Add db-solr-sync-next for catalog-next

Dockerfile

@@ -16,7 +16,7 @@ ENV CLASSPATH=${CLASSPATH}:/usr/lib/jvm/java-11-openjdk/saxon/saxon.jar
 RUN pip install --upgrade pip
 # RUN python3 -m pip install 'cython<3'
 # RUN python3 -m pip install --no-use-pep517 pyproj==3.4.1
-RUN python3 -m pip install pyproj@git+https://github.com/pyproj4/pyproj.git@main
+RUN python3 -m pip install pyproj@git+https://github.com/pyproj4/pyproj.git@3.6.1
 
 COPY . $APP_DIR/
 

Makefile

@@ -2,28 +2,28 @@ CKAN_VERSION ?= 2.10
 COMPOSE_FILE ?= docker-compose.yml
 
 build: ## Build the docker containers
-	CKAN_VERSION=$(CKAN_VERSION) docker-compose -f $(COMPOSE_FILE) build
+	CKAN_VERSION=$(CKAN_VERSION) docker compose -f $(COMPOSE_FILE) build
 debug:
-	CKAN_VERSION=$(CKAN_VERSION) docker-compose run --service-ports app
+	CKAN_VERSION=$(CKAN_VERSION) docker compose run --service-ports app
 
 lint: ## Lint the code
-	CKAN_VERSION=$(CKAN_VERSION) docker-compose -f docker-compose.yml run --rm app flake8 /srv/app/ckanext/ --count --max-line-length=127 --show-source --statistics --exclude ckan
+	CKAN_VERSION=$(CKAN_VERSION) docker compose -f docker-compose.yml run --rm app flake8 /srv/app/ckanext/ --count --max-line-length=127 --show-source --statistics --exclude ckan
 
 clean: ## Clean workspace and containers
 	find . -name *.pyc -delete
-	CKAN_VERSION=$(CKAN_VERSION) docker-compose -f $(COMPOSE_FILE) down -v --remove-orphans
+	CKAN_VERSION=$(CKAN_VERSION) docker compose -f $(COMPOSE_FILE) down -v --remove-orphans
 
 test: ## Run tests in a new container
-	CKAN_VERSION=$(CKAN_VERSION) docker-compose -f $(COMPOSE_FILE) run --rm app /srv/app/test.sh
+	CKAN_VERSION=$(CKAN_VERSION) docker compose -f $(COMPOSE_FILE) run --rm app /srv/app/test.sh
 
 java-test: ## Test java transformation command (java + saxon installed)
-	CKAN_VERSION=$(CKAN_VERSION) docker-compose -f $(COMPOSE_FILE) run --rm app bash -c "java net.sf.saxon.Transform -s:/app/ckanext/geodatagov/tests/data-samples/waf-fgdc/fgdc-csdgm_sample.xml -xsl:/app/ckanext/geodatagov/harvesters/fgdcrse2iso19115-2.xslt"
+	CKAN_VERSION=$(CKAN_VERSION) docker compose -f $(COMPOSE_FILE) run --rm app bash -c "java net.sf.saxon.Transform -s:/app/ckanext/geodatagov/tests/data-samples/waf-fgdc/fgdc-csdgm_sample.xml -xsl:/app/ckanext/geodatagov/harvesters/fgdcrse2iso19115-2.xslt"
 
 up: ## Start the containers
-	CKAN_VERSION=$(CKAN_VERSION) docker-compose -f $(COMPOSE_FILE) up
+	CKAN_VERSION=$(CKAN_VERSION) docker compose -f $(COMPOSE_FILE) up
 
 ci: ## Start the containers in the background
-	CKAN_VERSION=$(CKAN_VERSION) docker-compose -f $(COMPOSE_FILE) up -d
+	CKAN_VERSION=$(CKAN_VERSION) docker compose -f $(COMPOSE_FILE) up -d
 
 .DEFAULT_GOAL := help
 .PHONY: build clean help lint test up

README.md

@@ -46,18 +46,18 @@ All the tests live in the [/ckanext/geodatagov/tests](/ckanext/geodatagov/tests)
 ### Build Environment
 
 To start environment, run:
-```docker-compose build```
-```docker-compose up```
+```docker compose build```
+```docker compose up```
 
 CKAN will start at localhost:5000
 
 To shut down environment, run:
 
-```docker-compose down```
+```docker compose down```
 
 To docker exec into the CKAN image, run:
 
-```docker-compose exec app /bin/bash```
+```docker compose exec app /bin/bash```
 
 ### Testing
 

ckanext/geodatagov/cli.py

@@ -306,6 +306,41 @@ def index_for(_type):
         return NoopSearchIndex()
 
