Build applications that make decisions (chatbots, agents, simulations, etc...). Monitor, trace, persist, and execute on your own infrastructure.
Python aiburrchatbot-frameworkdagsgenerative-aigraphshacktoberfestllmopsllmsmlopspersistent-data-structurestate-machinestate-managementvisibility
Adds input visualization
elijahbenizzyd46e0c6670f9
diff --git a/burr/tracking/common/models.py b/burr/tracking/common/models.pyindex c0bab6bf..8ff760a9 100644--- a/burr/tracking/common/models.py+++ b/burr/tracking/common/models.py@@ -1,6 +1,6 @@import datetimeimport inspect-from typing import Any, Dict, Optional, Union+from typing import Any, Dict, List, Optional, Unionfrom pydantic import field_serializer@@ -37,6 +37,7 @@ class ActionModel(IdentifyingModel):writes: list[str]code: strtype: str = "action"+ inputs: List[str] = pydantic.Field(default_factory=list)@staticmethoddef from_action(action: Action) -> "ActionModel":@@ -54,6 +55,7 @@ class ActionModel(IdentifyingModel):reads=list(action.reads),writes=list(action.writes),code=code,+ inputs=action.inputs,)diff --git a/telemetry/ui/src/api/models/ActionModel.ts b/telemetry/ui/src/api/models/ActionModel.tsindex 83c470dc..ecda2183 100644--- a/telemetry/ui/src/api/models/ActionModel.ts+++ b/telemetry/ui/src/api/models/ActionModel.ts@@ -6,10 +6,10 @@* Pydantic model that represents an action for storing/visualization in the UI*/export type ActionModel = {- type?: string;- name: string;- reads: Array<string>;- writes: Array<string>;- code: string;+ type?: string;+ name: string;+ reads: Array<string>;+ writes: Array<string>;+ code: string;+ inputs?: Array<string>;};-diff --git a/telemetry/ui/src/components/routes/app/GraphView.tsx b/telemetry/ui/src/components/routes/app/GraphView.tsxindex 8324911f..c7aa0391 100644--- a/telemetry/ui/src/components/routes/app/GraphView.tsx+++ b/telemetry/ui/src/components/routes/app/GraphView.tsx@@ -32,10 +32,18 @@ const elkOptions = {'org.eclipse.elk.layered.crossingMinimization.strategy': 'LAYER_SWEEP'};-type NodeData = {+type ActionNodeData = {action: ActionModel;label: string;};++type InputNodeData = {+ input: string;+ label: string;+};++type NodeData = ActionNodeData | InputNodeData;+type NodeType = {id: string;type: string;@@ -70,19 +78,21 @@ const ActionNode = (props: { data: NodeData }) => {currentAction} = React.useContext(NodeStateProvider);const highlightedActions = [currentAction, ...(previousActions || [])].reverse();+ const data = props.data as ActionNodeData;+ const name = data.action.name;const indexOfAction = highlightedActions.findIndex(- (step) => step?.step_start_log.action === props.data.action.name+ (step) => step?.step_start_log.action === data.action.name);const shouldHighlight = indexOfAction !== -1;const step = highlightedActions[indexOfAction];- const isCurrentAction = currentAction?.step_start_log.action === props.data.action.name;+ const isCurrentAction = currentAction?.step_start_log.action === name;const bgColor =isCurrentAction && step !== undefined? backgroundColorsForIndex(0, getActionStatus(step)): shouldHighlight? 'bg-gray-100': '';- const opacity = hoverAction?.step_start_log.action === props.data.action.name ? 'opacity-50' : '';+ const opacity = hoverAction?.step_start_log.action === name ? 'opacity-50' : '';const additionalClasses = isCurrentAction? 'border-dwlightblue/50 text-white border-2': shouldHighlight@@ -94,12 +104,24 @@ const ActionNode = (props: { data: NodeData }) => {<divclassName={`${bgColor} ${opacity} ${additionalClasses} text-xl font-sans p-4 rounded-md border`}>- {props.data.action.name}+ {name}</div><Handle type="source" position={Position.Bottom} id="a" /></>);};++const InputNode = (props: { data: NodeData }) => {+ return (+ <>+ <div className=" text-xl font-sans p-4 rounded-md bg-white bg-opacity-0">+ {props.data.label}+ </div>+ <Handle type="source" position={Position.Bottom} id="a" className="w-0" />+ </>+ );+};+// TODO -- separate out into different edge typesexport const ActionActionEdge = ({sourceX,sourceY,@@ -224,25 +246,47 @@ const getLayoutedElements = (};const convertApplicationToGraph = (stateMachine: ApplicationModel): [NodeType[], EdgeType[]] => {- return [- stateMachine.actions.map((action) => ({- id: action.name,- type: 'action',- data: { action, label: action.name },+ const inputUniqueID = (action: ActionModel, input: string) => `${action.name}:${input}`; // Currently they're distinct by name++ const allActionNodes = stateMachine.actions.map((action) => ({+ id: action.name,+ type: 'action',+ data: { action, label: action.name },+ position: { x: 0, y: 0 }+ }));+ const allInputNodes = stateMachine.actions.flatMap((action) =>+ (action.inputs || []).map((input) => ({+ id: inputUniqueID(action, input),+ type: 'externalInput',+ data: { input, label: input },position: { x: 0, y: 0 }- })),- stateMachine.transitions.map((transition) => ({- id: `${transition.from_}-${transition.to}-${transition.condition}`,- source: transition.from_,- target: transition.to,+ }))+ );+ const allInputTransitions = stateMachine.actions.flatMap((action) =>+ (action.inputs || []).map((input) => ({+ id: `${action.name}:${input}-${action.name}`,+ source: inputUniqueID(action, input),+ target: action.name,markerEnd: { type: MarkerType.ArrowClosed, width: 20, height: 20 },- data: { from: transition.from_, to: transition.to, condition: transition.condition }+ data: { from: inputUniqueID(action, input), condition: input, to: action.name }}))+ );+ const allTransitionEdges = stateMachine.transitions.map((transition) => ({+ id: `${transition.from_}-${transition.to}`,+ source: transition.from_,+ target: transition.to,+ markerEnd: { type: MarkerType.ArrowClosed, width: 20, height: 20 },+ data: { from: transition.from_, to: transition.to, condition: transition.condition }+ }));+ return [+ [...allActionNodes, ...allInputNodes],+ [...allInputTransitions, ...allTransitionEdges]];};const nodeTypes = {- action: ActionNode+ action: ActionNode,+ externalInput: InputNode // if this is "input" it is reserved...};const edgeTypes = {