Enable use of `database_connection` in scripts without current_app

There is need to run external scripts using the same configurations as the
application but without the need to couple the script to the application.

In this case, we provide the needed configuration directly in the CLI, and
modify the existing `gn3.db_utils.database_connection` function to allow it to
work coupled to the app or otherwise.

2 files changed, 19 insertions, 12 deletions

diff --git a/gn3/db_utils.py b/gn3/db_utils.py
index 862d76c..4827358 100644
--- a/gn3/db_utils.py
+++ b/gn3/db_utils.py
@@ -7,20 +7,21 @@ import xapian
 from flask import current_app
 
 
-def parse_db_url() -> Tuple:
+def parse_db_url(sql_uri: str) -> Tuple:
     """function to parse SQL_URI env variable note:there\
     is a default value for SQL_URI so a tuple result is\
     always expected"""
-    parsed_db = urlparse(current_app.config["SQL_URI"])
+    parsed_db = urlparse(sql_uri)
     return (
         parsed_db.hostname, parsed_db.username, parsed_db.password,
         parsed_db.path[1:], parsed_db.port)
 
 
 # This function is deprecated. Use database_connection instead.
-def database_connector() -> mdb.Connection:
+def database_connector(sql_uri: str = "") -> mdb.Connection:
     """function to create db connector"""
-    host, user, passwd, db_name, db_port = parse_db_url()
+    host, user, passwd, db_name, db_port = parse_db_url(
+        sql_uri or current_app.config["SQL_URI"])
 
     return mdb.connect(host, user, passwd, db_name, port=(db_port or 3306))
 
@@ -33,10 +34,15 @@ class Connection(Protocol):
         ...
 
 
+## We need to decouple current_app from this module and function, but since this
+## function is used throughout the code, that will require careful work to update
+## all the code to pass the `sql_uri` argument, and make it a compulsory argument
+## rather than its current optional state.
 @contextlib.contextmanager
-def database_connection() -> Iterator[Connection]:
+def database_connection(sql_uri: str = "") -> Iterator[Connection]:
     """Connect to MySQL database."""
-    host, user, passwd, db_name, port = parse_db_url()
+    host, user, passwd, db_name, port = parse_db_url(
+        sql_uri or current_app.config["SQL_URI"])
     connection = mdb.connect(db=db_name,
                              user=user,
                              passwd=passwd or '',
diff --git a/scripts/index-genenetwork b/scripts/index-genenetwork
index b92fa5e..401ae1a 100755
--- a/scripts/index-genenetwork
+++ b/scripts/index-genenetwork
@@ -299,12 +299,12 @@ def worker_queue(number_of_workers: int = os.cpu_count() or 1) -> Generator:
         process.join()
 
 
-def index_query(index_function: Callable, query: SQLQuery, xapian_build_directory: pathlib.Path, start: int = 0) -> None:
+def index_query(index_function: Callable, query: SQLQuery, xapian_build_directory: pathlib.Path, sql_uri: str, start: int = 0) -> None:
     """Run SQL query, and index its results for Xapian."""
     i = start
     try:
         with worker_queue() as spawn_worker:
-            with database_connection() as conn:
+            with database_connection(sql_uri) as conn:
                 for chunk in group(query_sql(conn, serialize_sql(
                         # KLUDGE: MariaDB does not allow an offset
                         # without a limit. So, set limit to a "high"
@@ -325,7 +325,7 @@ def index_query(index_function: Callable, query: SQLQuery, xapian_build_director
     except MySQLdb._exceptions.OperationalError:
         logging.warning("Reopening connection to recovering from SQL operational error",
                         exc_info=True)
-        index_query(index_function, query, xapian_build_directory, i)
+        index_query(index_function, query, xapian_build_directory, sql_uri, i)
 
 
 @contextlib.contextmanager
@@ -352,17 +352,18 @@ def xapian_compact(combined_index: pathlib.Path, indices: List[pathlib.Path]) ->
 
 @click.command(help="Index GeneNetwork data and build Xapian search index in XAPIAN_DIRECTORY.")
 @click.argument("xapian_directory")
+@click.argument("sql_uri")
 # pylint: disable=missing-function-docstring
-def main(xapian_directory: str) -> None:
+def main(xapian_directory: str, sql_uri: str) -> None:
     logging.basicConfig(level=os.environ.get("LOGLEVEL", "DEBUG"),
                         format='%(relativeCreated)s: %(levelname)s: %(message)s')
     pathlib.Path(xapian_directory).mkdir(exist_ok=True)
     with temporary_directory("combined", xapian_directory) as combined_index:
         with temporary_directory("build", xapian_directory) as xapian_build_directory:
             logging.info("Indexing genes")
-            index_query(index_genes, genes_query, xapian_build_directory)
+            index_query(index_genes, genes_query, xapian_build_directory, sql_uri)
             logging.info("Indexing phenotypes")
-            index_query(index_phenotypes, phenotypes_query, xapian_build_directory)
+            index_query(index_phenotypes, phenotypes_query, xapian_build_directory, sql_uri)
             logging.info("Combining and compacting indices")
             xapian_compact(combined_index, list(xapian_build_directory.iterdir()))
         for child in combined_index.iterdir():