Merge pull request #57 from samply/feature/idManagementModule

Added Module for Id-Management
2023-02-09 10:22:30 +01:00 · 2023-02-09 10:22:30 +01:00 · e4ac27061a
parent c75c9ed673 5a7dd1d0ae
commit e4ac27061a
10 changed files with 281 additions and 3 deletions
--- a/README.md
+++ b/README.md
@ -139,6 +139,15 @@ Your Bridgehead will automatically and regularly check for updates. Whenever som

 If you would like to understand what happens exactly and when, please check the systemd units deployed during the [installation](#base-installation) via `systemctl cat bridgehead-update@<PROJECT>.service` and `systemctl cat bridgehead-update@<PROJECT.timer`.

+### Auto-Backups
+Some of the components in the bridgehead will store persistent data. For those components, we integrated an automated backup solution in the bridgehead updates. It will automatically save the backup in multiple files
+
+1) Last-XX, were XX represents a weekday to allow re-import of at least one version of the database for each of the past seven days.
+2) Year-KW-XX, were XX represents the calendar week to allow re-import of at least one version per calendar week
+3) Year-Month, to allow re-import of at least one version per month
+
+To enable the Auto-Backup feature, please set the Variable `BACKUP_DIRECTORY` in your sites configuration.
+
 ### Monitoring

