Print

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

[em312s0n]

Number of possible different roasts that carry only preferred food (or drink in 2nd, 3rd or 4th spot) in my fortress:
3661038000 (that's 3661 millions)

Woah! how many dwarves do you have! lol. I dont think it would get to that number I mean even if you have 240 dwarves(a ludricous amount IMO) with no dwarf preferring the same food as another you'd only need at least 60 varieties of roasts that everyone can choose from and be satisfied

EDIT: I keep checking my math if I'm wrong lol but yeah. 1 roast= 4 ingredients that can satisfy. 240 dwarves/ 4 = 60 varieties of roasts.

[Saiko Kila]

Yes, I don't need that many of course (and I have about 350 dwarves, fluctuating). It's the maximum of different combinations, where each of them would satisfy at least one dwarf (but almost always more). Problem is, that the best combination will not satisfy more than a handful of dwarves, so the minimum number is also big. How big exactly - I don't know, because it's much harder to calculate.

[em312s0n]

Well 350 dwarves(I have got to know your PC specs) assuming they don't have overlapping food preferences would require at least 350/4=87.5 round up to 88 varieties of roasts. at most well its easy 1 custom roast per dwarf = 350 varieties.

I do agree on the strategy of identifying whiner dwarves lets call em snowflakes lol. Having to work up the effort to maintain 88 varieties man that is tedious.

[Saiko Kila]

To me even working on more than 10 roasts would be overwhelming. What I tried is to create manually roasts with the most liked alcohols - but turned out that the most popular alcohol, and the one I have in sufficient quantity, is liked by 4 dwarves. It takes top spot ex aequo with several other alcohols, liked by 4 dwarves each, followed by a bunches of alcohols liked by 3 dwarves and so on. It's even worse with solid foods.

But there's also problems with availability - most of what they love to eat is simply not available, and never will be. Unless the trading overhaul will allow to order items from any biomes... Heck, I even live in a savage non-freezing biome, and don't have whip vine!

[mikekchar]

You can satisfy most of your dwarfs without *that* much effort (especially if you use Patrik Lundell's script).  The main trick is to ensure that you don't over produce foods (e.g. massive farms of plump helmets).  Having shortages of some foods means that you get variety automatically.  You can also potentially make several kitchens with different ingredients nearby to enforce some variety.  Also, you *must* avoid using containers in your kitchen stockpiles (or else you end up with the same ingredient in each slot).

There are a couple of issues that are unavoidable with the current system, though.  First, there is no way to satisfy all of your dwarfs.  I think this is a big problem because eventually some of them will start dwelling on it.  The other issue is that while it isn't actually a huge amount of effort to satisfy most of the dwarfs, it *is* a huge amount of effort to figure out how to set up a system to do it.  It also requires some players to fairly comprehensively change their play style.  Before, you could make a massive farm, set up a kitchen, make thousands of roasts and forget about it.  Now it's actually pretty tricky to set up the production line -- and while *I* find that fun, I can imagine that there are many others who don't.

I kind of wish we could separate discussions based on this kind of criteria: I want to find out how to do it OR I want to express my desire to avoid doing it.  Mixing the two makes discussions hard to navigate IMHO.

[PatrikLundell]

I've tried to attack the "decent meal" issue from a script angle, i.e. write a repeating script that orders the kitchens to make lavish meals that have 3 random components and 1 drink that's favored by at least one resident. Rather than trying to cram as many favored ingredients into each roast, I'm gunning for the booze cooking angle, for a number of reasons:
- There's a limited number of kinds of drinks
- My standard fortress design already aims at producing every kind of booze known to dwarf (but not any kind of milk...).
- It's tricky trying to optimize among several favored foods.

The ideal would be to optimize ingredients such that only a single ingredient is favored, and all the others are "meh" to everyone, as that would reduce the risk of "wasting" ingredients when the dorf who likes sewer brew grabs the only remaining roast that has lamprey eyelid eyelash dandruff in it, ignoring the 305 other roasts that has sewer brew in it, which Urist McPicky only likes lamprey eyelid eyelash dandruff to eat, and wants to drink water (I've had a dorf with water as favorite drink, and I've also had dorfs who have no favorite food at all), and don't have anything left he likes, so he randomly takes the only roast containing purring maggot brain stem in it (rare indeed, as I don't think purring maggots have brain stems).

