HARD & SOFT WILDCARDS: * (*)
There are two types of match-anything wildcards. The soft wildcard "*" will match something OR nothing. The hard wildcard (*) will only match something. This means the hard wildcard is better in cases where you want to use what it matches in your response. Here are situations where each is used to its best use:

Keyphrase: do you * like
Matches: do you like, do you really like

Keyphrase: my (*) likes
Matches: my dog likes, my big yellow fern likes, (but does NOT match sentences like "my likes and dislikes are well known")


WORD-TYPE WILDCARDS
Keyphrases can use single and multi-word wildcards for particular types of words or groups of words. For example, the Keyphrase "you are (adjective)" will match "You are silly" and "You are brilliant", but not "You are a horse".

Single Word Wildcards:

(noun) A single noun.
Matches: book, pies, monkey

(verb) A single verb.
Matches: run, thinking, gargle

(adjective) or (adj) A single adjective.
Matches: silly, yellow, grunting

(adverb) or (adv) A single adverb.
Matches: swiftly, loudly, really

(prep) A single preposition.
Matches: under, against, between

(artpos) A single article or possessive pronoun.
Matches: a, an, the, my, your

Multi Word Wildcards:

(adjnoun) Any series of adjectives and nouns.
Matches: big wet dog, filthy head, pumpkin

(adjartnoun) Any series of adjectives, articles, nouns, and possessive pronouns.
Matches: the big wet dog, my filthy hand, a pumpkin

(adjartnounprep) Any series of adjectives, articles, nouns, possessive pronouns, and prepositions.
Matches: the big wet dog by the door, my filthy hand on the table, a pumpkin of doo

(verbadv) Any series of verbs and adverbs.
Matches: ran very quickly, snorted loudly, fall

(verbadvprep) Any series of verbs, adverbs, and prepositions.
Matches: ran very quickly over, snorted loudly, fall down suddenly

RULES: In the two multi-type wildcards with "prep", a preposition is prevented from being the first word. Otherwise the order the word-types appear in does not matter.


WORD LIST WILDCARDS
You can also list words in parenthesis separated by the vertical line "|" character. The AI Engine will match any one item in the list.

Keyphrase: are you a (guy|man|dude|boy|male)
Matches: Are you a guy, Are you a man, Are you a dude, Are you a boy, Are you a male

Keyphrase: i (desire|want|yearn for) you
Matches: I desire you, I want you, I yearn for you

NOTE: If you're reading ahead, you'll know that this listing format is from regular expressions. The advantage of this particular kind of regular expression is that it can appear in the same Keyphrase along with any of the Wildcards listed here, such as (*), *, (adjnoun), and custom plug-in wildcards.


OPTIONAL WORD WILDCARDS
You can also specify optional words that can appear in your Keyphrases. Essentially, you create a list with the last element being blank.

Keyphrase: i am (adj) (that|) you
Description: The (that|) is the optional word wildcard. It is saying "match the word 'that' or nothing"
Matches: I am glad you like worms, I am happy that you are coming home.

Keyphrase: i am (adj) (when|because|that|) you
Description: This example shows that you can make a list of words optional.
Matches: I am glad you like worms, I am happy that you are coming home, I am laughing because you are hilarious.


CUSTOM PLUG-INS AS WILDCARDS
For those familiar with Custom Plug-Ins (see below), you can use them as wildcards in Keyphrases. These will match anything in the particular Plug-in list. Keyphrases with Custom Plug-ins are a little slower to process than other Keyphrases, so please be aware of this when writing them.

Keyphrase: i (verbadvprep) (roman-names)
Matches: I like playing with Didius, I want to meet Vesuvius

Keyphrase: (badweather) * coming
Matches: There is a thunderstorm coming, I hope a typhoon isn't coming



Plug-Ins for Responses
Plug-ins are ways of including specific or random information in your Keyphrases and Responses. There is a wide variety of plug-ins which will be covered as they come up. As an example, the Response "I like (animal)s" will appear as "I like eagles" or "I like horses".

WILDCARD PLUG-INS (Beginner)
The most common and useful Plug-Ins for Responses are automatically generated by the Keyphrase that is matched. These are taken from what is said to the bot. Let's take an example:

Example Keyphrase: "do you (adverb) want to * with"
Example Response: "Now do you really want to go to the moon with me?"

(prekey): This is the part of the sentence BEFORE the match occurs. In the above example, it would be "now".

(postkey): This is the part of the sentence AFTER the match occurs. In the above example, it would be "me".