 To keep all Bridgeheads up and working and detect any errors before a user does, a central monitoring 
--- a/ccp/modules/id-management-compose.yml
+++ b/ccp/modules/id-management-compose.yml
@ -0,0 +1,57 @@
+version: "3.7"
+services:
+  id-manager:
+    image: docker.verbis.dkfz.de/bridgehead/magicpl
+    container_name: bridgehead-id-manager
+    environment:
+      TOMCAT_REVERSEPROXY_FQDN: ${HOST}
+      MAGICPL_SITE: ${IDMANAGEMENT_FRIENDLY_ID}
+      MAGICPL_ALLOWED_ORIGINS: https://${HOST}
+      MAGICPL_LOCAL_PATIENTLIST_APIKEY: ${IDMANAGER_LOCAL_PATIENTLIST_APIKEY}
+      MAGICPL_CENTRAXX_APIKEY: ${IDMANAGER_UPLOAD_APIKEY}
+      MAGICPL_CONNECTOR_APIKEY: ${IDMANAGER_READ_APIKEY}
+      MAGICPL_CENTRAL_PATIENTLIST_APIKEY: ${IDMANAGER_CENTRAL_PATIENTLIST_APIKEY}
+      MAGICPL_CONTROLNUMBERGENERATOR_APIKEY: ${IDMANAGER_CONTROLNUMBERGENERATOR_APIKEY}
+      MAGICPL_OIDC_CLIENT_ID: ${IDMANAGER_AUTH_CLIENT_ID}
+      MAGICPL_OIDC_CLIENT_SECRET: ${IDMANAGER_AUTH_CLIENT_SECRET}
+    depends_on:
+      - patientlist
+    labels:
+      - "traefik.enable=true"
+      - "traefik.http.routers.id-manager.rule=PathPrefix(`/id-manager`)"
+      - "traefik.http.services.id-manager.loadbalancer.server.port=8080"
+      - "traefik.http.routers.id-manager.tls=true"
+
+  patientlist:
+    image: docker.verbis.dkfz.de/bridgehead/mainzelliste
+    container_name: bridgehead-patientlist
+    environment:
+      - TOMCAT_REVERSEPROXY_FQDN=${HOST}
+      - ML_SITE=${IDMANAGEMENT_FRIENDLY_ID}
+      - ML_DB_PASS=${PATIENTLIST_POSTGRES_PASSWORD}
+      - ML_API_KEY=${IDMANAGER_LOCAL_PATIENTLIST_APIKEY}
+      - ML_UPLOAD_API_KEY=${IDMANAGER_UPLOAD_APIKEY}
+      # Add Variables from /etc/patientlist-id-generators.env
+      - PATIENTLIST_SEEDS_TRANSFORMED
+    labels:
+      - "traefik.enable=true"
+      - "traefik.http.routers.patientlist.rule=PathPrefix(`/patientlist`)"
+      - "traefik.http.services.patientlist.loadbalancer.server.port=8080"
+      - "traefik.http.routers.patientlist.tls=true"
+    depends_on:
+      - patientlist-db
+
+  patientlist-db:
+    image: postgres:15.1-alpine
+    container_name: bridgehead-patientlist-db
+    environment:
+      POSTGRES_USER: "mainzelliste"
+      POSTGRES_DB: "mainzelliste"
+      POSTGRES_PASSWORD: ${PATIENTLIST_POSTGRES_PASSWORD}
+    volumes:
+      - "patientlist-db-data:/var/lib/postgresql/data"
+      # NOTE: Add backups here. This is only imported if /var/lib/bridgehead/data/patientlist/ is empty!!!
+      - "/tmp/bridgehead/patientlist/:/docker-entrypoint-initdb.d/"
+
+volumes:
+  patientlist-db-data:
--- a/ccp/modules/id-management-setup.sh
+++ b/ccp/modules/id-management-setup.sh
@ -0,0 +1,52 @@
+#!/bin/bash
+
+function idManagementSetup() {
+	if [ -n "$IDMANAGER_UPLOAD_APIKEY" ]; then
+		log INFO "id-management setup detected -- will start id-management (mainzelliste & magicpl)."
+		OVERRIDE+=" -f ./$PROJECT/modules/id-management-compose.yml"
+
+		# Auto Generate local Passwords
+		PATIENTLIST_POSTGRES_PASSWORD="$(echo \"id-management-module-db-password-salt\" | openssl rsautl -sign -inkey /etc/bridgehead/pki/${SITE_ID}.priv.pem | base64 | head -c 30)"
+		IDMANAGER_LOCAL_PATIENTLIST_APIKEY="$(cat /proc/sys/kernel/random/uuid | sed 's/[-]//g' | head -c 20)"
+
+		# Transform Seeds Configuration to pass it to the Mainzelliste Container
+		PATIENTLIST_SEEDS_TRANSFORMED="$(declare -p PATIENTLIST_SEEDS | tr -d '\"' | sed 's/\[/\[\"/g' | sed 's/\]/\"\]/g')"
+
+		# Ensure old ids are working !!!
+		export IDMANAGEMENT_FRIENDLY_ID=$(legacyIdMapping "$SITE_ID")
+	fi
+}
+