However, that gets a bit messy, so my initial approach would be to just go for the booze and let everything else get thrown in as they may be (which doesn't guard against other kinds/more booze being added). A fully fledged solution ought to specify the complete set of ingredients to make sure things aren't wasted. The drinks approach has the advantage from a complexity point of view that each dorf seems to have exactly one favored drink, while it can be quite messy to try to figure out which combinations you should go for if trying to take all preferences into consideration, especially given that you can't control who takes what.

My initial script attempt seems to be working, but I've got a groundhog day fortress that crashes after 2-6 days, so any real testing is hard to do.

[LordPorkins]

Roasted Plump helmet made from a sauce derivative from Elven Blood.

Oh your talking code. NVM.

[em312s0n]

My initial script attempt seems to be working, but I've got a groundhog day fortress that crashes after 2-6 days, so any real testing is hard to do.

☼«☼Script☼»☼

This is a masterwork script by PatrikLundell. It is encircled with bands of masterful lua code. On the item is an Image of Patrik Lundell the dwarf in unicorn leather. Patrik Lundell is typing code. The artwork relates to the creation of a masterwork script on the DF Forums in the early winter of 2018

[PatrikLundell]

This is the booze_cooker script in its current state:

Spoiler (click to show/hide)

Code: [Select]

local target_meal_count = 16  --  The number of meals to keep in stock per dorf.
                              --  16 should be 1 year's worth (1 meal every 3 weeks, according to wiki)
local verbose = true          --  Prints every meal added. You probably want to set it to false when/if
                              --  you believe it works correctly.

-----------------------------

function add (list, item, preference)
  for i, element in ipairs (list) do
    if element [1] == item then
      list [i] [2] = list [i] [2] + 1
      return
    end
  end
 
  table.insert (list, {item, 1, preference})
end

-----------------------------

function sort (list)
  local temp
 
  for i = 1, #list - 1 do
    for k = i + 1, #list do
      if list [k] [1] < list [i] [1] then
        temp = list [i]
        list [i] = list [k]
        list [k] = temp
      end
    end
  end
end

-----------------------------

function get_food_needs ()
  local preferences = {}
 
  for i, unit in ipairs (df.global.world.units.all) do
    if dfhack.units.isCitizen (unit) and
       dfhack.units.isAlive (unit) and
       unit.status.current_soul then
      for k, preference in ipairs (unit.status.current_soul.preferences) do
        if preference.active then
          if  preference.type == df.unit_preference.T_type.LikeFood then
            local found = false
           
            for l, pref in ipairs (preferences) do
              if pref.mattype == preference.mattype and
                 pref.matindex == preference.matindex and
                 pref.item_type == preference_item_type then
                found = true
                break
              end
            end
           
            if not found then
              table.insert (preferences, preference)
            end
          end
        end
      end
    end
  end

  local drink_list = {}
  local food_list = {}
 
  for i, preference in ipairs (preferences) do
    local material = dfhack.matinfo.decode (preference.mattype, preference.matindex)
           
    if preference.matindex ~= -1 and
       (material.mode == "plant" or
        material.mode == "creature") then
      if preference.item_type == df.item_type.DRINK or
         preference.item_type == df.item_type.LIQUID_MISC then  --  The state in the preferences seems locked to Solid
         add (drink_list, material.material.state_name.Liquid, preference)
             
      else               
        if material.material.prefix == "" then
          add (food_list, material.material.state_name.Solid, preference)
               
        else
          add (food_list, material.material.prefix .. " " .. material.material.state_name.Solid, preference)
        end
      end           
           
    else
      add (food_list, df.global.world.raws.creatures.all [preference.mattype].name [0], preference)
    end
  end
 
  sort (drink_list)
  sort (food_list)
 
