Trying things with music lyrics.

You need to set a python environment (tested version: 3.10.5). Then, in your shell, type:

git clone [email protected]:MiguelLaura/lyrics_cloud.git
cd lyrics_cloud
pip install .

To install the Git hooks:

pre-commit install

Before committing, black and generate_readme.py will automatically run.

To change the README file, change README.template.md first and generate the README after (or let the pre-commit do it). You'll need to add the new functions to generate_readme.py in DOCS.

usage: python -m lyrics_cloud.artists_lyrics [-h] [--method [list,file]] --artists ARTISTS --output-file OUTPUT_FILE --token TOKEN

options:
  -h, --help            show this help message and exit
  --method [list,file]
  --artists ARTISTS     list of artists or text file with one artist per line
  --output-file OUTPUT_FILE
                        csv file to write the results
  --token TOKEN         token for the genius API (https://docs.genius.com/)

usage: python -m lyrics_cloud.content_based_recommender [-h] --artist ARTIST --csv-file CSV_FILE --title TITLE

options:
  -h, --help           show this help message and exit
  --artist ARTIST      artist singing the song to get recommendations from
  --csv-file CSV_FILE  csv file containing the lyrics
  --title TITLE        song title to get recommendations from

usage: python -m lyrics_cloud.word_cloud [-h] --artist ARTIST --csv-file CSV_FILE --title TITLE

options:
  -h, --help           show this help message and exit
  --artist ARTIST      artist singing the song to get recommendations from
  --csv-file CSV_FILE  csv file containing the lyrics
  --title TITLE        song title to get recommendations from

• Scraper
  • get_artists_from_file
  • get_lyrics
• Recommender
  • build_recommender_word2vec
  • create_avg_word2vec_embeddings
  • recommend_with_word2vec
  • visualize_embeddings
• Word cloud
  • generate_word_cloud
• Utils
  • clean_lyrics
  • prepare_text

Function to get artists from a txt file.

Arguments

• file str - name of the input file (txt format).

Yields

str - artist's name.

Function to write lyrics from a list of artists into a csv file.

Arguments

• artists list[str] - list of artists name.
• output_file str - name of the output file (csv format).
• token str - token for the genius API (https://docs.genius.com/).

Function to load and train the word2vec model on the corpus.

Arguments

• corpus list[list[str] - corpus of songs.

Returns

model - trained word2vec model.

Function to create the averaged word2vec embeddings.

Arguments

• df_lyrics Series - lyrics.
• model - trained word2vec model.

Returns

list[float] - lyrics averaged word2vec embeddings.

Function to recommend a song based on a title and artist.

Arguments

• idx int - index of the item to get recommendations from.
• embeddings list[float] - averaged word2vec embeddings.
• nb_reco int - number of recommendations to return.

Function to visualize the word2vec embeddings.

Arguments

• w2v - trained word2vec model.

Returns

dataframe - PCA dataframe.

Function to plot a text word cloud.

Arguments

• text str - text to output in the word cloud.

Function to clean the lyrics.

Arguments

• lyrics str - lyrics of a song.

Returns

str - cleaned lyrics.

Function to prepare the text (remove stop words, punctuation, etc.).

Arguments

• text str - text to prepare.

Returns

str - cleaned and tokenized text.