Skip to content
Snippets Groups Projects
Commit 849dfc4d authored by Alessandro Cerioni's avatar Alessandro Cerioni
Browse files

Improvement.

The previous configuration files (config.yaml + config-docker.yaml) were replaced by a unique template configuration file.
parent adf40d61
Branches
Tags
No related merge requests found
session:
id: abcd1234
id: <the session id, ex.: abcd1234>
working_directory: output
log_level: DEBUG
log_level: <the log level: DEBUG, INFO, WARNING, ERROR, INFO, CRITICAL>
geonetwork:
url: https://download.data.grandlyon.com/catalogue/srv/fre/q
url: <the URL of GeoNetwork's q API, ex.: https://download.data.grandlyon.com/catalogue/srv/fre/q >
records_per_page: 100
wfs:
features_per_page: 1000
rabbitmq:
host: rabbitmq
exchange: download_data_grandlyon_com_index
host: <the RabbitMQ host. N.B.: in case docker-compose is used, this should correspond to the service name>
exchange: download_data_grandlyon_com_indexing
routing_key_1_suffix: metadata_pages_to_process
queue_name_1_suffix: metadata_pages_to_process
routing_key_2_suffix: doc_pages_to_index
......@@ -29,33 +29,35 @@ rabbitmq:
mongo:
username: root
password: example
host: mongo
host: <the MongoDB host. N.B.: in case docker-compose is used, this should correspond to the service name>
port: 27017
data-db: download_data_grandlyon_com
report-db: download_data_grandlyon_com
data-db: <the name of the database where data are stored, ex.: download_data_grandlyon_com>
report-db: <the name of the database where reporting data are stored, ex.: download_data_grandlyon_com>
metadata_getter:
uuids_to_get:
#- all
- 46eac9e4-33cb-45b7-a104-7bcc8e654c98
#- all # 'all' means that all the uuids will be retrieved
- <1st uuid to get>
- <2nd uuid to get>
- <...>
uuids_to_filter_out:
- ef5745f1-060d-496d-8aea-2461538964a8
- eff0fb32-d259-45b8-b662-9eb00c64ade5
- <1st uuid to filter out>
- <2nd uuid to filter out>
- <...>
field_type_detector:
wait: true
rewrite: false
debug: false
wait: true # whether the field-type-detector should wait for the no. of docs in MongoDB to be stable
rewrite: false # whether it should overwrite already existing output files
debug: false # whether it should run in debug mode
indexer:
url: https://elastic2.alpha.grandlyon.com
index: download.data.grandlyon.com.ingest.v1
url: <the root URL of Elasticsearch, ex.: https://elastic2.alpha.grandlyon.com>
index: <the destination index, ex.: download.data.grandlyon.com.ingest.v2>
reindexer:
source_url: https://elastic2.alpha.grandlyon.com:443 # N.B.: the port number MUST be explicit
destination_url: https://elastic2.alpha.grandlyon.com
source_index: download.data.grandlyon.com.ingest.v1
destination_index: download.data.grandlyon.com.v1
template_name: download.data.grandlyon.com.v1
template_index_pattern: download.data.grandlyon.com.v1
#source_url: https://elastic2.alpha.grandlyon.com:443 # N.B.: the port number MUST be explicit; this parameter is optional
destination_url: <ex.: https://elastic2.alpha.grandlyon.com>
source_index: <ex.: download.data.grandlyon.com.ingest.v2>
destination_index: <ex.: download.data.grandlyon.com.v2>
template_name: <ex.: download.data.grandlyon.com.v2>
template_index_pattern: <ex.: download.data.grandlyon.com.v2>
session:
id: abcd1234
working_directory: output
log_level: DEBUG
geonetwork:
url: https://download.data.grandlyon.com/catalogue/srv/fre/q
records_per_page: 100
wfs:
features_per_page: 1000