--  dfhack.color (COLOR_YELLOW)
--  dfhack.println ("Drink preferences found:")
--  dfhack.color (COLOR_RESET)
 
--  for i, drink in ipairs (drink_list) do
--    dfhack.println (drink [1], drink [2], drink [3].mattype, drink [3].matindex)
--  end
 
--  dfhack.println ()
--  dfhack.color (COLOR_YELLOW)
--  dfhack.println ("Food preferences found:")
--  dfhack.color (COLOR_RESET)
 
--  for i, food in ipairs (food_list) do
--    dfhack.println (food [1], food [2])
--  end
 
  return drink_list, food_list
end

-----------------------------

function matching_meals (preference)
  local count = 0
 
  for i, food in ipairs (df.global.world.items.other.FOOD) do
    for k, ingredient in ipairs (food.ingredients) do
      if ingredient.mat_type == preference.mattype and
         ingredient.mat_index == preference.matindex then
        count = count + 1
        break
      end
    end
  end
 
  return count
end

-----------------------------

local drink_list = {}
local food_list = {}
local current_drink = 1
local max_jobs = 10
 
-----------------------------

function drink_available (preference)
  if preference.item_type == df.item_type.DRINK then
    for i, drink in ipairs (df.global.world.items.other.DRINK) do
      if drink.mat_type == preference.mattype and
         drink.mat_index == preference.matindex then
        return true
      end
    end
 
  elseif preference.item_type == df.item_type.LIQUID_MISC then
    for i, drink in ipairs (df.global.world.items.other.LIQUID_MISC) do
      if drink.mat_type == preference.mattype and
         drink.mat_index == preference.matindex then
        return true
      end
    end
 
  else
    local material = dfhack.matinfo.decode (preference.mattype, preference.matindex)
   
    dfhack.color (COLOR_LIGHTRED)
    dfhack.println ("Unaccounted for 'drinks' df.item_type of " .. df.item_type [preference.item_type] .. " for " .. material.material.state_name.Liquid)
    dfhack.color (COLOR_RESET)
  end

  return false
end

-----------------------------

function callback ()
--  dfhack.println ("callback invoked")
  local current_kitchen = 0
 
  if #drink_list == 0 then
      drink_list, food_list = get_food_needs ()
  end
 
  for i = current_drink, #drink_list do
    if current_drink < #drink_list then
      current_drink = current_drink + 1
   
    else
      drink_list, food_list = get_food_needs ()
      current_drink = 1
    end

    if drink_available (drink_list [current_drink] [3]) and
       matching_meals (drink_list [current_drink] [3]) < target_meal_count * drink_list [current_drink] [2] then
      while #df.global.world.buildings.other.WORKSHOP_KITCHEN [current_kitchen].jobs >= max_jobs do
        current_kitchen = current_kitchen + 1
       
        if #df.global.world.buildings.other.WORKSHOP_KITCHEN <= current_kitchen then
          current_kitchen = 0
          dfhack.timeout (7, 'days', callback) 
          return
        end
      end
     
      local job = df.job:new ()

      job.job_type = df.job_type.PrepareMeal
      job.pos.x = df.global.world.buildings.other.WORKSHOP_KITCHEN [current_kitchen].centerx
      job.pos.y = df.global.world.buildings.other.WORKSHOP_KITCHEN [current_kitchen].centery
      job.pos.z = df.global.world.buildings.other.WORKSHOP_KITCHEN [current_kitchen].z
      job.mat_type = 4
      job.wait_timer = 0
      job.general_refs:insert ("#", df.general_ref_building_holderst:new ())
      job.general_refs [0].building_id = df.global.world.buildings.other.WORKSHOP_KITCHEN [current_kitchen].id
      job.job_items:insert ("#", df.job_item:new ())
      job.job_items:insert ("#", df.job_item:new ())
      job.job_items:insert ("#", df.job_item:new ())
      job.job_items:insert ("#", df.job_item:new ())
     
      job.job_items [0].flags1.solid = true
     
      for l, job_item in ipairs (job.job_items) do
        job.job_items [l].flags1.unrotten = true
        job.job_items [l].flags1.cookable = true
        job.job_items [l].quantity = 1
        job.job_items [l].vector_id = df.job_item_vector_id.ANY_COOKABLE
      end
     
