We use Plotly in conjunction with DVC, where DVC tracks the html file that Plotly generates as the output of one of our stages. DVC does some basic equality checking on the output files based on checksums. If a stage is run, and the Plotly-generated HTML changes its contents, we'll see that reflected in a different checksum in dvc.lock.

We found that Plotly will generate two different files, even if the input is identical. This is likely caused by this line, which is later emitted into the HTML:

What are the reasons that lead to using a random ID? Could this potentially be some non-changing value?