Files
insights/bin/create-notebook-node.sh
zeekay 1b367e2b62 chore: rebase the tree onto upstream master
The tree was a debranded snapshot of upstream taken around 2026-02-20, and
upstream has moved about five months since. This replaces it with upstream
master, mechanically re-derived: bin/debrand does the rename, and the
pipeline and squashed migrations are restored from the shipped tree.

Not a merge. The previous import replaced the tree wholesale in a single
commit, so there is no shared content history to merge against. Reproducing
the rename on the new tree is what actually converges.

What survives: .hanzo/workflows, the Dockerfiles and bin/docker* entrypoints,
compose files, LLM.md, and insights/migrations. The squashed initial
migration is load-bearing — the live database records it by name, so
upstream's ~918 individual migrations cannot come back.

Nine products are dropped as unreferenced. Twenty-one more that we do not run
are kept anyway because the core app still imports them; unpicking those is a
refactor and does not belong in a rebase.

This base does not build yet, and the reasons are not in the rename:
upstream now wants Python 3.13.13 against our 3.12.12 and Django 5.2 against
our 4.2, and the dependency manifests still name upstream packages that we
republish under our own names. Those are decisions, not renames.
2026-08-01 15:44:40 -07:00

161 lines
5.6 KiB
Bash
Executable File

#!/bin/bash
# Script to create a new Insights notebook node
# Usage: ./bin/create-notebook-node.sh
set -e
# Colors for output
GREEN='\033[0;32m'
BLUE='\033[0;34m'
NC='\033[0m' # No Color
# Get the script's directory
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
PROJECT_ROOT="$(dirname "$SCRIPT_DIR")"
# Ask for node name
echo -e "${BLUE}Creating a new Insights notebook node${NC}"
echo ""
read -p "Enter the node name (PascalCase, e.g., 'RelatedGroups'): " NODE_NAME
if [ -z "$NODE_NAME" ]; then
echo "Error: Node name cannot be empty"
exit 1
fi
# Convert PascalCase to kebab-case for the node type
# e.g., RelatedGroups -> related-groups
NODE_TYPE=$(echo "$NODE_NAME" | sed 's/\([A-Z]\)/-\1/g' | sed 's/^-//' | tr '[:upper:]' '[:lower:]')
NODE_TYPE="ph-$NODE_TYPE"
# Ask for human-readable label
echo ""
read -p "Enter the human-readable label (e.g., 'Related groups'): " NODE_LABEL
if [ -z "$NODE_LABEL" ]; then
echo "Error: Node label cannot be empty"
exit 1
fi
echo ""
echo -e "${GREEN}Creating notebook node:${NC}"
echo " Node name: $NODE_NAME"
echo " Node type: $NODE_TYPE"
echo " Label: $NODE_LABEL"
echo ""
# Define file paths
NODES_DIR="$PROJECT_ROOT/frontend/src/scenes/notebooks/Nodes"
NODE_FILE="$NODES_DIR/NotebookNode$NODE_NAME.tsx"
EDITOR_FILE="$PROJECT_ROOT/frontend/src/scenes/notebooks/Notebook/Editor.tsx"
TYPES_FILE="$PROJECT_ROOT/frontend/src/scenes/notebooks/types.ts"
FILTER_FILE="$PROJECT_ROOT/frontend/src/scenes/notebooks/NotebooksTable/ContainsTypeFilter.tsx"
UTILS_FILE="$PROJECT_ROOT/frontend/src/scenes/notebooks/utils.ts"
# Step 1: Create the node component file
echo -e "${BLUE}1. Creating node component file...${NC}"
cat > "$NODE_FILE" << 'EOF'
import { useValues } from 'kea'
import { NotebookNodeProps, NotebookNodeType } from '../types'
import { createInsightsWidgetNode } from './NodeWrapper'
import { notebookNodeLogic } from './notebookNodeLogic'
const Component = ({ attributes }: NotebookNodeProps<NotebookNode__NODE_NAME__Attributes>): JSX.Element | null => {
const { expanded } = useValues(notebookNodeLogic)