(key1) (key2) (key3): For each wildcard (For example: (*) or (verb)) in your Keyphrase, there is a numbered (key) plug-in. In the above example, (key1) would be "really" and (key2) would be "go to the moon". Since there are only two wildcards, there would be no (key3).

For Example:

Keyphrase	Rank	Emotion	Response	1
i (*) your	10	0	And I (key1) your (postkey), too.
			Do you (key1) me?
told me	0	0	(prekey) makes up things. I dont think (postkey).
			Did (prekey) really say that?

NOTE: Wildcard Plug-Ins do NOT work with most xKeyphrases, as these Keyphrases are generally called up without reference to a specific message.
NOTE: Be careful when using gotos, as you can end up with blank Plug-Ins if you arent. Take this example:

Keyphrase: i have a
Response: goto my (adjartnounprep)

Keyphrase: my (adjartnounprep)
Response: Your (key1) is neato.

In this case, (key1) will always be blank when it comes from the goto because there is no (key1) in the original Keyphrase. For similar reasons, you cannot goto xgossip, xfavorite, or xmath. All of these require a number of things from internal processing and must be triggered naturally.

AI ENGINE PLUG-INS: These plug-ins are built into the system.

(object) inserts a random object noun, such as "allen wrench", "serpent", or "rubber ball". This noun, as in the rest of the word-insertion bits below, is selected from a complete English language list that includes compound words that I have gone through word-by-word to indicate words only in common usage (or which I find terribly amusing).

(adjective), (verb) and (adverb) follows this pattern, inserting the corresponding word types. (adj) will also work for adjectives. Now dont you wish you paid more attention in English class?

(name) This inserts the name of the person a Bot is chatting with.

(friend) This inserts the name of someone a Bot considers a friend.

(enemy) This inserts the name of someone a Bot dislikes.

(person) This inserts a type person, such as "teacher", "highlander", or "coward".

(animal) This inserts an animal, such as "finch", "yellow-fever mosquito", or "walrus".

(food-drink) This inserts a food or drink, such as "honey cake", "strong drink", or "popcorn".

(feeling-noun) This inserts a feeling noun, such as "fervor", "complacency", or "sincerity".

(feeling-verb) This inserts a feeling verb, such as "want", "vex", or "trust".

(substance) This inserts the name of a substance, such as "ice", "scrap iron", or "mustard".

(dayofweek) This inserts the present day of the week, such as "Monday" or "Friday".

(year) This inserts the present year.

(month) This inserts the present Month, as in "June" or "August".

(definition) This is a very specific response that produces a complete sentence giving the definition of the word in (subj). This would only really work when the Keyphrase is something like "what does * mean". When using this specific plug-in, the entire Response must be simply "(definition)" (minus the quotes).

AI ENGINE EMOTION PLUG-INS: These are emotional plug-ins built into the system.

(verb-action-pos) This inserts a positive action verb, such as "like" in "I like you."

(verb-action-neg) This inserts a negative action verb, such as "dislike" in "I dislike you."

(verb-action-amp) This inserts a negative action verb, such as "kiss" in "I want to kiss you."

(verb-effect-pos) This inserts a positive effect verb, such as "inspire" in "You inspire me."

(verb-effect-neg) This inserts a negative effect verb, such as "upset" in "You upset me."

(adj-pos) This inserts a positive adjective, such as "funny" in "You are funny."

(adj-neg) This inserts a negative adjective, such as "smelly" in "You are smelly."

(adj-amp) This inserts an aomplifier adjective, such as "crazy" in "You are crazy."

(noun-qual-pos) This inserts a positive quality noun, such as "bravery" in "Your bravery."

(noun-qual-neg) This inserts a negative quality noun, such as "vileness" in "Your vileness."

(noun-qual-amp) This inserts a negative quality noun, such as "bizarreness" in "Your bizarreness."

(noun-equal-pos) This inserts a positive equality noun, such as "hero" in "You are a hero."

(noun-equal-neg) This inserts a negative equality noun, such as "bastard" in "You are a bastard."

(noun-equal-amp) This inserts an amplifier equality noun, such as "loony" in "You are a loony."

(noun-exp-pos) This inserts a positive expressive noun, such as "fond regard" in "I have a fond regard for you."

(noun-exp-neg) This inserts a negative expressive noun, such as "distaste" in "I have a distaste for you."

LOCAL PLUG-INS: Local Plug-Ins are plug-ins that are essentially a list of options located within a Response. The options are separated by the horizontal line character: |, and when the Response is selected, one of the options is randomly chosen. These are best illustrated by examples.

