"functions":[
##$gpt.loadInputFromTemplate($context, "/$mediaarchive.getMediaDbId()/gpt/functiondefs/checkweather.json"),
#foreach($module in $chatprofile.getEntities())
$context.putPageValue("module",$module)
$gpt.loadInputFromTemplate($context, "/$mediaarchive.getMediaDbId()/gpt/functiondefs/search-moduletemplate.json"),
#end
$gpt.loadInputFromTemplate($context, "/$mediaarchive.getMediaDbId()/gpt/functiondefs/emediasearchformat.json")
], See this - he changed it to emediasearchformat
ok, feel like they haven't complete all the formatting, at least is not throwing an ugly error now, I gonna wait for shakil to see if he made some more progress formatting.
Caused by: org.apache.velocity.exception.MacroOverflowException: Max calling depth of 20 was exceeded in macro 'showcats' with Call Stack:showcats->showcats->showcats->showcats->showcats->showcats->showcats->showcats->showcats->showcats->showcats->showcats->showcats->showcats->showcats->showcats->showcats->showcats->showcats->showcats at /WEB-INF/base/finder/mediadb/gpt/systemmessage/analyzeasset.html?filter=translation&locale=en_US[line 2, column 1]
fixed couple of things in the scipt and let it ran some time, is not tagging anything, looks like always depending in having a ""tool_calls" in the message response from ollama, which I can't find anywhere in the code, is this and old concept or something?