if (!expanded) {
return null
}
return (
<div className="p-4">
{/* TODO: Implement component */}
<p>__NODE_NAME__ component - implement your UI here</p>
</div>
)
}
type NotebookNode__NODE_NAME__Attributes = {
// TODO: Add your attributes here
}
export const NotebookNode__NODE_NAME__ = createInsightsWidgetNode<NotebookNode__NODE_NAME__Attributes>({
nodeType: NotebookNodeType.__NODE_NAME__,
titlePlaceholder: '__TITLE_PLACEHOLDER__',
Component,
resizeable: false,
expandable: true,
startExpanded: true,
attributes: {
// TODO: Add your attribute definitions here
},
})
EOF
# Replace placeholders
sed -i '' "s/__NODE_NAME__/$NODE_NAME/g" "$NODE_FILE"
sed -i '' "s/__TITLE_PLACEHOLDER__/$NODE_LABEL/g" "$NODE_FILE"
echo " Created: $NODE_FILE"
# Step 2: Add import and extension to Editor.tsx
echo -e "${BLUE}2. Updating Editor.tsx...${NC}"
# Add import statement after the last NotebookNode import
IMPORT_LINE="import { NotebookNode$NODE_NAME } from '../Nodes/NotebookNode$NODE_NAME'"
LAST_IMPORT_LINE=$(grep -n "^import { NotebookNode" "$EDITOR_FILE" | tail -1 | cut -d: -f1)
# Use awk to insert the import line
awk -v line="$LAST_IMPORT_LINE" -v text="$IMPORT_LINE" 'NR==line{print; print text; next}1' "$EDITOR_FILE" > "$EDITOR_FILE.tmp" && mv "$EDITOR_FILE.tmp" "$EDITOR_FILE"
# Add the node to the extensions array (after the last node, before the closing bracket)
# Find the line with the closing bracket of the extensions array
EXTENSIONS_CLOSE=$(grep -n "^ \]$" "$EDITOR_FILE" | head -1 | cut -d: -f1)
# Use awk to insert the extension line before the closing bracket
awk -v line="$EXTENSIONS_CLOSE" -v text=" NotebookNode$NODE_NAME," 'NR==line{print text; print; next}1' "$EDITOR_FILE" > "$EDITOR_FILE.tmp" && mv "$EDITOR_FILE.tmp" "$EDITOR_FILE"
echo " Updated: $EDITOR_FILE"
# Step 3: Add to NotebookNodeType enum
echo -e "${BLUE}3. Updating NotebookNodeType enum...${NC}"
ENUM_LINE=$(grep -n "ZendeskTickets = 'ph-zendesk-tickets'," "$TYPES_FILE" | cut -d: -f1)
# Use awk to insert the enum entry
awk -v line="$ENUM_LINE" -v text=" $NODE_NAME = '$NODE_TYPE'," 'NR==line{print; print text; next}1' "$TYPES_FILE" > "$TYPES_FILE.tmp" && mv "$TYPES_FILE.tmp" "$TYPES_FILE"
echo " Updated: $TYPES_FILE"
# Step 4: Add to fromNodeTypeToLabel
echo -e "${BLUE}4. Updating ContainsTypeFilter.tsx...${NC}"
FILTER_LINE=$(grep -n "\[NotebookNodeType.ZendeskTickets\]: 'Zendesk tickets'," "$FILTER_FILE" | cut -d: -f1)
# Use awk to insert the filter entry
awk -v line="$FILTER_LINE" -v text=" [NotebookNodeType.$NODE_NAME]: '$NODE_LABEL'," 'NR==line{print; print text; next}1' "$FILTER_FILE" > "$FILTER_FILE.tmp" && mv "$FILTER_FILE.tmp" "$FILTER_FILE"
echo " Updated: $FILTER_FILE"
# Step 5: Add to customNodeTextSerializers
echo -e "${BLUE}5. Updating utils.ts...${NC}"
UTILS_LINE=$(grep -n "\[NotebookNodeType.ZendeskTickets\]: customOrTitleSerializer," "$UTILS_FILE" | cut -d: -f1)
# Use awk to insert the serializer entry
awk -v line="$UTILS_LINE" -v text=" [NotebookNodeType.$NODE_NAME]: customOrTitleSerializer," 'NR==line{print; print text; next}1' "$UTILS_FILE" > "$UTILS_FILE.tmp" && mv "$UTILS_FILE.tmp" "$UTILS_FILE"
echo " Updated: $UTILS_FILE"
echo ""
echo -e "${GREEN}✓ Successfully created notebook node: $NODE_NAME${NC}"
echo ""
echo "Next steps:"
echo " 1. Implement the component in: $NODE_FILE"
echo " 2. Define the attributes for your node"
echo " 3. Add any custom settings if needed"
echo " 4. Test your new node in a notebook"
echo ""