rabbitmq:
host: localhost
exchange: download_data_grandlyon_com_index
routing_key_1_suffix: metadata_pages_to_process
queue_name_1_suffix: metadata_pages_to_process
routing_key_2_suffix: doc_pages_to_index
queue_name_2_suffix: doc_pages_to_index
routing_key_3_suffix: doc_pages_to_enrich
queue_name_3_suffix: doc_pages_to_enrich
routing_key_4_suffix: doc_pages_to_store_in_mongo
queue_name_4_suffix: doc_pages_to_store_in_mongo
routing_key_5_suffix: doc_pages_to_process
queue_name_5_suffix: doc_pages_to_process
#routing_key_6_suffix: reindex_tasks
#queue_name_6_suffix: reindex_tasks
mongo:
username: root
password: example
host: localhost
port: 27017
data-db: download_data_grandlyon_com
report-db: download_data_grandlyon_com
metadata_getter:
uuids_to_get:
#- all
- 46eac9e4-33cb-45b7-a104-7bcc8e654c98
uuids_to_filter_out:
- ef5745f1-060d-496d-8aea-2461538964a8
- eff0fb32-d259-45b8-b662-9eb00c64ade5
field_type_detector:
wait: true
rewrite: false
debug: false
indexer:
url: https://elastic2.alpha.grandlyon.com
index: download.data.grandlyon.com.ingest.v1
reindexer:
source_url: https://elastic2.alpha.grandlyon.com:443 # N.B.: the port number MUST be explicit
destination_url: https://elastic2.alpha.grandlyon.com
source_index: download.data.grandlyon.com.ingest.v1
destination_index: download.data.grandlyon.com.v1
template_name: download.data.grandlyon.com.v1
template_index_pattern: download.data.grandlyon.com.v1
......@@ -35,7 +35,7 @@ services:
image: data-grandlyon-com-indexer
command: python 0-reset-session.py
volumes:
- ./config-docker.yaml:/app/config.yaml:ro
- ./config.yaml:/app/config.yaml:ro
- working-directory:/app/output
metadata-getter:
......@@ -43,35 +43,35 @@ services:
image: data-grandlyon-com-indexer
command: python 1-metadata-getter.py
volumes:
- ./config-docker.yaml:/app/config.yaml:ro
- ./config.yaml:/app/config.yaml:ro
metadata-processor:
build: .
image: data-grandlyon-com-indexer
command: python 2-metadata-processor.py
volumes:
- ./config-docker.yaml:/app/config.yaml:ro
- ./config.yaml:/app/config.yaml:ro
doc-enricher:
build: .
image: data-grandlyon-com-indexer
command: python 3-doc-enricher.py
volumes:
- ./config-docker.yaml:/app/config.yaml:ro
- ./config.yaml:/app/config.yaml:ro
docs-to-mongodb:
build: .
image: data-grandlyon-com-indexer
command: python 4-docs-to-mongodb.py
volumes:
- ./config-docker.yaml:/app/config.yaml:ro
- ./config.yaml:/app/config.yaml:ro
field-type-detector:
build: .
image: data-grandlyon-com-indexer
command: python 5-field-type-detector.py
volumes:
- ./config-docker.yaml:/app/config.yaml:ro
- ./config.yaml:/app/config.yaml:ro
- working-directory:/app/output
doc-processor:
......@@ -79,7 +79,7 @@ services:
image: data-grandlyon-com-indexer
command: python 6-doc-processor.py
volumes:
- ./config-docker.yaml:/app/config.yaml:ro
- ./config.yaml:/app/config.yaml:ro
- working-directory:/app/output
doc-indexer:
......@@ -87,14 +87,14 @@ services:
image: data-grandlyon-com-indexer
command: python 7-doc-indexer.py
volumes:
- ./config-docker.yaml:/app/config.yaml:ro
- ./config.yaml:/app/config.yaml:ro
reindexer:
build: .
image: data-grandlyon-com-indexer
command: python 8-reindexer.py
volumes:
- ./config-docker.yaml:/app/config.yaml:ro
- ./config.yaml:/app/config.yaml:ro
volumes:
rabbitmq:
......
0% Loading or .
You are about to add 0 people to the discussion. Proceed with caution.
Please register or to comment