abrax edited this page 5 years ago


Is lizard only for Linux?

Since Python is cross platform it should run if Python runs, but I only use it on Linux so currently instructions are for that. They shouldn't be hard to adapt to other OSes, but if anyone is actually interested in running this on something else I can try to provide specific instructions.

Is there a GUI?

No. An earlier version of this program had a GUI, but it turned out to be more trouble than it's worth.

Why doesn't Lizard automatically refresh threads at set intervals?

The user is expected to use the scheduling system of their OS to accomplish this.

For instance, on Linux you can set up cron to run lizard rc regularly, and maybe redirect output to your own log file as well.

How can I inspect the database?

lizard l will print out a summary, but if you want the nitty gritty you need a SQLite browser such as sqliteman or DB Browser for SQLite.

Why does it create so many .json and .html files in the data directory?

These are saved at each refresh in case the thread dies or any replies are deleted. Deleting these files when they are no longer needed is the user's responsibility.

How do you handle file names?

I use the original filenames with the first 10 characters of 8chan's hash-based name. Adding the hash is necessary because 8ch allows duplicate filenames in the same thread, and they would otherwise conflict with each other when saved with the original name.

How does the conservative refresh predict replies?

If the last reply was x seconds ago, and the thread was last refreshed y seconds ago, lizard doesn't bother refreshing unless x/y>10. Internally, the 10 is called conservative_refresh_criterion. Increasing this will make it less conservative, decreasing will make it more conservative.

In reality I think replies would follow a bimodal (people who browse only the front page and people who browse the catalog) or trimodal (bimodal plus people who watch individual threads, such as with this tool) distribution, with each mode itself being a Poisson. Interesting as that is, I think my simple heuristic works well enough so I'm not planning to implement such rigorous statistical analysis for the time being.