Survival Curve Risk Table Generator

Automatically add "Number at risk" tables to Kaplan-Meier survival curves. Aligns time points precisely with the curve X-axis and generates publication-quality combined figures for clinical oncology journals.

Quick Check

python -m py_compile scripts/main.py
python scripts/main.py --help
python scripts/main.py --input survival_data.csv --time-col time --event-col event --group-col treatment --output risk_table.png

When to Use

• Add "Number at risk" tables to existing KM survival curves
• Generate survival plots meeting NEJM, Lancet, or JCO requirements
• Produce risk tables for clinical trial reports
• Standardize medical paper charts before journal submission

Workflow

1. Confirm input CSV path, time column, event column, and optional group column.
2. Validate that the request involves KM survival curve risk table generation; stop early if not.
3. Run scripts/main.py with the appropriate style and time-point flags.
4. Return a structured result separating assumptions, output files, and unresolved items.
5. If execution fails or inputs are incomplete, switch to the Fallback Template below.

Group count check: If more than 6 groups are provided, emit a warning, split into pages of 6 groups each, and name paginated output files as risk_table_page1.png, risk_table_page2.png, etc.

Fallback Template

If scripts/main.py fails or required fields are missing, respond with:

FALLBACK REPORT
───────────────────────────────────────
Objective        : <risk table generation goal>
Inputs Available : <file, columns, style provided>
Missing Inputs   : <list exactly what is missing>
Partial Result   : <any steps completed safely>
Blocked Steps    : <what could not be completed and why>
  Note: If lifelines is not installed, survival analysis features are limited.
        Install with: pip install lifelines
Next Steps       : <minimum info needed to complete>
───────────────────────────────────────

Stress-Case Output Checklist

For complex multi-constraint requests, always include these sections explicitly:

• Assumptions: style defaults (NEJM), time points auto-calculated, DPI=300; always state the journal style applied, noting if it differs from the NEJM default
• Constraints: max 6 groups per page; >6 groups triggers pagination (see Group count check above)
• Risks: lifelines optional but required for full survival analysis
• Unresolved Items: missing columns, unsupported file formats

CLI Usage

# Basic risk table
python scripts/main.py \
  --input survival_data.csv \
  --time-col time --event-col event --group-col treatment \
  --output risk_table.png

# Journal style
python scripts/main.py \
  --input survival_data.csv \
  --time-col time --event-col status --group-col arm \
  --style NEJM --time-points 0,6,12,18,24,30,36 \
  --output figure_1a.pdf

# Combined figure (KM curve + risk table)
python scripts/main.py \
  --input survival_data.csv \
  --time-col months --event-col death --group-col group \
  --km-plot km_curve.png --combine \
  --output combined_figure.png

Required Input Columns

Supported file formats: CSV, Excel (.xlsx/.xls), SAS (.sas7bdat), RData (.rda/.rds), pickle (.pkl)

Key Parameters

→ Full parameter and journal style reference: references/parameters.md

Input Validation

This skill accepts: CSV or tabular survival data files with time-to-event and event indicator columns, for generating Kaplan-Meier risk tables.

If the request does not involve KM survival curve risk table generation — for example, asking to perform Cox regression, compute hazard ratios, or analyze non-survival time-series data — do not proceed. Instead respond:

"survival-curve-risk-table is designed to generate 'Number at risk' tables for Kaplan-Meier survival curves. Your request appears to be outside this scope. Please provide a survival data CSV with time and event columns, or use a more appropriate tool."

Error Handling

• If required inputs are missing, state exactly which fields are missing and request only the minimum additional information.
• If lifelines is not installed, report the warning and provide the install command: pip install lifelines.
• If the task goes outside documented scope, stop instead of guessing.
• If scripts/main.py fails, use the Fallback Template above.
• Do not fabricate survival statistics, risk counts, or execution outcomes.

Output Requirements

Every final response must include:

1. Objective — what risk table was generated and for which journal style
2. Inputs Received — file path, column names, style, time points
3. Assumptions — defaults applied (style, DPI, time points); always state the journal style applied, noting if it differs from NEJM default; if no group column, label as "All patients"
4. Result — output file paths generated
5. Risks and Limits — lifelines dependency, group count limits, pagination if >6 groups
6. Next Checks — manually spot-check number-at-risk calculations

Dependencies

numpy >= 1.20.0
pandas >= 1.3.0
matplotlib >= 3.4.0
seaborn >= 0.11.0
lifelines >= 0.27.0   # optional but recommended
Pillow >= 8.0.0