Python script for collecting basic YouTube search results. This script uses the YouTube API to gather information from videos based on specified queries and release date constraints, collected into .csv, .html and .pdf files. It also uses ChatGPT to generate a few recommendations of video titles for the specified queries.
This is Python3 code. I used Python 3.11.
Install required packages with
pip install -r requirements.txt
Python 3 is available on the Microsoft Store.
The WeasyPrint package that generates the PDF reports requires the Pango library. The recommended way to do this is to install MSYS2 and run
pacman -Smingw-w64-x86_64-pango
In the MSYS2 shell, then close the shell. More details on the WeasyPrint documentation.
If you encounter issues with missing modules, it may be because Windows is not looking for them in the right directory.
Check the path of pip and python with
get-command pip
get-command python
If they do not match or do not look correct (for me that's C:\Users\<myusername>\AppData\Local\Microsoft\WindowsApps\...), check the system environment variables on Windows System Properties.
Make a copy of the configuration file template
cp conf_template.py conf.py
You must add your API keys for YouTube and OpenAI to the conf.py file.
You can use the name of the input and logo files provided or change accordingly.
The script reads from the QUERIES_FILE specified in conf.py. The file should contain on each line a query phrase; the
desired maximum number of days elapsed since the video release; and the maximum number of subscribers of the video channels. Separate them with ;.
Edit the queries_template.txt file as needed or make a
copy and change the file name in conf.py.
For the .html and .pdf files, the script uses the logo file stored in static/images. Set it in conf.py accordingly.
The script generates collected_result_<date>.{csv,html,pdf} files into the output directory.
python3 main.py