Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Tachytaenius]

This is all for DF 0.47.05.
A while back, I started a DFHacked-up hermit save. I set up manager orders for everything that needed doing, and let my hermit research and write books (with the game's FPS increased to as much as my computer could handle) for about 16 centuries (the hermit has a no aging syndrome). I had a script to check progress towards discovery of every topic:

Code: (knowledge-progress.lua) [Select]

local knowledge = df.historical_figure.find(dfhack.gui.getSelectedUnit().hist_figure_id).info.known_info.knowledge

local keys = {
"philosophy", "philosophy2",
"math", "math2",
"history",
"astronomy",
"naturalist",
"chemistry",
"geography",
"medicine", "medicine2", "medicine3",
"engineering", "engineering2"
}

local knownThings, thingsToKnow = 0, 0

for _, key in ipairs(keys) do
for knowledge, known in pairs(knowledge[key]) do
if not tonumber(knowledge) then -- skip unnamed bits
thingsToKnow = thingsToKnow + 1
if known then
knownThings = knownThings + 1
end
end
end
end

print("Knowledge: " .. knownThings .. "/" .. thingsToKnow)
But eventually she stopped at 310/312 topics discovered, doing Research! instead of Ponder [remaining topic]!
I also burned out because I had too many scripts and such to migrate to version 50 when it came out.

But yesterday I tried again to get back into DF, and have now managed to successfully force my hermit to ponder the remaining topics!
I did some research about topic mechanics on the wiki, and made a hacky little script to force research so that a breakthrough could happen (explanation is also in the script lol):

Code: (knowledge-fix.lua) [Select]

-- knowledge-fix
-- By Tachytaenius

-- I had a hermit save where a hermit was working on discovering every single topic in the game.
-- She stopped at 310/312 topics learned-- Research! instead of Ponder [remaining topic]!
-- The topics (at least in my case) are philosophy's philosophy_logic_propositional_logic and engineering2's engineering_design_models_and_templates.
-- This script forces your hermit (currently selected unit) to ponder those two remaining topics by unlearning an already-known topic, causing the hermit to ponder it,
-- and then switching the pondering to be about the actual topic (also relearning the unlearned topic).
-- To start the script, run (while your hermit is selected): knowledge-fix start [name of topic to unlearn to trigger pondering] [name of topic to try to learn]
-- and allow research to happen. An example topic to unlearn is engineering_machine_trip_hammer.
-- To stop it, just run knowledge-fix.
-- You should set breakthrough announcements to pause and stop this script once a discovery happens. Otherwise they continue pondering the broken topic after discovering it.
-- To do this temporarily, run:
-- gui/gm-editor df.global.d_init.announcements.flags.RESEARCH_BREAKTHROUGH
-- and set DO_MEGA, PAUSE, and RECENTER to true.
-- This script is hacky and focused on a specific case, so be sure to use it right.

local repeatUtil = require("repeat-util")
local repeatName = "knowledgeFix"

local start = select(1, ...) == "start"
if not start then
repeatUtil.cancel(repeatName)
print("Cancelling")
return
end

local knowledgeCategories = { -- Indices here start at 1 but should start at 0 for DF
"philosophy", "philosophy2",
"math", "math2",
"history",
"astronomy",
"naturalist",
"chemistry",
"geography",
"medicine", "medicine2", "medicine3",
"engineering", "engineering2"
}
local function getTopicInfo(knowledgeName) -- Given topic name, returns knowledge category name, knowledge category id, and topic id within category
for luaCategoryIndex, categoryName in ipairs(knowledgeCategories) do
local categoryIndex = luaCategoryIndex - 1
local bitfieldType = df["knowledge_scholar_flags_" .. categoryIndex]
for topicIndex, topicName in ipairs(bitfieldType) do
if topicName == knowledgeName then
return categoryName, categoryIndex, topicIndex
end
end
end
error("Couldn't find this topic")
end

local unlearnTopicName = select(2, ...) -- Anything of the working ones they've already learned
local learnTopicName = select(3, ...) -- philosophy_logic_propositional_logic or engineering_design_models_and_templates

local unlearnCategoryName, unlearnCategoryId, unlearnTopicId = getTopicInfo(unlearnTopicName)
local learnCategoryName, learnCategoryId, learnTopicId = getTopicInfo(learnTopicName)

local unit = dfhack.gui.getSelectedUnit()
local figure = df.historical_figure.find(unit.hist_figure_id)
local knowledge = figure.info.known_info.knowledge

local function tickFunction()
-- Make sure research is going okay (temporary?)
knowledge.knowledge_goal_category = learnCategoryId
knowledge.unk_1 = learnCategoryId
knowledge.unk_2 = learnCategoryId
knowledge.unk_3 = learnCategoryId
for i in ipairs(knowledge.knowledge_goal) do
knowledge.knowledge_goal[i] = false
end
knowledge.knowledge_goal[learnTopicId] = true
knowledge.research_points = 100000
knowledge.times_pondered = 100

if #unit.social_activities == 0 then
return
end

local activity = df.activity_entry.find(unit.social_activities[0])
if activity.type ~= df.activity_entry_type.Research then
return
end

local ponderEvent
for _, event in ipairs(activity.events) do
if event._type == df.activity_event_ponder_topicst then
ponderEvent = event
break
end
end

if not ponderEvent then
-- Unlearn a topic so that the hermit starts pondering it
-- We relearn it after and switch the pondering's topic to the topic we actually want to learn
knowledge[unlearnCategoryName][unlearnTopicName] = false
return
end

-- We've managed to create a ponder event previously, so if we're currently still on the unlearned topic, unset it from the unlearned topic and set it to the learned topic
if ponderEvent.knowledge.flag_type == unlearnCategoryId and ponderEvent.knowledge.flag_data["flags_" .. unlearnCategoryId][unlearnTopicName] then
-- Unset unlearned
ponderEvent.knowledge.flag_data["flags_" .. unlearnCategoryId][unlearnTopicName] = false
-- Set learned
ponderEvent.knowledge.flag_type = learnCategoryId
ponderEvent.knowledge.flag_data["flags_" .. learnCategoryId][learnTopicName] = true

-- And relearn the unlearned topic
knowledge[unlearnCategoryName][unlearnTopicName] = true
end
end

repeatUtil.scheduleEvery(repeatName, 1, "ticks", tickFunction)

Maybe the bug that causes propositional logic and models & templates to not be researchable isn't applicable to other versions (or saves!) of the game, but it was for me and I'm pleased and proud that I managed to fix it in a reasonable manner.

She now knows all 312/312 topics.