Pick a date, scrape the Billboard Hot 100 of that day and get a private YouTube Music playlist of the songs. A solution to day 46 of Udemy's 100 Days of Code: The Complete Python Pro Bootcamp, for fellow course participants and anyone browsing for ideas.

📚 Table of Contents

Overview
Requirements
Setup
Run
Tests
License
Links
Appendix: Project documents

🧭 Overview

The program follows the four steps of the lecture:

Scrape the chart. Ask for a date (YYYY-MM-DD) and read the 100 song titles from the Bakeboard clone of the Billboard Hot 100 (https://appbrewery.github.io/bakeboard-hot-100/) with requests and BeautifulSoup.
Authenticate. Open a YouTube Music session with ytmusicapi from a browser.json file.
Create the playlist. Create a private playlist named {date} Billboard 100, unless it already exists.
Add the songs. Search every song and add the top result; a song that fails is skipped.

Layout of the repository:

Folder	What is in it
`src/musical_time_machine/`	The program, in the layers `domain`, `application`, `infrastructure` and `presentation`; all constants are in `constants.py`
`tests/`	The automated tests (pytest); they use a stored chart page and a fake music service, never the network
`docs/`	Business case, plan, milestones, reviews and the wiki sources

The wiki (docs/wiki/, published to the repository wiki) is written in English, Danish and Bahasa Malaysia and explains what each lecture step teaches: English, Dansk, Bahasa Melayu.

📋 Requirements

Python 3.13 or later.
Three runtime libraries, each needed by the assignment: requests (fetch the page), beautifulsoup4 (read the HTML) and ytmusicapi (talk to YouTube Music).
A Google account and Firefox, only for steps 2 to 4 (to create browser.json).
Optional: Doxygen to build the API documentation.

🛠️ Setup

Create and use a local virtual environment named .venv, then upgrade pip:

python -m venv .venv

Activate it (Windows PowerShell):

.venv\Scripts\Activate.ps1

Activate it (Linux and macOS):

source .venv/bin/activate

python -m pip install --upgrade pip
python -m pip install -e ".[dev]"

Leave out [dev] when you only want to run the program.

browser.json (steps 2 to 4 only) holds your YouTube Music login cookies. It is ignored by git: never commit it.

Open Firefox, go to https://music.youtube.com and log in.
Open the developer tools (Ctrl+Shift+I, on macOS Cmd+Option+I), choose the Network tab and type browse in the filter.
Pick a successful browse?... request to music.youtube.com, right-click it and choose Copy Value, then Copy Request Headers.
In the project folder run ytmusicapi browser and paste the headers, then finish with Ctrl+D (Ctrl+Z and Enter on Windows). On macOS run pbpaste | ytmusicapi browser.

The cookies expire after a while; repeat the steps to get a fresh file.

If --check-login says Expecting value, the file holds entries that are not request headers (a Chrome copy can add Request URL, Decoded, music.youtube.com, a path entry or content-encoding). Copy only the Request Headers block, or use Firefox's Copy Request Headers, and create the file again.

▶️ Run

The program is added step by step (one pull request per lecture step):

Step	Status
1. Scrape the Billboard Hot 100	done (below)
2. YouTube Music authentication	done, with a live run screenshot
3. Create the playlist	done, with a live run screenshot
4. Search and add songs	done, with a live run screenshot

Step 1 - Scrape the Billboard Hot 100

python -m musical_time_machine --list-only

The program asks for a date in the format YYYY-MM-DD and prints the 100 songs of that chart as rank. title - artist. Use a date that Bakeboard has, for example 2026-04-18. scrape_titles returns the titles as a Python list; a date in the wrong format or a chart that does not exist ends the program with a clear message.

The chart page on Bakeboard:

The program's output for the same date (rendered from the real output of the run, not a photo of a terminal window):

Step 2 - YouTube Music authentication

Create browser.json as described in Setup, then check the login:

python -m musical_time_machine --check-login

It prints Found N playlists in your library.. With no or an expired browser.json it prints the renewal steps and exits with code 1 instead of a stack trace. --browser-file PATH uses another file.

A live run with the author's browser.json (the output shows only the count, never playlist names):

The file holds personal cookies and is never committed. The next screenshot shows the run without a browser.json and the tests that use a fake YouTube Music service:

Step 3 - Create the playlist

python -m musical_time_machine --create-playlist

It asks for the date and creates the private playlist YYYY-MM-DD Billboard 100 in your YouTube Music library. It first looks through get_library_playlists() for a playlist with the same title, so a second run for the same date creates nothing and says so.

A live run with the author's browser.json: the first run creates the playlist, the second run creates nothing:

The next screenshot shows the tests of this step, which use a fake YouTube Music service (no account needed):

Step 4 - Search and add songs

python -m musical_time_machine

The full program: it asks for the date, scrapes the chart, creates the playlist (or reuses the existing one), then searches YouTube Music for each song (yt.search(..., filter="songs")) and adds the top result with yt.add_playlist_items(). A song that is not found or fails is skipped and the run goes on; the summary says how many songs were added and which were skipped. A failed login stops the run, because every following song would fail too. Use --list-only to print the chart without logging in (step 1).

A live run with the author's browser.json: all 100 songs were added:

The next screenshot shows the tests of this step with a fake YouTube Music service:

🧪 Tests

python -m pytest

To build the API documentation (written as Doxygen comments in the source):

doxygen Doxyfile

📄 License

GNU Affero General Public License v3.0, see LICENSE.

🔗 Links

📎 Appendix: Project documents

Document	What it tells you
AGENTS.md	The rules AI agents follow in this repository
Business Case	Why the project exists, its scope and success criteria
Stakeholder Analysis	Who is involved; the stakeholder IDs used for owners and reviewers
Project Plan	The gateways and their schedule
Milestones	One document per gateway: Go/No-Go criteria and tasks
Artifact registry	Where each artifact type lives, the next version, the PO language and domain
Traceability matrix	How the artifacts link and when each was last reviewed
Governance	The review workflow, who decides and how disputes escalate
Framework README	The framework this project uses
Framework changelog	What changed in each framework release