The HTML report

cf.to_html("report.html") writes a single self-contained HTML file (no server, no external assets) documenting one pipeline run. It is the primary per-run artifact.

import conformare as cf

cf.trackSpark()
cf.set_profiles({"*": [cf.rowCount], "filter": [cf.histogram(columns="all")]})
# ... your pipeline ...
cf.to_html("report.html", title="Customer scoring")

What’s in it

• Overview — KPI cards (nodes, operations, columns, sources, sinks, sensitive columns, risks, expectations) and the describe_process() description.
• Process diagram — the lineage graph, with a settings menu:
  • Layout: Sequential (the default — disconnected chains stacked by execution order), Grouped, Layered, Swimlane.
  • Compress chained operations — roll a run of steps on one dataframe into a single expandable node.
  • Compress context-linked nodes — contract a describe() group into one node.
  • Show function calls — annotate nodes with the functions they pass through.
  • Expand operation details — show each operation’s captured expression inline (and, for collapsed chains / contracted contexts, every folded step plus the functions used).
  • Edge style, zoom, light/dark.
• Column index — every column and the nodes it appears in (long lists collapse to the first few with “… and N more”).
• Data sensitivity — flagged columns, their tags, and whether they reach a written output.
• Risk register — one row per risk: severity, mitigation, owner, and where it occurs.
• Context register — the describe() contexts, their purpose, owner and risks.
• Created columns — a catalog of created columns and a diagram showing every existing column that contributes to each one.
• Node profiles — per-node detail (rows, size, columns, distributions, expectations), collapsible by context / function / chain, with a “Hide non-profiled” toggle (on by default) so the section focuses on what was actually measured.

Other per-run exports

Function	Output
cf.to_html(path)	the interactive report above
cf.to_json(path)	the full model (build_model) as JSON, for your own tooling
cf.to_mermaid()	a Mermaid lineage diagram (embed in Markdown / wikis)
cf.to_risk_checklist(path)	a formal, sign-off-ready Markdown risk checklist
cf.build_model(cf.store)	the in-memory model dict the report renders from

These describe a single run. To track results across runs, record each run to a fleet store — see Fleet recording.