+# Transform into single string array, e.g. 'dktk-test' to 'dktk test'
+# Usage: transformToSingleStringArray 'dktk-test' -> 'dktk test'
+function transformToSingleStringArray() {
+	echo "${1//-/ }";
+}
+
+# Ensure all Words are Uppercase
+# Usage: transformToUppercase 'dktk test' -> 'Dktk Test'
+function transformToUppercase() {
+	result="";
+	for word in $1; do
+		result+=" ${word^}";
+	done
+	echo "$result";
+}
+
+# Handle all execeptions from the norm (e.g LMU, TUM)
+# Usage: applySpecialCases 'Muenchen Lmu Test' -> 'Muenchen LMU Test'
+function applySpecialCases() {
+	result="$1";
+	result="${result/Lmu/LMU}";
+	result="${result/Tum/TUM}";
+	echo "$result";
+}
+
+# Transform current siteids to legacy version
+# Usage: legacyIdMapping "dktk-test" -> "DktkTest"
+function legacyIdMapping() {
+	single_string_array=$(transformToSingleStringArray "$1");
+	uppercase_string=$(transformToUppercase "$single_string_array");
+	normalized_string=$(applySpecialCases "$uppercase_string");
+	echo "$normalized_string" | tr -d ' '
+}
--- a/ccp/modules/id-management.md
+++ b/ccp/modules/id-management.md
@ -0,0 +1,66 @@
+# Module: Id-Management
+This module provides integration with the CCP-Pseudonymiziation Service. To learn more on the backgrounds of this service, you can refer to the [CCP Data Protection Concept](https://dktk.dkfz.de/klinische-plattformen/documents-download).
+
+## Getting Started
+The following configuration variables are added to your sites-configuration repository:
+
+```
+IDMANAGER_UPLOAD_APIKEY="<random-string>"
+IDMANAGER_READ_APIKEY="<random-string>"
+IDMANAGER_CENTRAL_PATIENTLIST_APIKEY="<given-to-you-by-ccp-it>"
+IDMANAGER_CONTROLNUMBERGENERATOR_APIKEY="<given-to-you-by-ccp-it>"
+IDMANAGER_AUTH_CLIENT_ID="<given-to-you-by-ccp-it>"
+IDMANAGER_AUTH_CLIENT_SECRET="<given-to-you-by-ccp-it>"
+
+IDMANAGER_SEEDS_BK="<three-numbers>"
+IDMANAGER_SEEDS_MDS="<three-numbers>"
+IDMANAGER_SEEDS_DKTK000001985="<three-numbers>"
+```
+> NOTE: Additionally, the CCP-IT adds lines declaring the `PATIENTLIST_SEEDS` array in your site configuration. This will contain the seeds for the different id-generators used in all projects.
+
+Once your Bridgehead is updated and restarted, you're all set!
+
+## Additional information you may want to know
+
+### Services
+
+Upon configuration, the Bridgehead will spawn the following services:
+
+- The `bridgehead-id-manager` at https://bridgehead.local/id-manager, provides a common interface for creating pseudonyms in the bridgehead.
+- The `bridgehead-patientlist` at https://bridgehead.local/patientlist is a local instance of the open-source software [Mainzelliste](https://mainzelliste.de). This service's primary task is to map patients IDAT to pseudonyms identifying them along the different CCP projects.
+- The `bridgehead-patientlist-db` is only accessible within the Bridgehead itself. This is a local postgresql instance storing the database for `bridgehead-patientlist`. The data is persisted as a named volume `patientlist-db-data`.
+
+### How to import an existing database (e.g from Legacy Windows or from Backups)
+First you must shutdown your local bridgehead instance:
+```
+systemctl stop bridgehead@ccp
+```
+
+Next you need to remove the current patientlist database:
+```
+docker volume rm patientlist-db-data;
+```
+
+Third, you need to place your postgres dump in the import directory `/tmp/bridgehead/patientlist/some-dump.sql`. This will only be imported, then the volume `patientlist-db-data` was removed previously. 
