LabCAS Backend

Repository containing back-end services and configuration for executing EDRN LabCAS data processing workflows.

Starting the Service

Create a .env file (or specify one with --env) and launch it with labcas-backend. Run labcas-backend --help for more information.

By default, API endpoints are served from the root path, such as /auth. If the service is behind a reverse proxy that keeps a subpath in forwarded requests, set LABCAS_SUBPATH_PREFIX in .env, for example LABCAS_SUBPATH_PREFIX=labcas-backend-data-access-api, to serve endpoints such as /labcas-backend-data-access-api/auth.

Zipperlab Integration

To test sending queries to Zipperlab, first get your JWT (above) then do:

curl --request POST --verbose --insecure \
    --cookie "JasonWebToken=`cat /tmp/jwt`" \
    --header "Authentication: Bearer `cat /tmp/jwt`" \
    --header 'Content-Type: application/x-www-form-urlencoded' \
    --data 'email=hello@a.co&query=id:Pre-diagnostic_PDAC_Images/City_of_Hope/COH_0171/COH_01710003/DICOM/I883*' \
    https://localhost:8444/labcas-backend-data-access-api/zip

Make sure you have Zipperlab running and set its URL in ~/labcas.properties.

If you want to send file IDs instead, do:

curl --request POST --verbose --insecure \
    --cookie "JasonWebToken=`cat /tmp/jwt`" \
    --header "Authentication: Bearer `cat /tmp/jwt`" \
    --header 'Content-Type: application/x-www-form-urlencoded' \
    --data 'email=hello@a.co&id=FILE1&id=FILE2&id=FILE3' \
    https://localhost:8444/labcas-backend-data-access-api/zip

Loading Solr Data

If you have downloaded backups of Solr data you can reload it generally as follows:

curl --insecure --verbose "https://localhost:8984/solr/collections/replication?command=restore&location=$BACKUP_PATH/collections"
curl --insecure --verbose "https://localhost:8984/solr/datasets/replication?command=restore&location=$BACKUP_PATH/datasets"
curl --insecure --verbose "https://localhost:8984/solr/files/replication?command=restore&location=$BACKUP_PATH/files"

Replace $BACKUP_PATH with the location of the downloaded backups.# LabCAS Backend (Python)

This package is the Python rewrite of the LabCAS Data Access API. It mirrors the existing Java functionality while providing a modern FastAPI-based stack with an extensible directory abstraction, Redis-backed session management, and a pluggable search layer.

Development

python -m venv .venv
source .venv/bin/activate
pip install --upgrade pip
pip install --editable .
labcas-backend --help

Create a .env file based on env.example before running the service or tests.

Solr Migration

This guide explains how to export data from your old Solr instance (https://localhost:8984/solr) and import it into a newer Solr instance.

The migration is handled by the migrate-solr console script, which is part of the jpl.labcas.backend package. After installing the package, you can use migrate-solr directly.

Prerequisites

• Python 3.13+ (as required by the package)
• The jpl.labcas.backend package installed (includes required dependencies like httpx and tqdm)

Quick Start

The migration tool operates in two modes: export (to save cores to files) and import (to load cores from files).

Export Mode

# Export all cores from a Solr instance
migrate-solr --mode export https://localhost:8984/solr

# Export specific cores only
migrate-solr --mode export https://localhost:8984/solr --cores collections,datasets

# Specify backup directory (default: /tmp/solr-backup-<timestamp>)
migrate-solr --mode export https://localhost:8984/solr --backup-dir /path/to/backup

Import Mode

# Import all cores into a Solr instance (from current directory)
migrate-solr --mode import https://newhost:8983/solr

# Import from a specific backup directory
migrate-solr --mode import https://newhost:8983/solr --backup-dir /tmp/solr-backup-20240101-120000

# Import specific cores only
migrate-solr --mode import https://newhost:8983/solr --backup-dir /tmp/solr-backup-20240101-120000 --cores collections,datasets

Your Solr Cores

The LabCAS Solr engine has the following cores:

• collections - Collection metadata
• datasets - Dataset metadata
• files - File metadata
• userdata - User data

Migration Process

The migration process consists of two phases:

1. Export Phase

• Connects to your old Solr instance
• Exports all documents from each core using cursorMark pagination
• Saves data as JSON files in the backup directory

2. Import Phase

• Connects to your new Solr instance
• Imports documents into corresponding cores
• Commits changes after import

Important Notes

Before Migration

1. Create cores in new Solr: The new Solr instance must have the cores created first. You can:

   • Copy the core configuration from solr/src/main/resources/solr-home/ to your new Solr instance
   • Or use the Solr Admin UI: http://newhost:8983/solr/admin/cores?action=CREATE&name=<core>&instanceDir=<core>

2. Schema compatibility: Ensure the schema.xml files are compatible between old and new Solr versions. You may need to update field types or configurations.

3. Backup first: Always backup your data before migration!

During Migration

• The tool will prompt you if a core doesn't exist in the target instance
• Large cores may take significant time to export/import
• Progress is shown during the process with progress bars

After Migration

1. Verify data: Check document counts match:

   curl "https://oldhost:8984/solr/collections/select?q=*:*&rows=0"
   curl "https://newhost:8983/solr/collections/select?q=*:*&rows=0"
   

2. Update configuration: Update your application's LABCAS_SOLR_URL environment variable

3. Test thoroughly: Run your application tests to ensure everything works

Alternative: Using Solr Backup API

For very large datasets, you might prefer using Solr's built-in backup API:

Export (on old Solr)

# For each core
curl "https://localhost:8984/solr/collections/replication?command=backup&location=/path/to/backup"

Import (on new Solr)

# Copy backup files to new Solr server, then restore
curl "https://newhost:8983/solr/collections/replication?command=restore&location=/path/to/backup"

Note: This method requires filesystem access to both Solr servers and the cores must have identical configurations.

Troubleshooting

SSL Certificate Errors

If you get SSL errors, the tool skips verification by default. Use --verify-ssl if you want to verify certificates.

Core Not Found

If a core doesn't exist in the new instance, the script will prompt you to create it. Make sure to:

1. Copy the conf/ directory from the old core
2. Create the core using Solr Admin API or UI
3. Then continue with the import

Large Datasets

For very large cores (>1M documents), consider:

• Running exports during off-peak hours
• Running export and import separately (export first, then import in a separate command)
• The tool processes documents in batches of 1000 to manage memory efficiently

Memory Issues

If you encounter memory issues:

• The tool processes documents in batches of 1000
• Consider migrating one core at a time using the --cores option
• Ensure sufficient disk space for backup files

Support

For issues or questions, check:

• Solr documentation: https://solr.apache.org/guide/
• Your project's Solr configuration in solr/src/main/resources/solr-home/