      job.job_items [3].item_type = drink_list [current_drink] [3].item_type
      job.job_items [3].mat_type = drink_list [current_drink] [3].mattype
      job.job_items [3].mat_index = drink_list [current_drink] [3].matindex
     
      df.global.world.buildings.other.WORKSHOP_KITCHEN [current_kitchen].jobs:insert ("#", job)
      dfhack.job.linkIntoWorld (job, true)
     
      if verbose then
        dfhack.println ("Adding job to prepare " .. drink_list [current_drink] [1] .. " roasts.")
      end
    end
  end
 
  dfhack.timeout (7, 'days', callback) 
end

-----------------------------

function booze_cooker () 
  if #df.global.world.buildings.other.WORKSHOP_KITCHEN == 0 then
    dfhack.color (COLOR_LIGHTRED)
    dfhack.println ("No kitchen found!")
    dfhack.color (COLOR_RESET)
    return
  end
 
  callback ()
end

booze_cooker ()


It can be noted that it does in fact cook non booze potable liquids (such as milk) as well, but I have no idea what it will do when water is called for. In my testing I ran the script on a 1 day repeat, while the posted version has a 7 day one. Also, I only have one kitchen, and I haven't been able to test it wrap around to start from the beginning.
Anyway, it might be useful in itself, or might be used as inspiration for others to develop further.
The script currently prints every roast order entered, which is probably useful initially to see what's happening, but annoying later on. Early in the script there's the definition of the 'verbose' variable, and changing "true" to "false" should turn off that output-

[em312s0n]

I like how you just nonchallantly solved the problem. will be testing once I get home from work. Cheers

[PatrikLundell]

I like how you just nonchallantly solved the problem. will be testing once I get home from work. Cheers

It's something I've been thinking of for quite some time for starters, and it's far from a proper solution, but rather the beginning of a path towards one. Also, those who look at the script will see it's built around the earlier script for determining fortress food preferences (the printout of that is commented away), so it's not really something just whipped up from nothing.
Anyway, I hope it may be of some use.

[Saiko Kila]

You can satisfy most of your dwarfs without *that* much effort (especially if you use Patrik Lundell's script).  The main trick is to ensure that you don't over produce foods (e.g. massive farms of plump helmets).  Having shortages of some foods means that you get variety automatically.  You can also potentially make several kitchens with different ingredients nearby to enforce some variety.  Also, you *must* avoid using containers in your kitchen stockpiles (or else you end up with the same ingredient in each slot).

I can't agree, unfortunately, after examining my dwarves. Note that Patrik's script takes only full citizens into account. When I modified it to count every guy from the Citizen tab (i.e. both citizens and long-term residents), I've got these numbers:
All citizens and long-term residents: 358
 Different Drinks: 101
 Different Foods: 347

It's simply not possible to meet needs of most of my dwarves, even if I try to make only meals with drinks. The variability is too high, almost everyone needs something different, and the number of unobtainable drinks (not to mention foods) is higher than number of obtainable ones, by wide margin. It's rare than dwarves need more than three things, usually they need 1, 2 or 3 things.

Some examples of unobtainable drinks (or liquids, since honey or moghopper juice don't really count as alcohols) in my population:
whip vine (biome doesn't contain despite being savage, not obtainable otherwise, I've got a guy who likes only this, no other food or drink, and total of six dwarves need it as alcohol)
bumblebee honey (not obtainable)
bumblebee mead (source ingredient not obtainable, I've got four guys who drink only this, and eat something not obtainable)
mead (can't use as food ingredient, I've got THREE guys who like only this, no other food or drink, and two guys who like mead and some unobtainable food)
rice beer (can't make rice beer from rice, and it's not plantable)
moghopper juice (can't find moghoppers, I've got a guy who likes both this and gutter cruor)
gutter cruor (evil savage biome only; aka sliver barb alcohol, one of my starting dwarves needs it)
sunshine (good biome only I think; aka sun berry alcohol, never seen it, I've got a guy who likes only this, no other food or drink)
tuber beer (only marshy biome, no merchants from such biome, two of my starting dwarves need it, very popular for some reason)

It's still better than the solid stuff, where statistical dwarf likes something like worm gut, i.e. part of vermin which is not obtainable. In case of plants though most alcohol-source plants are simply not brought by merchants, or not seen in my biome.