+> NOTE: Please create the postgres dump with the options "--no-owner" and "--no-privileges". Additionally ensure the dump is created in the plain format (SQL).
+
+After this, you can restart your bridgehead and the dump will be imported:
+```
+systemctl start bridgehead@ccp
+```
+
+### How to connect your local data-management
+Typically, the sites connect their local data-management for the pseudonym creation with the id-management in the bridgehead. In the following two sections, you can read where you can change the configuration:
+#### Sites using CentraXX
+On your CentraXX Server, you need to change following settings in the "centraxx-dev.properties" file.
+```
+dktk.idmanagement.url=https://<your-linux-bk-host>/id-manager/translator/getId
+dktk.idmanagement.apiKey=<your-setting-for-IDMANAGER_UPLOAD_APIKEY>
+```
+They typically already exist, but need to be changed to the new values!
+#### Sites using ADT2FHIR
+@Pierre
+
+
+### How to connect the legacy windows bridgehead
+You need to change the configuration file "..." of your Windows Bridgehead. TODO... 
--- a/ccp/mtba-compose.yml
+++ b/ccp/mtba-compose.yml
@ -0,0 +1,36 @@
+version: "3.7"
+
+services:
+  mtba:
+    image: samply/mtba:develop
+    container_name: bridgehead-mtba
+    environment:
+      BLAZE_STORE_URL: http://blaze:8080
+      # NOTE: Aktuell Berechtigungen wie MagicPL!!!
+      # TODO: Add separate ApiKey to MagicPL only for MTBA!
+      ID_MANAGER_API_KEY: ${IDMANAGER_UPLOAD_APIKEY}
+      ID_MANAGER_PSEUDONYM_ID_TYPE: BK_${IDMANAGEMENT_FRIENDLY_ID}_L-ID
+      ID_MANAGER_URL: http://id-manager:8080/id-manager
+      PATIENT_CSV_FIRST_NAME_HEADER: ${MTBA_PATIENT_CSV_FIRST_NAME_HEADER}
+      PATIENT_CSV_LAST_NAME_HEADER: ${MTBA_PATIENT_CSV_LAST_NAME_HEADER}
+      PATIENT_CSV_GENDER_HEADER: ${MTBA_PATIENT_CSV_GENDER_HEADER}
+      PATIENT_CSV_BIRTHDAY_HEADER: ${MTBA_PATIENT_CSV_BIRTHDAY_HEADER}
+      CBIOPORTAL_URL:  http://cbioportal:8080
+      FILE_CHARSET: ${MTBA_FILE_CHARSET}
+      FILE_END_OF_LINE: ${MTBA_FILE_END_OF_LINE}
+      CSV_DELIMITER: ${MTBA_CSV_DELIMITER}
+    labels:
+      - "traefik.enable=true"
+      - "traefik.http.routers.mtba.rule=PathPrefix(`/`)"
+      - "traefik.http.services.mtba.loadbalancer.server.port=80"
+      - "traefik.http.routers.mtba.tls=true"
+    volumes:
+      - /tmp/bridgehead/mtba/input:/app/input
+      - /tmp/bridgehead/mtba/persist:/app/persist
+
+  # TODO: Include CBioPortal in Deployment ...
+  # NOTE: CBioPortal can't load data while the system is running. So after import of data bridgehead needs to be restarted!
+  # TODO: Find a trigger to let mtba signal a restart for CBioPortal
+
+volumes:
+  mtba-data:
--- a/ccp/nngm-setup.sh
+++ b/ccp/nngm-setup.sh
@ -7,3 +7,15 @@ function nngmSetup() {
 	fi
 	CONNECTOR_POSTGRES_PASSWORD="$(echo \"This is a salt string to generate one consistent password. It is not required to be secret.\" | openssl rsautl -sign -inkey /etc/bridgehead/pki/${SITE_ID}.priv.pem | base64 | head -c 30)"
 }
+
+function mtbaSetup() {
+	# TODO: Check if ID-Management Module is activated!
+	if [ -n "$ENABLE_MTBA" ];then
+		log INFO "MTBA setup detected -- will start MTBA Service and CBioPortal."
+		if [ ! -n "$IDMANAGER_UPLOAD_APIKEY" ]; then
+			log ERROR "Detected MTBA Module configuration but ID-Management Module seems not to be configured!"
+			exit 1;
+		fi
+		OVERRIDE+=" -f ./$PROJECT/mtba-compose.yml"
+	fi
+}
--- a/ccp/vars
+++ b/ccp/vars
@ -8,8 +8,12 @@ REPORTHUB_BEAM_SECRET_LONG="ApiKey report-hub.${PROXY_ID} ${REPORTHUB_BEAM_SECRE
 SUPPORT_EMAIL=support-ccp@dkfz-heidelberg.de
 PRIVATEKEYFILENAME=/etc/bridgehead/pki/${SITE_ID}.priv.pem

