Configure synthea-alldiseases for WES execution

- Rewrite params.json to match WES tool registry format
- Update main.nf to use Harbor container image
- Add k8s profile to nextflow.config for WES/Kubernetes execution
- Use s3://omic/eureka paths for output
This commit is contained in:
2026-03-25 12:31:34 +01:00
parent e10ae0cf81
commit d468509ec3
3 changed files with 195 additions and 160 deletions

146
main.nf
View File

@@ -2,127 +2,69 @@
nextflow.enable.dsl=2
/*
* Synthea Disease Module Generator Pipeline
*
* A Nextflow pipeline to generate and manage Synthea disease modules
*/
// Load API key from .env file if it exists
def envFile = file('.env')
if (envFile.exists()) {
envFile.eachLine { line ->
def (key, value) = line.tokenize('=')
if (key && value && key.trim() == 'ANTHROPIC_API_KEY') {
params.anthropic_api_key = value.trim()
}
}
}
// Default parameters
params.disease_name = null // Disease name to generate patients for
params.output_dir = "/mnt/OmicNAS/private/old/olamide/synthea/output/new" // Output directory
params.modules_dir = "src/main/resources/modules" // Directory for module files
params.population = 100 // Number of patients to generate
params.gender = 0.5 // Decimal representing proportion female (0.0-1.0)
params.min_age = 0 // Minimum age of generated patients
params.max_age = 90 // Maximum age of generated patients
params.seed = null // Random seed for reproducibility
params.help = false // Show help message
// Show help message
if (params.help) {
log.info """
Synthea Patient Generator
========================
Usage: nextflow run main.nf --disease_name "Disease Name"
Required Arguments:
--disease_name Disease name to generate patients for
Optional Arguments:
--modules_dir Module directory (default: modules)
--output_dir Output directory (default: output)
--population Number of patients (default: 100)
--gender Gender ratio - female proportion 0.0-1.0 (default: 0.5)
--min_age Minimum age (default: 0)
--max_age Maximum age (default: 90)
--seed Random seed (default: random)
"""
exit 0
}
params.disease_name = null
params.outdir = null
params.population = 10
params.gender = 0.5
params.min_age = 0
params.max_age = 90
params.seed = null
// Validate required parameters
if (!params.disease_name && !params.help) {
if (!params.disease_name) {
error "Disease name is required. Please specify with --disease_name"
}
// Process to check if module exists and generate it if needed
process checkAndGetModule {
container 'synthea-module-generator'
publishDir "${params.modules_dir}", mode: 'copy'
input:
val diseaseName
output:
path "*.json", emit: module_file
script:
def moduleFilename = diseaseName.toLowerCase().replaceAll(' ', '_') + '.json'
def fullPath = "/app/src/main/resources/modules/${moduleFilename}"
"""
echo "Looking for module at ${fullPath}"
if [ -f "${fullPath}" ]; then
echo "Module exists, copying..."
cp "${fullPath}" .
else
echo "Module not found, generating..."
python3 /app/module_generator/module_generator.py --disease "${diseaseName}" --output "${moduleFilename}"
if [ -f "${moduleFilename}" ]; then
echo "Successfully generated module"
else
echo "Error: Failed to generate module"
exit 1
fi
fi
"""
if (!params.outdir) {
error "Output directory is required. Please specify with --outdir"
}
// Process to generate synthetic patients
process generatePatients {
container 'synthea-module-generator'
publishDir "${params.output_dir}/${diseaseName.toLowerCase().replaceAll(' ', '_')}", mode: 'copy'
container 'harbor.cluster.omic.ai/omic/synthea-alldiseases:latest'
publishDir params.outdir, mode: 'copy'
input:
val diseaseName
path moduleFile
output:
path "m/*", optional: true
path "f/*", optional: true
path "fhir/*.json", optional: true, emit: fhir_output
path "run.log", emit: log_file
script:
def moduleBasename = diseaseName.toLowerCase().replaceAll(' ', '_')
def genderArg = params.gender < 0.5 ? "M" : (params.gender > 0.5 ? "F" : "B")
def seedValue = params.seed ?: new Random().nextInt(1000000)
def seedArg = params.seed ? "-s ${params.seed}" : ""
"""
# Copy module and run Synthea
cp "${moduleFile}" /app/modules/
cd /app && ./run_synthea -p ${params.population} -g ${genderArg} -m ${moduleBasename} -a ${params.min_age}-${params.max_age} -s ${seedValue}
# Organize output by gender
mkdir -p m f
find /app/output/fhir -type f -name "*.json" ! -name "*hospital*" ! -name "*practitioner*" | xargs grep -l '"gender":"male"' | xargs -I{} cp {} m/
find /app/output/fhir -type f -name "*.json" ! -name "*hospital*" ! -name "*practitioner*" | xargs grep -l '"gender":"female"' | xargs -I{} cp {} f/
# Check if a custom module exists, otherwise use built-in Synthea modules
MODULE_FILE="/app/src/main/resources/modules/${moduleBasename}.json"
if [ -f "\${MODULE_FILE}" ]; then
echo "Found custom module: \${MODULE_FILE}" | tee run.log
else
echo "Using built-in Synthea modules for: ${diseaseName}" | tee run.log
fi
# Run Synthea patient generation
cd /app && ./run_synthea \
-p ${params.population} \
-g ${genderArg} \
-a ${params.min_age}-${params.max_age} \
${seedArg} \
-- ${diseaseName} 2>&1 | tee -a run.log
# Collect FHIR output
mkdir -p fhir
if [ -d /app/output/fhir ]; then
cp /app/output/fhir/*.json fhir/ 2>/dev/null || true
echo "Copied \$(ls fhir/*.json 2>/dev/null | wc -l) FHIR bundles" | tee -a run.log
else
echo "Warning: No FHIR output generated" | tee -a run.log
fi
"""
}
// Define workflow
// Workflow
workflow {
// First check if the module exists
checkAndGetModule(params.disease_name)
// Then generate patients
generatePatients(params.disease_name, checkAndGetModule.out.module_file)
generatePatients(params.disease_name)
}

