Various changes

docs/docs/flows.mdx

@@ -0,0 +1,294 @@
+---
+id: flows
+sidebar_label: Business Logic with Flows
+title: Business Logic with Flows
+abstract: |
+  Flows in Rasa offer a structured way to design conversation-driven business
+  logic. The documentation details the rules for effectively using and managing
+  Flows within Rasa.
+---
+
+import RasaDiscoveryBanner from "@theme/RasaDiscoveryBanner";
+
+<RasaDiscoveryBanner/>
+
+## Overview
+
+**Flow** is a novel Rasa feature designed to build intuitive,
+conversation-centric business logic for your chatbots. Flows are primarily
+devised using Studio's Flow builder. However, as a Rasa core feature, they also
+offer a YAML counterpart, yielding a specification for creating Flow primitives
+in YAML.
+
+In essence, a **Flow** embodies a conversational state machine in the form of a
+directed graph. The primary building block of a Flow is a **Step**, equivalent
+to a node in the graph. Each step outlines the potential subsequent steps, which
+may be associated with certain conditions. Hence, graph vertices are embedded in
+steps.
+
+### YAML Specification for Flow
+
+A Flow is defined using the following structure:
+
+```yaml
+flows:
+  a_flow: # an id
+    name: A flow # a name
+    description: optional description of a flow
+    steps: [] # a list of steps
+```
+
+## Steps
+
+A `Flow` comprises `step`s. All steps share these common properties:
+
+```yaml
+id: a_required_unique_id
+description: an optional human readable description of the step
+metadata: {} # an optional dict
+next: an id or conditionals to determine which step should be executed next
+```
+
+Steps can manifest in one of three types: `user_message`, `action`, or
+`question`.
+
+### User Message (`user_message`)
+
+This step signifies a user utterance. Here's an example:
+
+```yaml
+- id: a_required_unique_id
+  intent: open_bank_account
+  entities: [] # optional
+  description: open a bank account
+  metadata: {} # an optional dict
+  next: "5"
+```
+
+In this example, the user intent is to open a bank account. The `next` field
+indicates the ID of the next step to be taken in the flow, which would be the
+step with the ID "5".
+
+A `user_message` step can also support `OR` statements. This means that the flow
+can proceed if either of the conditions in the `OR` statement are met:
+
+```yaml
+- id: a_required_unique_id
+  or:
+    - intent: complaint
+      entities: [] # optional
+    - intent: talk_to_agent
+      entities: [] # optional
+  description: handle a complaint or talk to an agent
+  metadata: {} # an optional dict
+  next: "5"
+```
+
+Here, if the user intent is either to make a complaint or to talk to an agent,
+the flow would proceed to the step with ID "5".
+
+### Action (`action`)
+
+An `action` step is where the bot performs a specific action or utters a
+message. Here are two examples:
+
+```yaml
+action: action_xyz
+```
+
+This could be a custom action where the bot performs a specific function defined
+in your custom actions script.
+
+```yaml
+action: utter_xyz
+```
+
+This is an utterance action where the bot sends a predefined message to the
+user. The message for "utter_xyz" would be defined in the domain file under
+responses.
+
+### Question (`question`)
+
+In a `question` step, the bot asks a question, with the aim usually being to
+fill a slot. Here is an example:
+
+```yaml
+question: account_type # slot name
+```
+
+In this example, the bot is
+
+asking the user to provide their account type. The response will be used to fill
+the "account_type" slot.
+
+A question can be configured to be skipped if the slot is already filled, as
+shown below:
+
+```yaml
+question: first_name # slot name
+skip_if_filled: true
+```
+
+In this scenario, if the "first_name" slot is already filled, the bot will skip
+this question.
+
+The slot can also be configured to be cleared at the end of the flow execution.
+The following configuration would clear the slot when a leaf of the flow is
+reached OR when exiting a `link`ed node:
+
+```yaml
+question: yes_or_no # slot name
+scope: flow | global
+```
+
+For the `slot_type`, it's recommended to use `text` slots in Flows.
+
+### Scope (`scope`)
+
+Scopes determine the lifespan of a slot or variable. There are two types of
+scopes available:
+
+- **flow**: The variable exists only for the duration of the flow. Once the flow
+  ends, the variable ceases to exist.
+- **global**: The variable persists across flows until it is explicitly changed.
+
+### Fill (`fill`)
+
+`Fill` refers to the process of filling slots based on user responses. It is a
+simplified version of slot mapping. Below is an example of mapping from
+entities:
+
+```yaml
+fill:
+  entity: entity_name
+  window: question | flow | global
+```
+
+Here, the "entity_name" entity from the user's message is extracted and used to
+fill a corresponding slot.
+
+Or mapping from intent (or button payload):
+
+```yaml
+fill:
+  intent: affirm
+  value: "true"
+  window: question
+```
+
+Here, if the user's intent is "affirm", the corresponding slot will be filled
+with the value "true".
+
+### Next (`next`)
+
+The `next` field in steps allows for branching logic. For instance:
+
+```yaml
+- id: "2"
+  question: Do you have any other questions?
+  next:
+    - intent: affirm
+      id: "3"
+    - intent: deny
+      id: "4"
+```
+
+In this example, depending on whether the user's intent is "affirm" or "deny",
+the flow will proceed to step "3" or "4", respectively. This allows you to
+create complex and dynamic conversational flows that adapt to user inputs.
+
+## Examples
+
+### Hello World Flow
+
+```yaml
+flows:
+  hello_world:
+    name: Hello World!
+    description: A basic flow
+    steps:
+      - id: "1"
+        intent: greet
+        next: "2"
+      - id: "2"
+        action: utter_greet
+```
+
+The intent `greet` and the response template `utter_greet` need to be defined in
+the domain file.
+
+The `next` property determines the id of the next step and links all the 
+different steps together.
+
+### A basic flow with branching
+
+```yaml
+flows:
+  basic_flow_with_branching:
+    name: basic flow with branching
+    description: a flow with a branch
+    steps:
+      - id: "1"
+        intent: greet
+        next: "2"
+      - id: "2"
+        action: utter_greet
+        next: "3"
+      - id: "3"
+        question: age
+        next:
+          - if: age < 18
+            then: "4"
+          - else: "5"
+      - id: "4"
+        action: utter_too_young
+      - id: "5"
+        action: utter_old_enough
+```
+
+### Asking for confirmation
+- TODO: include flow but also slot configuration from domain
+- TODO: include branching on the result of the user selection
+
+### Multiple Flows with links
+
+```yaml
+flows:
+  flow_with_links:
+    name: flow with links
+    description: optional description of a flow with links
+    steps:
+      - id: "1"
+        intent: greet
+        next: "2"
+      - id: "2"
+        action: utter_greet
+        next: "3"
+      - id: "3"
+        link: collect_personal_details
+        next:
+          - if: age < 18
+            then: "4"
+          - else: "5"
+      - id: "4"
+        action: utter_too_young
+        next: "6"
+      - id: "5"
+        action: utter_old_enough
+        next: "6"
+      - id: "6"
+        link: another_flow
+
+  collect_personal_details:
+    name: collect details
+    description: gather personal details from the user
+    steps:
+      - id: "1"
+        question: first_name
+        next: "2"
+      - id: "2"
+        question: last_name
+        next: "3"
+      - id: "3"
+        question: age
+```

