create-recipe

name	create-recipe
description	Create new growth media or solution YAML records from input data or documents
category	workflow
requires_database	false
requires_internet	false
version	1.0.0

Create Recipe Skill

Overview

Purpose: Generate properly formatted CultureMech YAML records for growth media or solutions from various input formats (text descriptions, papers, protocols, structured data).

Why: Streamlines recipe creation, ensures schema compliance, maintains data quality, and provides proper ID assignment.

Scope: Creates new MediaRecipe or Solution YAML files with full validation and proper placement.

When to Use This Skill

Use this skill when you need to:

• Create a new growth medium YAML record from a paper or protocol
• Convert a text recipe description into structured YAML
• Import recipes from external sources (papers, databases, lab notes)
• Generate solution records for stock solutions
• Create test/example recipes for development
• Batch import recipes from documents

Input Formats Supported

1. Text Description

"LB Broth: Mix 10 g/L tryptone, 5 g/L yeast extract, 10 g/L NaCl in water.
Autoclave at 121°C for 15 min. pH 7.0."

2. Structured Data (JSON)

{
  "name": "LB Broth",
  "ingredients": [
    {"name": "Tryptone", "concentration": "10 g/L"},
    {"name": "Yeast extract", "concentration": "5 g/L"},
    {"name": "NaCl", "concentration": "10 g/L"}
  ],
  "ph": 7.0,
  "sterilization": "121°C, 15 min"
}

3. PDF/Document

• Research papers with media recipes
• Lab protocols
• Supplier specifications
• Culture collection datasheets

4. Existing Recipe (for modification)

• Path to existing YAML file
• Recipe ID to use as template

Workflow

Step 1: Analyze Input

Your Task: Understand the input format and extract recipe information

Actions:

1. Read the input (file, text, JSON, etc.)
2. Identify recipe type (medium vs solution)
3. Extract key components:
   • Recipe name
   • Ingredients with concentrations
   • pH, temperature, preparation steps
   • Source/reference information
   • Target organisms (if mentioned)

Example:

Input: "M9 minimal medium: Na2HPO4 (6 g/L), KH2PO4 (3 g/L), NaCl (0.5 g/L),
NH4Cl (1 g/L), glucose (4 g/L). Autoclave base salts, add sterile glucose."

Extracted:
- Name: M9 minimal medium
- Type: Defined/Minimal medium
- Ingredients: 5 components with concentrations
- Preparation: Autoclave base, add glucose separately

Step 2: Generate YAML Structure

Your Task: Create a valid CultureMech YAML record

Required Fields:

• name - Recipe name
• medium_type - DEFINED, COMPLEX, SEMI_DEFINED, etc.
• physical_state - LIQUID, SOLID, SEMI_SOLID
• ingredients - List with preferred_term and concentration

Template:

name: Recipe Name
original_name: Recipe Name
category: bacterial  # or algae, archaea, fungal, specialized
medium_type: DEFINED
physical_state: LIQUID

ingredients:
  - preferred_term: Ingredient 1
    concentration: 10 G_PER_L
  - preferred_term: Ingredient 2
    concentration: 5 G_PER_L

ph_value: 7.0

sterilization:
  method: AUTOCLAVE
  temperature: 121
  duration: 15
  notes: Standard autoclave cycle

preparation_steps:
  - action: DISSOLVE
    description: Dissolve all ingredients in distilled water
  - action: AUTOCLAVE
    description: Sterilize at 121°C for 15 minutes

notes: Additional preparation notes

curation_history:
  - timestamp: CURRENT_TIME
    curator: create-recipe-skill
    action: Created new recipe from input

Schema Validation: Always validate against src/culturemech/schema/culturemech.yaml

Step 3: Assign CultureMech ID

Your Task: Get next available CultureMech ID

Actions:

1. Use manage-identifiers skill to find highest ID
2. Mint next sequential ID
3. Add to YAML: id: CultureMech:NNNNNN

Example:

# Find highest ID
python scripts/find_highest_id.py --prefix CultureMech

# Output: CultureMech:015431
# Next ID: CultureMech:015432

Step 4: Determine File Location

Your Task: Choose correct category directory

Category Mapping:

• bacterial/ - Bacterial growth media
• algae/ - Algae/cyanobacteria media
• archaea/ - Archaeal media
• fungal/ - Fungal/yeast media
• specialized/ - Cross-kingdom or specialized
• solutions/ - Stock solutions (not complete media)

Filename Format:

{sanitized_name}.yaml

Example: "LB Broth" → "LB_Broth.yaml"

Full Path:

data/normalized_yaml/{category}/{sanitized_name}.yaml

Step 5: Validate and Save

Your Task: Validate schema and write file

Actions:

1. Validate YAML against schema
2. Check for duplicate names
3. Write file to correct location
4. Regenerate indexes

Validation:

# Validate single file
just validate-schema data/normalized_yaml/bacterial/LB_Broth.yaml

# Validate all
just validate-recipes

Complete:

# Regenerate indexes
just generate-indexes

Output Format

Success Output

# File: data/normalized_yaml/bacterial/LB_Broth.yaml
id: CultureMech:015432
name: LB Broth
original_name: LB Broth
category: bacterial
medium_type: COMPLEX
physical_state: LIQUID

ingredients:
  - preferred_term: Tryptone
    concentration: 10 G_PER_L
  - preferred_term: Yeast extract
    concentration: 5 G_PER_L
  - preferred_term: Sodium chloride
    concentration: 10 G_PER_L
  - preferred_term: Water
    concentration: 1000 G_PER_L

ph_value: 7.0

sterilization:
  method: AUTOCLAVE
  temperature: 121
  duration: 15
  temperature_unit: CELSIUS
  duration_unit: MINUTE

preparation_steps:
  - action: DISSOLVE
    description: Dissolve all ingredients in distilled water
  - action: ADJUST_PH
    description: Adjust pH to 7.0 if necessary
  - action: AUTOCLAVE
    description: Sterilize at 121°C for 15 minutes

curation_history:
  - timestamp: 2026-03-15T04:30:00.000000+00:00
    curator: create-recipe-skill
    action: Created new recipe from text input
    notes: Generated from user-provided recipe description

