphabetic',
                            'press_note', 'public_note']
  --genre TEXT              Filter for genre.
  --year TEXT               Filter for year.
  --country TEXT            Filter for country.
  -v, --verbose BOOLEAN     The person to greet.
  --help                    Show this message and exit.
```

**Examples**

```bash
# Scrap first current most popular movie in Allociné
scrap-movies -n 1 -p .
```

```bash
# Scrap top 15 most popular movies in Allociné in 2018 from France
scrap-movies -n 15 -p . --year 2018 --country France
```

```bash
# Scrap top 50 with the best public notes
scrap-movies -n 50 -p . --sort-by public_note
```

bechdelai

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README