From f31989c2d0460d86860b39f6faeeb09c6e70b297 Mon Sep 17 00:00:00 2001
From: Matt Brockman <mattbrockman@gmail.com>
Date: Mon, 18 Jan 2021 19:34:07 -0800
Subject: [PATCH] added readme on evaluationBots

---
 .../EvaluationBotData/README.md               | 29 +++++++++++++++++++
 examples/evaluationBots.js                    |  1 +
 2 files changed, 30 insertions(+)
 create mode 100644 examples/contextModifiers/EvaluationBotData/README.md

diff --git a/examples/contextModifiers/EvaluationBotData/README.md b/examples/contextModifiers/EvaluationBotData/README.md
new file mode 100644
index 0000000..c26efb0
--- /dev/null
+++ b/examples/contextModifiers/EvaluationBotData/README.md
@@ -0,0 +1,29 @@
+# Evaluation Bots
+Bots and their usage
+
+# Input
+- `DCStatDifficultyBot` - Evaluates the stat required and difficulty of an action (returns hard/medium/easy difficulty)
+- `InputDCattributeBot` - Evaluate the attribute and difficulty of an action (returns numeric difficulty)
+
+# Input Special Usage
+- `SimplePossiblityBot` - Evalutes two possible actions, PossibleAction1 and PossibleAction2.
+  - Special Usage: When `state.inputBot='SimplePossiblityBot'` you need to provide a character description in `state.contextualString`. This can be up to 75 characters (additional length will be truncated). `SimplePossiblityBot` will use the character description with the actions to provide information on what the character can do in the situation. Useful for enforcing limitations such as missing limbs or weapons.
+
+# Output
+The following bots are best used on output.
+
+## Multipurpose Evaluation Bots
+
+- `GoblinBot` - Evaluates whether it detected any Goblins killed, health was lost, and loot gained
+- `KittenBot` - Evaluates whether rapport was gained or a kitten got hungrier
+- `SpaceLootBot` - Evaluates whether loot would be found (on a spaceship)
+- `HungerBot` - Evaluates whether the player would get hungrier or not
+
+## Simple Evaluation Bots
+There are several evaluation score bots. These bots can be used on either input or output to score different text. They all return a `reason` and a `score` which is roughly along the below descriptions.
+- `KillBot` - Evaluates whether the player is killing things
+- `JudgeBot` - Evaluates whether an action was just
+- `EmpathyBot` - Evaluates whether an action showed empathy
+- `SuccessBot` - Evaluates whether an action succeeded
+- `SantaBot` - Evaluates whether an action showed christmas spirit (An overly judgemental SantaBot also exists but is not currently available)
+
diff --git a/examples/evaluationBots.js b/examples/evaluationBots.js
index 013934f..b9fa1b4 100644
--- a/examples/evaluationBots.js
+++ b/examples/evaluationBots.js
@@ -31,4 +31,5 @@ modifier(text)
 ['SpaceLootBot'] -> Loot
 ['DCStatDifficultyBot'] -> Stat, Difficulty
 ['HungerBot'] -> Fullness
+['SimplePossiblityBot'] -> PossibleAction1, PossibleAction2
 */
\ No newline at end of file