Streamlit_Backend_Core.md

graph LR
    Runtime["Runtime"]
    AppSession["AppSession"]
    ScriptRunner["ScriptRunner"]
    DeltaGenerator["DeltaGenerator"]
    ForwardMsgQueue["ForwardMsgQueue"]
    SessionState["SessionState"]
    Caching["Caching"]
    Secrets["Secrets"]
    Runtime -- "manages" --> AppSession
    Runtime -- "routes messages to" --> AppSession
    Runtime -- "flushes messages from" --> ForwardMsgQueue
    AppSession -- "initiates" --> ScriptRunner
    AppSession -- "enqueues messages to" --> ForwardMsgQueue
    AppSession -- "accesses" --> Secrets
    ScriptRunner -- "signals to" --> AppSession
    ScriptRunner -- "interacts with" --> DeltaGenerator
    ScriptRunner -- "manages state via" --> SessionState
    ScriptRunner -- "utilizes" --> Caching
    DeltaGenerator -- "enqueues deltas to" --> ForwardMsgQueue

Loading

Details

The Streamlit backend runtime is orchestrated by the Runtime component, which manages user sessions and message flow. Each user interaction is handled by an AppSession, responsible for initiating script execution via the ScriptRunner. The ScriptRunner executes the user's Python code, interacting with DeltaGenerator to produce UI updates, managing application state through SessionState, and optimizing performance with Caching. UI updates and other messages are buffered in the ForwardMsgQueue before being sent to the frontend. AppSession also securely accesses sensitive information through the Secrets component. This architecture ensures a clear separation of concerns, enabling efficient and responsive Streamlit application execution.

Runtime

The central orchestrator of the Streamlit backend. It manages the main event loop, handles incoming connections from the frontend, and coordinates interactions between AppSession instances and the ForwardMsgQueue.

Related Classes/Methods:

• streamlit.runtime.runtime

AppSession

Represents a single user's interactive session with a Streamlit application. It manages the session lifecycle, processes messages received from the frontend, initiates script execution, and handles file change detection to trigger reruns.

Related Classes/Methods:

• streamlit.runtime.app_session

ScriptRunner

Executes the user's Streamlit Python script in a dedicated thread. It controls the script's lifecycle, manages execution flow (reruns, stops), and interacts with other runtime components to update the UI and manage state.

Related Classes/Methods:

• streamlit.runtime.scriptrunner.script_runner

DeltaGenerator

Provides the core st.* API functions that users call to generate UI elements and updates (deltas). It manages the current position within the element tree and enqueues UI messages for transmission.

Related Classes/Methods:

• streamlit.delta_generator

ForwardMsgQueue

Buffers and manages ForwardMsgs (UI updates, session state changes, etc.) that need to be sent from the backend to the frontend. It optimizes message delivery by composing multiple deltas into single messages where possible.

Related Classes/Methods:

• streamlit.runtime.forward_msg_queue

SessionState

Manages the state of widgets and user-defined session variables across script reruns. It provides a dictionary-like interface for accessing and modifying this state.

Related Classes/Methods:

• streamlit.runtime.state.session_state

Caching

Implements Streamlit's caching mechanisms (st.cache_data and st.cache_resource) to optimize script execution by storing and reusing computation results and resources.

Related Classes/Methods:

• streamlit.runtime.caching

Secrets

Manages access to user-defined secrets, typically loaded from secrets.toml files and environment variables, providing a secure way to access sensitive information within the application.

Related Classes/Methods:

• streamlit.runtime.secrets

FAQ