docs/docs/llms/flow-policy.mdx

@@ -0,0 +1,11 @@
+---
+id: flow-policy
+sidebar_label: Flow Policy
+title: Flow Policy
+abstract: |
+  Running Flows with the FlowPolicy.
+---
+
+import RasaDiscoveryBanner from "@theme/RasaDiscoveryBanner";
+
+<RasaDiscoveryBanner/>

docs/docs/llms/large-language-models.mdx

@@ -0,0 +1,70 @@
+---
+id: large-language-models
+sidebar_label: LLMs in Rasa
+title: Using LLMs with Rasa
+className: hide
+abstract:
+---
+
+import llmUsageArchitecture from "./llm-usage-architecture.png";
+import RasaLabsLabel from "@theme/RasaLabsLabel";
+import RasaLabsBanner from "@theme/RasaLabsBanner";
+
+<RasaLabsLabel />
+
+<RasaLabsBanner />
+
+Large Language Models (LLMs), pretrained models using retrieval-augmented
+generation, can be utilized with Rasa in several ways. This document offers an
+overview of those usage patterns and how to handle them.
+
+## A future-proof approach using generative models
+
+The recent advances in large language models (LLMs) have opened up new
+possibilities for conversational AI. LLMs are pretrained models that can be
+fine-tuned to perform a variety of tasks, including intent classification,
+dialogue handling, and natural language generation (NLG).
+
+### An adjustable risk profile
+
+The potential and risks of LLMs vary per use case, with more tolerance for error
+internally than for customer-facing applications. As your organization becomes
+more familiar with the technology, you may be more open to its uses, such as
+enhancing the NLU pipeline, managing long-tail topics, or subtly altering bot
+responses in low-risk situations.
+
+It's essential that your system provides full
+control over these processes. Understanding how LLMs and other components
+behave and have the power to override any decision.
+
+### Orchestration at the center
+
+A robust future-proof architecture leverages multiple LLMs for multiple
+jobs. And it puts the orchestration of components at the center, with the LLMs
+as exchangeable utilities / components in various places.
+
+<Image
+  img={llmUsageArchitecture}
+  caption="LLM Component usage in Rasa"
+  alt=""
+/>
+
+LLMs are a key component of a future-proof architecture. More importantly,
+we have the flexibility to use different LLMs for different purposes
+and can plug the models of our choice as the technology evolves.
+
+## Where to go from here
+
+This section of the documentation guides you through the diverse ways you can
+integrate LLMs into Rasa. We will delve into the following topics:
+
+1. [Setting up LLMs](./llm-setup.mdx)
+2. [Intent Classification with LLMs](./llm-intent.mdx)
+3. [Dialogue Handling using LLMs](./llm-intentless.mdx)
+4. [NLG using LLMs](./llm-nlg.mdx)
+5. [Custom usage of LLMs](./llm-custom.mdx)
+
+Each link will direct you to a detailed guide on the respective topic, offering
+further depth and information about using LLMs with Rasa. By the end of this
+series, you'll be equipped to effectively use LLMs to augment your Rasa
+applications.

docs/docs/llms/llm-custom.mdx

@@ -0,0 +1,13 @@
+---
+id: llm-custom
+sidebar_label: Customizing LLM Components
+title: Customizing LLM based Components
+abstract:
+---
+
+import RasaLabsLabel from "@theme/RasaLabsLabel";
+import RasaLabsBanner from "@theme/RasaLabsBanner";
+
+<RasaLabsLabel />
+
+<RasaLabsBanner />