Dashboard Reference

models

Dashboard models shared by the Streamlit app and service layer.

DashboardActivityEvent

Bases: BaseModel

A normalized live workflow event shown in the dashboard activity panel.

display_line property

display_line

Return a compact human-facing timeline line.

DashboardCitation

Bases: BaseModel

Normalized citation rendered by the dashboard.

DashboardRunRequest

Bases: BaseModel

Input payload for one dashboard workflow run.

DashboardRunResult

Bases: BaseModel

Normalized dashboard result for any workflow.

evidence_count property

evidence_count

Return the normalized evidence count if present.

key_findings property

key_findings

Return normalized key findings for the current result.

primary_summary property

primary_summary

Return the best available one-line summary for thread displays.

unresolved_questions property

unresolved_questions

Return normalized unresolved questions for the current result.

DashboardThreadRecord

Bases: BaseModel

Thread summary used by the dashboard sidebar.

display_label property

display_label

Return the compact label used in thread selectors.

clear

clear()

Return a cleared thread record that keeps the thread identifier.

create classmethod

create(workflow, *, thread_id=None)

Create a new dashboard thread for the given workflow.

record_turn

record_turn(*, question, result)

Return an updated thread record after one completed turn.

DashboardTurnRecord

Bases: BaseModel

One persisted dashboard turn stored in Streamlit session state.

SearchWorkflow

Bases: StrEnum

Supported research workflows exposed in the dashboard.

description property

description

Return a short workflow description.

graph_mermaid property

graph_mermaid

Return a Mermaid diagram describing the workflow.

ideal_for property

ideal_for

Return the best-fit usage for the workflow.

input_placeholder property

input_placeholder

Return the dashboard input placeholder for the workflow.

label property

label

Return a human-readable workflow label.

stages property

stages

Return the major execution stages for the workflow.

starter_questions property

starter_questions

Return workflow-specific starter prompts.

presentation

Presentation helpers for the Streamlit dashboard.

build_mermaid_embed

build_mermaid_embed(mermaid_text, *, title, subtitle)

Return embeddable HTML for an interactive Mermaid workflow panel.

build_mermaid_iframe_src

build_mermaid_iframe_src(mermaid_text, *, title, subtitle)

Return a data URI that Streamlit can render via st.iframe.

format_thread_label

format_thread_label(thread)

Return the user-facing label for a thread selector option.

service

Dashboard service that drives live workflow runs and persistent dashboard state.

DashboardPersistenceError

Bases: RuntimeError

Raised when Postgres-backed dashboard persistence is unavailable.

DashboardService

Run the repository workflows behind a dashboard-friendly interface.

default_model_for_workflow

default_model_for_workflow(workflow)

Return the configured default model for a workflow.

list_persistent_threads async

list_persistent_threads(workflow)

Return dashboard thread metadata stored in Postgres for one workflow.

load_persistent_history async

load_persistent_history(workflow, thread_id)

Return the stored dashboard turn history for one thread.

load_persistent_run_state async

load_persistent_run_state(workflow, *, thread_id, timezone_name='America/Toronto', debug=False)

Load the latest persisted LangGraph state for one workflow thread.

run async

run(request, *, on_event=None)

Run one dashboard request and normalize the response.

save_persistent_thread async

save_persistent_thread(workflow, thread, turns)

Persist a thread record and its turn history to Postgres.

launcher

Launcher for the packaged Streamlit dashboard.

build_streamlit_command

build_streamlit_command(*, host='127.0.0.1', port=8501, headless=False)

Build the subprocess command used to launch the dashboard.

launch_dashboard

launch_dashboard(*, host='127.0.0.1', port=8501, headless=False)

Launch the Streamlit dashboard and return the process exit code.

main

main(argv=None)

Parse launcher args and run the packaged dashboard.