Poetry is very charged (high potential for negative sentiment words). Is there a flag I can set to use an alternate dictionary?

[bean5]

Other than the title, the project seems to work really well. I like the idea of using poetry to increase memorability. I would like to use this in an environment where it is used in place of guids where there are likely to be exposed to users/customers and I have tested a few runs with it. The results seem more negative than silly. My experience with random phrases has been with docker. The generator docker uses tends to be silly and generates phrases such as "wheezy pencil" while this project seems to make things like "frozen corpse" or "mean insult war".

[ggozad]

😂😂😂

I guess "frozen corpse" and "mean insult war" are far easier to remember than "wheezy pencil" which is the whole point...
I think you just had bad luck.

That said, google says it comes from https://github.com/moby/moby/blob/master/pkg/namesgenerator/names-generator.go
I will happily accept a PR that introduces an optional alternative set of words (you need 1626 of them).

[bean5]

They have maybe 400 words and they combine an adjective with a name, which is easier to control, which may be why they stopped updating that code.

[bean5]

This is getting interesting. I decided to locate the bitcoin word list. If you put the entire set into https://monkeylearn.com/sentiment-analysis-online/ the result is Negative 98.5% confidence. After removing a few words that are more difficult in English (words that contain "ph" and "kn") and some other words I felt might be a bit charged, the result is Neutral 73.7% confidence.

I am going to continue pruning words to see what I get in the end. If I were more efficient at this, I'd just rank all the words and take the top-most positive programmatically. Manual seems to be working though.

[bean5]

@ggozad Pull request is created.