View File

@@ -1,54 +1,44 @@
// Nextflow configuration file
manifest {
description = 'Synthea Module Generator Pipeline'
name = 'synthea-alldiseases'
description = 'Synthea synthetic patient generator pipeline'
mainScript = 'main.nf'
version = '1.0.0'
}
// Load parameters from params.json
def paramsJson = new File("$baseDir/params.json").text
def paramsData = new groovy.json.JsonSlurper().parseText(paramsJson)
// Keep work directories for debugging
cleanup = false
// Merge with defaults
params {
local_output_dir = "$baseDir/local_output"
disease_name = paramsData.disease_name
modules_dir = paramsData.modules_dir ?: "/Users/richman/workspace/synthea-alldiseases/modules"
output_dir = paramsData.output_dir ?: "/mnt/OmicNAS/private/old/olamide/synthea/output/new"
population = paramsData.population ?: 100
gender = paramsData.gender ?: 0.5
min_age = paramsData.min_age ?: 0
max_age = paramsData.max_age ?: 90
seed = paramsData.seed
generate_patients = paramsData.generate_patients ?: true
publish_dir = paramsData.publish_dir ?: "published_output"
// Additional params
max_cost = 5.0
timeout = 300
anthropic_api_key = null
batch_size = 1
help = false
}
docker {
enabled = true
runOptions = "-v $baseDir/$params.modules_dir:/app/src/main/resources/modules -v $baseDir/src/main/python:/app/src/main/python -v $baseDir/src/main/resources:/app/src/main/resources -v $baseDir/.env:/app/.env -v $baseDir/module_generator:/app/module_generator -v $baseDir/modules:/app/modules"
}
process {
container = 'synthea-module-generator'
containerOptions = "-e MODULES_DIR=/app/src/main/resources/modules -e PYTHONPATH=/app -e ANTHROPIC_API_KEY=${params.anthropic_api_key}"
}
trace {
enabled = true
overwrite = true
file = "$baseDir/trace.txt"
disease_name = null
outdir = null
population = 10
gender = 0.5
min_age = 0
max_age = 90
seed = null
}
profiles {
standard {
process.executor = 'local'
}
standard {
docker {
enabled = true
temp = 'auto'
}
}
k8s {
process {
executor = 'k8s'
container = 'harbor.cluster.omic.ai/omic/synthea-alldiseases:latest'
}
docker {
enabled = true
}
k8s {
storageClaimName = 'eureka-pvc'
storageMountPath = '/omic/eureka'
}
}
}
process {
cpus = 2
memory = '4 GB'
}