 
+def get_all_entity_ids_date():
+    """
+    Return a list of the IDs and metadata_modified of all indexed packages.
+
+    Solr result is to be processed in batches at 10000
+    pagination to avoid out-of-memory.
+    """
+    query = "*:*"
+    fq = '+site_id:"%s" ' % config.get("ckan.site_id")
+    fq += "+state:active "
+    fq += "+type:dataset "
+
+    ret_all = []
+
+    start = 0
+    page_size = 10000
+    conn = make_connection()
+
+    log.info(f"Now loading SOLR packages using page size {page_size}...")
+
+    while True:
+        log.info(f"loading packages starting from {start}")
+        data = conn.search(query, fq=fq, start=start, rows=page_size, fl="id, metadata_modified")
+
+        if not data:
+            break
+
+        for r in data.docs:
+            ret_all.append((r.get("id"), r.get("metadata_modified")))
+
+        start += page_size
+
+    return ret_all
+
+
 def get_all_entity_ids_date_hoid():
     """
     Return a list of the IDs and metadata_modified of all indexed packages.
@@ -484,6 +519,87 @@ def db_solr_sync(dryrun, cleanup_solr, update_solr):
         print(*active_package_id_wo_ho, sep='\n')
 
 
+@geodatagov.command()
+@click.option("--dryrun", is_flag=True, help="inspect what will be updated")
+@click.option(
+    "--cleanup_solr", is_flag=True, help="Only remove orphaned entries in Solr"
+)
+@click.option(
+    "--update_solr",
+    is_flag=True,
+    help=(
+        "(Update solr entries with new data from DB) OR (Add DB data to Solr that is missing)"
+    ),
+)
+def db_solr_sync_next(dryrun, cleanup_solr, update_solr):
+    """db solr sync next for catalog-next"""
+    if dryrun:
+        log.info("Starting dryrun to update index.")
+
+    package_index = index_for(model.Package)
+
+    # get active packages from DB
+    active_package = {
+        (r[0], r[1].replace(microsecond=0))
+        for r in model.Session.query(
+            model.Package.id,
+            model.Package.metadata_modified
+        )
+        .filter(
+            model.Package.type == "dataset",
+            model.Package.state == "active"
+        )
+        .all()
+    }
+
+    log.info(f"total {len(active_package)} DB active_package")
+
+    # get indexed packages from solr
+    indexed_package = set(get_all_entity_ids_date())
+    log.info(f"total {len(indexed_package)} solr indexed_package")
+
+    solr_package = indexed_package - active_package
+    db_package = active_package - indexed_package
+
+    work_list = {}
+    for id, *_ in solr_package:
+        work_list[id] = "solr"
+    for id, *_ in db_package:
+        if id in work_list:
+            work_list[id] = "solr-db"
+        else:
+            work_list[id] = "db"
+
+    both = cleanup_solr == update_solr
+    set_cleanup = {i if work_list[i] == "solr" else None for i in work_list} - {None}
+    set_update = work_list.keys() - set_cleanup
+    log.info(f"{len(set_cleanup)} packages need to be removed from Solr")
+    log.info(f"{len(set_update)} packages need to be updated/added to Solr")
+
+    if not dryrun and set_cleanup and (cleanup_solr or both):
+        log.info("Deleting indexes")
+        delete_packages(set_cleanup)
+        package_index.commit()
+        log.info("Finished cleaning solr entries.")
+
+    if not dryrun and set_update and (update_solr or both):
+        log.info("Rebuilding indexes")
+        try:
+            rebuild(package_ids=set_update, defer_commit=True)
+        except Exception as e:
+            log.error("Error while rebuild index %s: %s" % (id, repr(e)))
+        package_index.commit()
+        log.info("Finished updating solr entries.")
+        log.info("Here is the first a few dataset ids that are rebuilt:")
+        count = 0
+        max = 10
+        for id in set_update:
+            count = count + 1
+            if count > max:
+                break
+            log.info(f"{count}: {id}")
+
+
 @geodatagov.command()
 def check_stuck_jobs():
     """check stuck harvest jobs"""

setup.py

@@ -10,7 +10,7 @@
 
 setup(
     name="ckanext-geodatagov",
-    version="0.2.8",
+    version="0.2.9",
     description="",
     long_description=long_description,
     long_description_content_type="text/markdown",

test.sh

@@ -2,7 +2,7 @@
 # Setup and run extension tests. This script should be run in a _clean_ CKAN
 # environment. e.g.:
 #
-#     $ docker-compose run --rm app ./test.sh
+#     $ docker compose run --rm app ./test.sh
 #
 
 set -o errexit