GRAPH

Example: Graph Generation Workflow

# Step 1: Load corpus with documents that have keywords
crispt --inp crisp_input --graph --out crisp_input

# Step 2: Visualize the graph (all node types)
crispviz --inp crisp_input --out visualizations --graph

# Step 3: Visualize only documents and keywords
crispviz --inp crisp_input --out visualizations --graph --graph-nodes document,keyword

# Step 4: Try different graph layouts
crispviz --inp crisp_input --out visualizations --graph --graph-layout circular

About --graph-nodes:

The --graph-nodes option allows you to filter which node types are included in the graph visualization. For example, to show only documents and keywords, use:

crispviz --inp crisp_input --out visualizations --graph --graph-nodes document,keyword

Valid node types: document, keyword, cluster, metadata. If omitted or set to all, all node types are included. Edges are only shown if both endpoints are present in the filtered node set.

The graph visualization shows: - Documents (red nodes): Your corpus documents - Keywords (teal nodes): Keywords extracted from documents - Clusters (light green nodes): Document clusters (if clustering analysis was performed) - Metadata (yellow nodes): Metadata from DataFrame (if present with aligning ID field)

Note: If documents don't have keywords assigned, run keyword assignment first using text analysis features before generating the graph.