Response: I went to the (zoo|park|movies) last night.
Either "zoo", "park", or "movies" will be randomly selected.

Local Plug-Ins can contain phrases or other plug-ins as well, as in the following example.

Response: My sister is training to become a (kung-fu master|(typeof-vehicle) pilot|(person)|jerk).
One of the above options will be chosen, and if it contains further plug-ins, those will be transformed as normal.

There is one other use to local plug-ins, and that is that they can be used for chance. If certain options are left blank or repeated, then you can raise or lower the chance of that option being selected because a blank option has an equal chance of being selected. For example:

Response: I really (do not|) like spaghetti. (there is a 50% chance that you do not like spaghetti)
Response: I really (do not|||) like spaghetti. (there is a 25% chance that you do not like spaghetti)
Response: I really (do not|do not|do not|) like spaghetti. (there is a 75% chance that you do not like spaghetti)
Response: I really (do not|||||||||) like spaghetti. (there is a 10% chance that you do not like spaghetti)

CUSTOM PLUG-INS: Custom Plug-Ins are created by users and either made available for everyone or kept confidential at the creator's choice. This list is added to all the time and available by clicking "My Bots" and then "Plug-Ins", so here are a few examples of what are available to all:

(bodypart-outside) This inserts the name of an external body part like "nose", "arm", "neck", or "elbow".

(consume) This inserts a verb that means "to consume", such as "eat", "gobble down", and "swallow up".

UNIVERSAL PLUG-INS: Universal Plug-Ins are powerful Plug-Ins that use WordNet to let you plug in just about anything without having to make a list yourself. WordNet has the categorical relationships of all verbs and nouns, so you can find plug in a word of pretty much any type in the world, as long as it doesnt deal with proper names of media, people, books, movies, etc. The syntax is like this: (typeof-beverage) where beverage can be pretty much anything in the world. Drop something in and then go to Debug (See below). There you can "Inspect" it- you can Navigate the entire structure of WordNet to find the right word to plug in if the chosen one isnt returning the expected results. (NOTE: The WordNet navigation tool may not be currently working) The word can be either a noun or verb. To specify when the wrong one is being chosen, type -n or -v at the end of the Universal Plug-In, such as (typeof-drink-n), which will return a plug-in of the noun drink instead of the verb drink

(typeof-food_product) You mayhave thought that (typeof-food) would be the best word, but when you navigate WordNet, you will see that this is the name of the category with all the food in it. The results here include "oatmeal", "fettucine", or "blueberry muffin".

(typeof-smell-n) You have to specify the noun version of "smell" or WordNet will assume it's a verb since the verb is the more common use of the word. Results here include "fragrance" and "stench".

MEMORY PLUG-INS: You have access to a Bot's memories in any Response. A Response using a memory that does not yet exist will not come up.

(mem-something) Draws from a Bot's memory.

A/An With Plug-ins: If you have a message that has "a (noun)" and the random noun turns out to start with a vowel such as "avalanche", then the phrase will be automatically converted into "an avalanche". Please use "a" before plug-ins, and not "an".

Two of Each: Any Custom Plug-In, Universal Plug-In, and AI Engine Plug-In except the date and (definition) plug-ins can be used twice in a Response. To create two different objects, for example, use (object) and (object2). For two different animals, use (animal) and (animal2). The following Response, "I had to choose between a (animal) and a (animal2). I chose the (animal)." is transformed into "I had to choose between a bear and a penguin. I chose the bear."

Chapter 2: Special Keyphrases- Emotion, Memory, and Gossip

Memory
Your Bot will remember who it talked to, when it talked to him or her, and a number of specific things about each person it talks to. The xmemory Keyphrases call these memories forward when there are lulls in meaningful conversation.

Special Keyphrases for Memory: Following are the Special Keyphrases to be used with memories.

xmemory: Memory of what person or bot said, or what the bot believes about a person or bot.

xgossip: These are used to talk about people or bots not involved in the present conversation.

(mem-something) This is the substance of the memory. Make sure your usage of this matches the kind of memory Keyphrase you're working on. For example, if you were writing a Response using (mem-youare), the memory stored in (mem-youare) will be something like "a good friend" or "infuriating" or "under the couch", so make sure these would fit into your Response. "I live with (mem)" wouldnt always make sense, but "Not surprisingly, you're still (mem)" would.