+# This will load id-management setup. Effective only if id-management configuration is defined.
+source $PROJECT/modules/id-management-setup.sh
+idManagementSetup
 # This will load nngm setup. Effective only if nngm configuration is defined.
 source $PROJECT/nngm-setup.sh
 nngmSetup
 source $PROJECT/exliquid-setup.sh
 exliquidSetup
+mtbaSetup
--- a/lib/functions.sh
+++ b/lib/functions.sh
@ -131,11 +131,22 @@ fail_and_report() {

 setHostname() {
 	if [ -z "$HOST" ]; then
-		export HOST=$(hostname -f)
+		export HOST=$(hostname -f | tr "[:upper:]" "[:lower:]")
 		log DEBUG "Using auto-detected hostname $HOST."
 	fi
 }

+# Takes 1) The Backup Directory Path 2) The name of the Service to be backuped
+# Creates 3 Backups: 1) For the past seven days 2) For the current month and 3) for each calendar week
+createEncryptedPostgresBackup(){
+  docker exec "$2" bash -c 'pg_dump -U $POSTGRES_USER $POSTGRES_DB --format=p --no-owner --no-privileges' | \
+      # TODO: Encrypt using /etc/bridgehead/pki/${SITE_ID}.priv.pem | \
+      tee "$1/$2/$(date +Last-%A).sql" | \
+      tee "$1/$2/$(date +%Y-%m).sql" > \
+      "$1/$2/$(date +%Y-KW%V).sql"
+}
+
+
 # from: https://gist.github.com/sj26/88e1c6584397bb7c13bd11108a579746
 # ex. use: retry 5 /bin/false
 function retry {
--- a/lib/update-bridgehead.sh
+++ b/lib/update-bridgehead.sh
@ -81,7 +81,7 @@ done
 # Check docker updates
 log "INFO" "Checking for updates to running docker images ..."
 docker_updated="false"
-for IMAGE in $(cat $PROJECT/docker-compose.yml | grep -v "^#" | grep "image:" | sed -e 's_^.*image: \(.*\).*$_\1_g; s_\"__g'); do
+for IMAGE in $(cat $PROJECT/docker-compose.yml ${OVERRIDE//-f/} | grep -v "^#" | grep "image:" | sed -e 's_^.*image: \(.*\).*$_\1_g; s_\"__g'); do
  log "INFO" "Checking for Updates of Image: $IMAGE"
  if docker pull $IMAGE | grep "Downloaded newer image"; then
    CHANGE="Image $IMAGE updated."
@ -103,6 +103,37 @@ else
  hc_send log "$RES"
 fi

+if [ -n "${BACKUP_DIRECTORY}" ]; then
+  if [ ! -d "$BACKUP_DIRECTORY" ]; then
+    message="Performing automatic maintenance: Attempting to create backup directory $BACKUP_DIRECTORY."
+    hc_send log "$message"
+    log INFO "$message"
+    mkdir -p "$BACKUP_DIRECTORY"
+    chown -R "$BACKUP_DIRECTORY" bridgehead;
+  fi
+  checkOwner "$BACKUP_DIRECTORY" bridgehead || fail_and_report 1 "Automatic maintenance failed: Wrong permissions for backup directory $(pwd)"
+  # Collect all container names that contain '-db'
+  BACKUP_SERVICES="$(docker ps --filter name=-db --format "{{.Names}}" | tr "\n" "\ ")"
+  log INFO "Performing automatic maintenance: Creating Backups for $BACKUP_SERVICES";
+  for service in $BACKUP_SERVICES; do
+    if [ ! -d "$BACKUP_DIRECTORY/$service" ]; then
+      message="Performing automatic maintenance: Attempting to create backup directory for $service in $BACKUP_DIRECTORY."
+      hc_send log "$message"
+      log INFO "$message"
+      mkdir -p "$BACKUP_DIRECTORY/$service"
+    fi
+    if createEncryptedPostgresBackup "$BACKUP_DIRECTORY" "$service"; then
+      message="Performing automatic maintenance: Stored encrypted backup for $service in $BACKUP_DIRECTORY."
+      hc_send log "$message"
+      log INFO "$message"
+    else
+      fail_and_report 5 "Failed to create encrypted update for $service"
+    fi
+  done
+else
+  log WARN "Automated backups are disabled (variable AUTO_BACKUPS != \"true\")"
+fi
+
 exit 0

 # TODO: Print last commit explicit