Web Scraping Lyrics

This repository contains a series of functions that can be used to scrape AZLyrics.com and download lyrics in txt format.

Lyrics can be used for many Natural Language Processing tasks: as an example, you can read on my website the report of my Sentiment Analysis of Bruce Springsteen's songs.

Functions

• save_file(path, text, replace=False)
  This is an auxiliary function used to save a given text in a txt file.

• get_lyrics(song_url, save=True, by_decade=False, replace=False, folder="songs")
  This function can be used to download the lyrics of a single song, given its page url.
  Function Parameters:

  • song_url: The AZLyrics url of the page containing the song.
  • save: if True (default), the lyrics are saved in a txt file named as the song title. If False, the function just returns the song title, lyrics and year as a 3-dimensional tuple.
  • by_decade: if True, and save=True, the lyrics are saved in a folder named as the decade when the song was produced. If False (default), and save=True, the lyrics are just stored in a generic folder.
  • folder: the name of the folder where the txt lyrics will be saved.

• scrape_artist(az_url, sleep="random", by_decade=True, replace=False, folder="songs")
  This function downloads all the lyrics of a given artist, starting from their page url.
  Function Parameters:

  • az_url: The artist main page url on AZLyrics.
  • sleep: The sleeping time (in seconds) between iterations. By default it is set to "random", which means that the sleeping time is randomly selected at each iteration between 5 and 15 seconds. This has been tested and it avoids being recognized as a bot, resulting in your IP to be temporarly banned. You can also set this to a custom time, but it is recommended to keep it around 10 seconds as shorter intervals may be problematic.
  • by_decade: if True (default), the lyrics are saved in a folder named as the decade when the song was produced. If False, the txt lyrics are just stored in a generic folder.
  • replace: If False (default), if two or more songs have the same name, all lyrics are saved in separate files. If True, then only the latest one gets saved.
  • folder: the name of the folder where the txt lyrics will be saved.

• get_artists(letter, home="https://www.azlyrics.com/")
  Another auxiliary function, which returns the urls and names of all the artists whose names start with a given letter.

• scrape_all(letters="all", sleep="random", by_decade=True, replace=False, folder="songs")
  This function downloads all the lyrics of all artists whose names start with a given letter. Note: I have estimated that in order to download each song of every artist on AZLyrics, the function should run non-stop for something like 27 weeks, given an average sleep time of 10 seconds between iterations. So this is just for fun, and not meant for actual use.
  Function Parameters:

  • letters: A list containing all the letters whose corresponding artists' lyrics should be downloaded. By default, it is set to "all", which means that every lyric contained on AZLyrics is downloaded.
  • sleep: The sleeping time (in seconds) between iterations. By default it is set to "random", which means that the sleeping time is randomly selected at each iteration between 5 and 15 seconds. This has been tested and it avoids being recognized as a bot, resulting in your IP to be temporarly banned. You can also set this to a custom time, but it is recommended to keep it around 10 seconds as shorter intervals may be problematic.
  • by_decade: if True (default), the lyrics are saved in a folder named as the decade when the song was produced. If False, the txt lyrics are just stored in a generic folder.
  • replace: If False (default), if two or more songs have the same name, all lyrics are saved in separate files. If True, then only the latest one gets saved.
  • folder: the name of the folder where the txt lyrics will be saved. If set to "names", then each artist will be downloaded in a separate folder with the corresponding name. Otherwise, all lyrics are collected in the same folder.

Examples:

Downloading all lyrics on AZLyrics.com: (See warning above!)

scrape_all(letters="all")

---

Downloading all lyrics of every artist starting with "a", each artist in a separate folder: (ETA = 1 week)

letter_list = ["a"]
scrape_all(letter_list, folder="names")

---

Downloading all lyrics of a given artist:

bruce = "https://www.azlyrics.com/s/springsteen.html"
scrape_artist(bruce, folder="bruce")

---

Downloading a single song lyrics:

the_chain = "https://www.azlyrics.com/lyrics/fleetwoodmac/thechain.html"
get_lyrics(the_chain, folder="fwm")