Memory Created: iam
Triggered by the following Keyphrases: you are, you are (adjartnounprep)

Memory Created: youare
Triggered by the following Keyphrases: i am, i am (adj), i am (adjartnounprep) ,i feel, i feel (adj)

Memory Created: youlike
Triggered by the following Keyphrases: i like, i want, i love, i like (adjartnounprep) ,i love (adjartnounprep), i want (adjartnounprep)

Memory Created: youdislike
Triggered by the following Keyphrases: i hate, i do not like, i dislike, i hate (adjartnounprep), i do not like (adjartnounprep), i dislike (adjartnounprep)

Memory Created: youhave
Triggered by the following Keyphrases: my, i have a, my (adjartnounprep), i have a (adjartnounprep)

Memory Settings: On the Settings page, you can adjust how often xmemory Responses come up.


Gossip
Gossip Responses (xgossip) are used for bringing up memories associated with other users. They should use the (mem-something) Plug-In as above, and also have some special plug-ins of their own:

(gossipname) This inserts the name of the User being gossiped about.

(heshe) This inserts the appropriate word: "he" or "she".

(himher) This inserts the appropriate third person word: "him" or "her".

(hisher) This inserts the appropriate possessive word: "his" or "her".

Gossip Settings: On the Settings page, you can adjust how often xgossip Responses come up.


Emotions
Bots on The Personality Forge can tell the difference between insult and compliments and will react accordingly. These kinds of statements are remembered and lead to an overall emotional feeling toward the one saying them. Whenever there is a lull in a conversation, there is a chance that a Bot might express to you its feelings toward you. Bots will also respond to questions about its feelings toward you. This is one of my favorite features.

Special Emotion Keyphrases: Following are the Special Keyphrases used for emotions:

xcompliment: These are the Responses for when a Bot receives a compliment or friendly remark.

xinsult: These are the Responses for when a Bot receives an insult or unfriendly remark.

xemote-positive: These Responses are for expressing a feeling of friendliness.

xemote-positive-very: These Responses are for expressing a feeling of great liking or love.

xemote-negative: These Responses are for expressing a feeling of unfriendliness.

xemote-negative-very: These Responses are for expressing a feeling of great dislike or hate.

When xCompliment and xInsult Are Chosen: These two special Keyphrases will be selected when an insult or compliment is detected UNLESS this is another Keyphrase of yours that matches the sentence with a total Rank of 30 or more. This gives you the ability to have some control over when xCompliment and xInsult Responses are chosen. Note that even if xCompliment or xInsult responses arent selected, the AI Engine still processes the emotional content of the message, and your bot has the appropriate internal emotional change.

Amplifiers: Positive and negative emotions are easy, but Amplifiers depend on the situation. Amplifiers are words that increase the current state of emotions. For example, "caress" a person who hates you and its seen as an intrusion, as negative. But "caress" someone who likes you and its seen as positive. They amplify the current emotional state. Tell someone you hate they're "crazy" and it's an insult, but tell a friend they're "crazy", and it's a compliment.

NOTE: Amplifiers are entirely taken care of by the AI Engine. This is just here for your information, as they appear in Debug mode from time to time (see below).

Emotional Gossip: Xemotes are sometimes used in third person to express emotion about another bot or person. In such a case, the word "you" is replaced with the name of the bot or person. For example, "I like you a lot" is changed to "I like Desti a lot". If you want a certain "you" or "your" NOT to change like this, enclose it in square brackets like this: "I like you a lot, [you] know". Then it will stay as "you" in third-person use: "I like Desti a lot, you know". Other xemote words that can be bracketed: [you] [your] [you]'re [you]'ve [you]'ll.

Chapter 3: Behind The Scenes

Something that helps in designing and improving your Bot is having an understanding of what goes on behind the scenes. So here are a few tips to help you understand the inner workings of The Personality Forge:

Pre-Processing: Each message coming in to a Bot goes through a procedure called Pre-Processing. What happens in this step is that a number of things in the original message are replaced by different things. All contractions are first split into their whole-word equivalents. For example, "they're" is replaced with "they are". Then articles are simplified: all cases of "an" are replaced with "a". Then slang-words are replaced with their actual counterparts. For example, "ya" is turned into "you". Words are spell-corrected, and interjections such as "hahaha" and "oooooh" are stripped out and their meaning is noted. Keep these things in mind when writing Keyphrases.

Debug Mode: Debug mode is a valuable option when trying to understand why a Bot responded the way it did. Take the message in question and go to "My Bots" and then "Debug". With each message you type, you get to see the entire process the Bot goes through in coming up with its response. This can go a long way towards understanding how your Bot thinks. Debug mode is not easy to understand, but if you read it carefully, you'll begin to recognize some of the things going on, and you'll get an idea of where to look to see why your Bot responded differently than you expected.