Replace Hardcoded Values with Variables

Purpose

Systematically audit all QMD files for hardcoded numbers that should be replaced with {{< var variable_name >}} references from _variables.yml.

When to Use

• After batch adding new parameters to parameters.py
• Before major commits or releases
• When auditing specific chapters for variable consistency
• After noticing hardcoded values in a file

Process

Phase 1: Build Variable Lookup Map

First, load the parameter summary for quick lookups:

# Get all variable names and their display values
head -200 _analysis/parameter-summary.md

Key patterns to match:

• Currency: $14M, $929, $27.2B, $113.55B
• Percentages: 86.1%, 10%, 92% (skip 1% - treaty concept)
• Counts with units: 15,076 patients, 184.6M deaths, 4.83B DALYs
• Years: 1948, 2024 (skip years in citations @source-2024)

Phase 2: Find Hardcoded Values

For each QMD file in knowledge/ (excluding _build_temp/):

# Find currency values
grep -rn '\$[0-9,]\+\(\.[0-9]\+\)\?[KMB]\?' --include="*.qmd" knowledge/ | grep -v "_build_temp" | grep -v "{{< var"

# Find percentages (excluding 1%)
grep -rn '[0-9]\+\(\.[0-9]\+\)\?%' --include="*.qmd" knowledge/ | grep -v "_build_temp" | grep -v "{{< var" | grep -v "1%"

# Find large numbers with commas
grep -rn '[0-9]\{1,3\}\(,[0-9]\{3\}\)\+' --include="*.qmd" knowledge/ | grep -v "_build_temp" | grep -v "{{< var"

Phase 3: Match Against Variables

For each hardcoded value found:

1. Normalize the value: Remove commas, standardize format
2. Search parameter summary: Look for matching values

   grep -i "14M\|14,000,000\|14000000" _analysis/parameter-summary.md
   

3. Verify semantic match: Ensure variable meaning matches context
4. Document if no match: Create list for potential new parameters

Phase 4: Replace Values

For each confirmed match:

Old: The ADAPTABLE trial cost $14M and enrolled 15,076 patients
New: The ADAPTABLE trial cost {{< var adaptable_trial_total_cost >}} and enrolled {{< var adaptable_trial_patients >}}

Use the Edit tool to make replacements, preserving surrounding context.

Phase 5: Validation

After replacements:

# Verify all variable references are valid
npx tsx scripts/validate-variable-refs.ts

# Render to check for errors
quarto render knowledge/path/to/file.qmd --to html

Common Replacements Reference

Output Format

After completing audit, report:

## Hardcoded Value Audit Report

### Replacements Made
- file.qmd:123 - Replaced `$14M` with `{{< var adaptable_trial_total_cost >}}`
- file.qmd:456 - Replaced `92%` with `{{< var trial_exclusion_rate >}}`

### Values Without Variable Match
- file.qmd:789 - `$5.2B` - Context: "annual research budget"
  - Recommendation: Create ANNUAL_RESEARCH_BUDGET parameter

### Skipped (Intentional Hardcoding)
- file.qmd:101 - `1%` - Treaty percentage (conceptual, not a variable)
- file.qmd:202 - `2024` - Year in citation

Important Rules

1. NEVER replace 1% - This is the treaty concept, not a variable
2. Skip years in citations - @source-2024 should keep the year
3. Verify semantic context - A $14M budget is different from $14M trial cost
4. One file at a time - Process, validate, then move to next
5. Create new params if needed - If value should be a variable but doesn't exist, note it for parameter creation

Related Tools

• npm run review-hardcoded <file.qmd> - Generate detailed review for single file
• /validate-and-regenerate-parameters - After adding new parameters
• /qmd-consistency-check - Full book validation