View File

@@ -1,14 +1,117 @@
{
"disease_name": "Parkinson's Disease",
"modules_dir": "modules",
"output_dir": "/mnt/OmicNAS/private/old/olamide/synthea/output/new",
"generate_patients": true,
"population": 10,
"gender": 0.5,
"min_age": 0,
"max_age": 90,
"analyze_patient_data": false,
"report_format": "html",
"force_generate": false,
"publish_dir": "published_output"
"params": {
"disease_name": {
"type": "string",
"description": "Disease name to generate synthetic patients for",
"default": "Diabetes",
"required": true,
"pipeline_io": "parameter",
"var_name": "params.disease_name",
"examples": [
"Diabetes",
"Hypertension",
"Lung Cancer"
],
"pattern": ".*",
"enum": [],
"validation": {},
"notes": "The disease name used to find or generate a Synthea disease module. Case-insensitive."
},
"population": {
"type": "integer",
"description": "Number of synthetic patients to generate",
"default": 10,
"required": false,
"pipeline_io": "parameter",
"var_name": "params.population",
"examples": [
10,
100,
1000
],
"pattern": "^\\d+$",
"enum": [],
"validation": {
"min": 1,
"max": 10000
},
"notes": "Higher numbers take longer to generate."
},
"gender": {
"type": "number",
"description": "Proportion of female patients (0.0 = all male, 1.0 = all female)",
"default": 0.5,
"required": false,
"pipeline_io": "parameter",
"var_name": "params.gender",
"examples": [
0.5,
0.0,
1.0
],
"pattern": "^[01]\\.?\\d*$",
"enum": [],
"validation": {
"min": 0.0,
"max": 1.0
},
"notes": "Decimal between 0.0 and 1.0 representing the proportion of female patients."
},
"min_age": {
"type": "integer",
"description": "Minimum age of generated patients",
"default": 0,
"required": false,
"pipeline_io": "parameter",
"var_name": "params.min_age",
"examples": [
0,
18,
40
],
"pattern": "^\\d+$",
"enum": [],
"validation": {
"min": 0,
"max": 140
},
"notes": "Minimum patient age in years."
},
"max_age": {
"type": "integer",
"description": "Maximum age of generated patients",
"default": 90,
"required": false,
"pipeline_io": "parameter",
"var_name": "params.max_age",
"examples": [
90,
65,
100
],
"pattern": "^\\d+$",
"enum": [],
"validation": {
"min": 1,
"max": 140
},
"notes": "Maximum patient age in years."
},
"outdir": {
"type": "folder",
"description": "Output directory for generated patient data",
"default": "s3://omic/eureka/synthea-alldiseases/output",
"required": true,
"pipeline_io": "output",
"var_name": "params.outdir",
"examples": [
"s3://omic/eureka/synthea-alldiseases/output",
"s3://omic/eureka/synthea-alldiseases/results"
],
"pattern": ".*",
"enum": [],
"validation": {},
"notes": "Directory where generated FHIR patient bundles will be stored."
}
}
}