cli: add a new 'db' cli group with an 'init' subcommand

that allows to initialize the database for a given swh package with db connection credentials/dsn read from a standard swh config file.

Closes #1896 (closed).

---

Migrated from D1703 (view on Phabricator)

swh/core/cli/db.py

@@ -4,17 +4,95 @@
 # License: GNU General Public License version 3, or any later version
 # See top-level LICENSE file for more information
 
+import glob
 import logging
+from os import path
+import subprocess
 import warnings
+
 warnings.filterwarnings("ignore")  # noqa prevent psycopg from telling us sh*t
 
 import click
 
 from swh.core.cli import CONTEXT_SETTINGS
+from swh.core.config import read as config_read
 
 logger = logging.getLogger(__name__)
 
 
+@click.group(name="db", context_settings=CONTEXT_SETTINGS)
+@click.option("--config-file", "-C", default=None,
+              type=click.Path(exists=True, dir_okay=False),
+              help="Configuration file.")
+@click.pass_context
+def db(ctx, config_file):
+    """Software Heritage database generic tools.
+    """
+    ctx.ensure_object(dict)
+    cfg = config_read(config_file)
+    ctx.obj["config"] = cfg
+
+
+@db.command(name="init", context_settings=CONTEXT_SETTINGS)
+@click.pass_context
+def init(ctx):
+    """Initialize the database for every Software Heritage module found in the
+    configuration file. For every configuration section in the config file
+    that:
+
+    1. has the name of an existing swh package,
+    2. has credentials for a local db access,
+
+    it will run the initialization scripts from the swh package against the
+    given database.
+
+    Example for the config file::
+
+    \b
+      storage:
+        cls: local
+        args:
+          db: postgresql:///?service=swh-storage
+      objstorage:
+        cls: remote
+        args:
+          url: http://swh-objstorage:5003/
+
+    the command:
+
+      swh db -C /path/to/config.yml init
+
+    will initialize the database for the `storage` section using initialization
+    scripts from the `swh.storage` package.
+    """
+
+    for modname, cfg in ctx.obj["config"].items():
+        if cfg.get("cls") == "local" and cfg.get("args"):
+            try:
+                sqlfiles = get_sql_for_package(modname)
+            except click.BadParameter:
+                logger.info(
+                    "Failed to load/find sql initialization files for %s",
+                    modname)
+
+            if sqlfiles:
+                conninfo = cfg["args"]["db"]
+                for sqlfile in sqlfiles:
+                    subprocess.call_call(
+                        [
+                            "psql",
+                            "--quiet",
+                            "--no-psqlrc",
+                            "-v",
+                            "ON_ERROR_STOP=1",
+                            "-d",
+                            conninfo,
+                            "-f",
+                            sqlfile,
+                        ]
+                    )
+
+
 @click.command(context_settings=CONTEXT_SETTINGS)
 @click.argument('module', nargs=-1, required=True)
 @click.option('--db-name', '-d', help='Database name.',
@@ -41,10 +119,6 @@ def db_init(module, db_name, create_db):
     """
     # put import statements here so we can keep startup time of the main swh
     # command as short as possible
-    from os import path
-    import glob
-    from importlib import import_module
-    from swh.core.utils import numfile_sortkey as sortkey
     from swh.core.db.tests.db_testing import (
         pg_createdb, pg_restore, DB_DUMP_TYPES,
         swh_db_version
@@ -54,21 +128,7 @@ def db_init(module, db_name, create_db):
     dump_files = []
 
     for modname in module:
-        if not modname.startswith('swh.'):
-            modname = 'swh.{}'.format(modname)
-        try:
-            m = import_module(modname)
-        except ImportError:
-            raise click.BadParameter(
-                'Unable to load module {}'.format(modname))
-
-        sqldir = path.join(path.dirname(m.__file__), 'sql')
-        if not path.isdir(sqldir):
-            raise click.BadParameter(
-                'Module {} does not provide a db schema '
-                '(no sql/ dir)'.format(modname))
-        dump_files.extend(sorted(glob.glob(path.join(sqldir, '*.sql')),
-                                 key=sortkey))
+        dump_files.extend(get_sql_for_package(modname))
 
     if create_db:
         # Create the db (or fail silently if already existing)
@@ -89,3 +149,22 @@ def db_init(module, db_name, create_db):
 
     click.secho('DONE database is {} version {}'.format(db_name, db_version),
                 fg='green', bold=True)
+
+
+def get_sql_for_package(modname):
+    from importlib import import_module
+    from swh.core.utils import numfile_sortkey as sortkey
+
+    if not modname.startswith("swh."):
+        modname = "swh.{}".format(modname)
+    try:
+        m = import_module(modname)
+    except ImportError:
+        raise click.BadParameter("Unable to load module {}".format(modname))
+
+    sqldir = path.join(path.dirname(m.__file__), "sql")
+    if not path.isdir(sqldir):
+        raise click.BadParameter(
+            "Module {} does not provide a db schema "
+            "(no sql/ dir)".format(modname))
+    return list(sorted(glob.glob(path.join(sqldir, "*.sql")), key=sortkey))