Summary Report

✅ Recipe Created Successfully

**File**: data/normalized_yaml/bacterial/LB_Broth.yaml
**ID**: CultureMech:015432
**Name**: LB Broth
**Category**: bacterial
**Ingredients**: 4 components
**Validation**: ✅ Schema valid

**Next Steps**:
1. Review the generated YAML file
2. Add source reference if available
3. Enrich with MediaIngredientMech (if desired)
4. Run quality pipeline: `just fix-all-data-quality`

Common Patterns

Pattern 1: From Paper

Input: PDF with recipe in methods section

Process:

1. Extract text from PDF
2. Parse ingredients and concentrations
3. Generate YAML with source citation
4. Add to references field

Example:

references:
  - citation: "Smith et al. (2025). Journal of Microbiology."
    doi: "10.1234/jmicro.2025.001"
    notes: "Recipe described in Materials & Methods, page 3"

Pattern 2: From Culture Collection

Input: ATCC/DSMZ medium specification

Process:

1. Extract from datasheet
2. Add media_term with source ID
3. Link to organism if specified

Example:

media_term:
  preferred_term: ATCC Medium 1
  term:
    id: atcc.medium:1
    name: ATCC Medium 1

target_organisms:
  - preferred_term: Escherichia coli

Pattern 3: Stock Solution

Input: "10× PBS: 80 g NaCl, 2 g KCl, 14.4 g Na2HPO4, 2.4 g KH2PO4 per liter"

Process:

1. Identify as solution (not complete medium)
2. Save to solutions/ directory
3. Mark as stock solution

Output Location: data/normalized_yaml/solutions/10x_PBS.yaml

Pattern 4: Batch Import

Input: CSV file with multiple recipes

Process:

1. Read CSV rows
2. Generate one YAML per row
3. Assign sequential IDs
4. Validate all files
5. Generate batch report

Validation Checklist

Before saving, verify:

• ✅ Valid YAML syntax
• ✅ Schema compliance
• ✅ Required fields present (name, medium_type, physical_state, ingredients)
• ✅ CultureMech ID assigned and unique
• ✅ Correct category directory
• ✅ No duplicate names in category
• ✅ Concentration units valid
• ✅ Enum values valid (medium_type, physical_state, etc.)
• ✅ Curation history entry added

Error Handling

Common Issues

Issue: Missing ingredient concentrations Solution: Mark as approximate or add data quality flag

Issue: Unclear medium type Solution: Use COMPLEX as default, add note

Issue: Multiple recipes in input Solution: Create separate files for each

Issue: Incomplete information Solution: Create with data_quality_flags and notes

Integration with Pipeline

After creating recipe:

# 1. Validate
just validate-schema data/normalized_yaml/bacterial/New_Recipe.yaml

# 2. Run quality pipeline
just fix-all-data-quality

# 3. Enrich (optional)
# Run MediaIngredientMech enrichment if desired

# 4. Regenerate indexes
just generate-indexes

# 5. Commit
git add data/normalized_yaml/bacterial/New_Recipe.yaml
git commit -m "Add New_Recipe medium"

Examples

Example 1: Simple Recipe from Text

Input:

Create a recipe for TSB (Tryptic Soy Broth):
- Tryptone: 17 g/L
- Soy peptone: 3 g/L
- NaCl: 5 g/L
- K2HPO4: 2.5 g/L
- Glucose: 2.5 g/L
pH 7.3, autoclave 121°C for 15 min

Output: data/normalized_yaml/bacterial/TSB.yaml with CultureMech ID

Example 2: From JSON

Input:

{
  "name": "Nutrient Agar",
  "type": "complex",
  "state": "solid",
  "ingredients": [
    {"name": "Peptone", "amount": "5 g/L"},
    {"name": "Beef extract", "amount": "3 g/L"},
    {"name": "Agar", "amount": "15 g/L"}
  ]
}

Output: Validated YAML with proper enums and structure

Example 3: Solution

Input: "Create 1 M Tris-HCl pH 8.0 stock solution"

Output: data/normalized_yaml/solutions/1M_Tris_HCl_pH8.yaml

Tips for Best Results

1. Provide Complete Information: More details = better YAML
2. Include Source: Always cite where recipe came from
3. Specify Units: Clear concentration units help parsing
4. Note Variations: Document any modifications from original
5. Validate Early: Check schema before committing
6. Use Templates: Start from similar recipes when possible

Related Skills

• manage-identifiers - ID assignment and management
• manage-ingredient-hierarchy - MediaIngredientMech integration

Script Support

Helper scripts available:

• scripts/find_highest_id.py - Get next CultureMech ID
• scripts/cleanup_recipe_ingredients.py - Clean duplicates
• scripts/generate_recipe_indexes.py - Regenerate indexes

Quick Reference

# Full workflow
1. Parse input → Extract recipe data
2. Generate YAML → Validate against schema
3. Assign ID → Use manage-identifiers skill
4. Save file → data/normalized_yaml/{category}/{name}.yaml
5. Validate → just validate-schema {file}
6. Update indexes → just generate-indexes

---

Remember: Always validate against the schema and regenerate indexes after creating new recipes!