Rice beer and mead seem to be quite popular in my population, I don't know why mead can't be made into meals, or why rice can't be planted (it is both [DRY] and [WET]), but it prevents use of it as food even though I do have mead and I do have rice seeds, so it's a good example of problems the overseer has to face.

[PatrikLundell]

I haven't had problems with making rice wine. It's made from Rice Plants, while the seeds are called only Rice (and can't be brewed), according to the UI in my fortress (I don't have any Rice Wine at the moment, though, but Rice Plant is marked by the UI as Brewable, and I've had a fortress where I did at least once have every kind of booze on the wiki it the same time). Rice is ANY_TROPICAL according to the wiki. Banana beer can only be produced from elven imports because of a bug that causes "trees" that don't produce wood to never appear in an embark even if it should. Hacking Banana to produce wood allows it to appear in suitable embarks, though.

Mead can be used for cooking, but the DFHack interface for it doesn't support it, possibly because it's produced from honey via a reaction rather than through normal brewing of a plant. I believe my script can use mead, though, although I haven't been able to verify it. I haven't tried to set up a use of it the "normal" way (that's mikekchar's domain...).

A problem is that elves bring only fruit, which means they leave out berries (such a Sun Berries, should they have access to Good biomes). Since I embark in a good savage embark (with a slice of evil for Gutter Cruor and Glumprong) I typically don't have any problem with this one.

I've considered whether there's a need/use for a script that would replace specific food preferences with the generic animal one (i.e. cow, rather than cow heart) matching what the UI displays, and a vermin one with a randomly selected real creature, and possibly add a random food preference if the number of food preferences is zero.

[mikekchar]

I have to admit that I've never run a fortress with more than about 120 dwarfs.  I think the needs system in general would make it impossible to satisfy 350 dwarfs in just about anything.  However, there are special challenges for a large fortress like that.

When I look at my dwarfs, about 80% of them have either a food *or* a drink that is obtainable.  20% have all foods/drinks that are unobtainable.  The order screen from the dwarf caravan is crucial -- you have to order ahead of time to ensure that you get as much as you can.  And then for fruits, you have to trade with the elves.  Living in a biome that grows many plants is helpful too.   You should also try to get as much livestock as possible for making various cheeses -- again, for some animals order from the dwarfs and for others get what you can from the elves.  You can even grow livestock for meat and catch whatever fish you can get in your biome.
 Finally get various things from the humans.

With a smaller fortress the above is manageable (even without scripts) because it's all about setup.  You start with 7 dwarfs and you make sure that you have access to as much food as possible for them.  Then as you get migrants, you have to keep improving your setup.  It takes hours (like everything in DF), but once you have it set up, there isn't that much you need to do.  However, with 350 dwarfs (especially if you are just starting to set it up now), then you are in for a world of hurt.   I even limit my migrants to 10 or so per year so I have time to set up everything for their needs.

The other problem with a really large fortress is that you are probably going to want multiple kitchens, multiple farms, multiple breweries, etc, etc.  It's going to cover a large area.  I think there is considerable possibility that even if you make all the food you need, you won't easily be able to store your food in a small enough area that the dwarfs will search for the food they want.  Or if you do, it won't be possible to get the distribution to the kitchens right so that you can hack the production methods.

So I can definitely see that for your fortress it's probably going to be very difficult or impossible to fulfil their food needs.  That doesn't mean it's impossible to satisfy the majority of food needs in general, though.

[Sylverone]

Okay raiding parties need to be able to start raiding NATURE for these missing ingredients. Problem half-solved then. It would also be neat if someday distant you could make requests to foreign companies who would send parties to gather things and then trade them to you. Not that this is the suggestions forum, just thinking aloud.