icu: move token deduplication into TokenInfo

[nominatim.git] / nominatim / tools / exec_utils.py

diff --git a/nominatim/tools/exec_utils.py b/nominatim/tools/exec_utils.py
index 9e16e29382473694d8ba7de193eba4cce36cf44e..10d8fbc08acbf398f5d0a28c799395b1d20b2518 100644 (file)
--- a/nominatim/tools/exec_utils.py
+++ b/nominatim/tools/exec_utils.py
@@ -1,10 +1,22 @@
+# SPDX-License-Identifier: GPL-2.0-only
+#
+# This file is part of Nominatim. (https://nominatim.org)
+#
+# Copyright (C) 2022 by the Nominatim developer community.
+# For a full list of authors see the git log.
 """
 Helper functions for executing external programs.
 """
 import logging
 import subprocess
+import urllib.request as urlrequest
 from urllib.parse import urlencode
 
+from nominatim.version import NOMINATIM_VERSION
+from nominatim.db.connection import get_pg_env
+
+LOG = logging.getLogger()
+
 def run_legacy_script(script, *args, nominatim_env=None, throw_on_fail=False):
     """ Run a Nominatim PHP script with the given arguments.
 
@@ -12,16 +24,16 @@ def run_legacy_script(script, *args, nominatim_env=None, throw_on_fail=False):
         then throw a `CalledProcessError` on a non-zero exit.
     """
     cmd = ['/usr/bin/env', 'php', '-Cq',
-           nominatim_env.phplib_dir / 'admin' / script]
+           str(nominatim_env.phplib_dir / 'admin' / script)]
     cmd.extend([str(a) for a in args])
 
     env = nominatim_env.config.get_os_env()
     env['NOMINATIM_DATADIR'] = str(nominatim_env.data_dir)
-    env['NOMINATIM_BINDIR'] = str(nominatim_env.data_dir / 'utils')
-    if not env['NOMINATIM_DATABASE_MODULE_PATH']:
-        env['NOMINATIM_DATABASE_MODULE_PATH'] = nominatim_env.module_dir
+    env['NOMINATIM_SQLDIR'] = str(nominatim_env.sqllib_dir)
+    env['NOMINATIM_CONFIGDIR'] = str(nominatim_env.config_dir)
+    env['NOMINATIM_DATABASE_MODULE_SRC_PATH'] = str(nominatim_env.module_dir)
     if not env['NOMINATIM_OSM2PGSQL_BINARY']:
-        env['NOMINATIM_OSM2PGSQL_BINARY'] = nominatim_env.osm2pgsql_path
+        env['NOMINATIM_OSM2PGSQL_BINARY'] = str(nominatim_env.osm2pgsql_path)
 
     proc = subprocess.run(cmd, cwd=str(nominatim_env.project_dir), env=env,
                           check=throw_on_fail)
@@ -64,7 +76,9 @@ def run_api_script(endpoint, project_dir, extra_env=None, phpcgi_bin=None,
     else:
         cmd = [str(phpcgi_bin)]
 
-    proc = subprocess.run(cmd, cwd=str(project_dir), env=env, capture_output=True,
+    proc = subprocess.run(cmd, cwd=str(project_dir), env=env,
+                          stdout=subprocess.PIPE,
+                          stderr=subprocess.PIPE,
                           check=False)
 
     if proc.returncode != 0 or proc.stderr:
@@ -80,3 +94,67 @@ def run_api_script(endpoint, project_dir, extra_env=None, phpcgi_bin=None,
     print(result[content_start + 4:].replace('\\n', '\n'))
 
     return 0
+
+
+def run_php_server(server_address, base_dir):
+    """ Run the built-in server from the given directory.
+    """
+    subprocess.run(['/usr/bin/env', 'php', '-S', server_address],
+                   cwd=str(base_dir), check=True)
+
+
+def run_osm2pgsql(options):
+    """ Run osm2pgsql with the given options.
+    """
+    env = get_pg_env(options['dsn'])
+    cmd = [str(options['osm2pgsql']),
+           '--hstore', '--latlon', '--slim',
+           '--with-forward-dependencies', 'false',
+           '--log-progress', 'true',
+           '--number-processes', str(options['threads']),
+           '--cache', str(options['osm2pgsql_cache']),
+           '--output', 'gazetteer',
+           '--style', str(options['osm2pgsql_style'])
+          ]
+    if options['append']:
+        cmd.append('--append')
+    else:
+        cmd.append('--create')
+
+    if options['flatnode_file']:
+        cmd.extend(('--flat-nodes', options['flatnode_file']))
+
+    for key, param in (('slim_data', '--tablespace-slim-data'),
+                       ('slim_index', '--tablespace-slim-index'),
+                       ('main_data', '--tablespace-main-data'),
+                       ('main_index', '--tablespace-main-index')):
+        if options['tablespaces'][key]:
+            cmd.extend((param, options['tablespaces'][key]))
+
+    if options.get('disable_jit', False):
+        env['PGOPTIONS'] = '-c jit=off -c max_parallel_workers_per_gather=0'
+
+    if 'import_data' in options:
+        cmd.extend(('-r', 'xml', '-'))
+    elif isinstance(options['import_file'], list):
+        for fname in options['import_file']:
+            cmd.append(str(fname))
+    else:
+        cmd.append(str(options['import_file']))
+
+    subprocess.run(cmd, cwd=options.get('cwd', '.'),
+                   input=options.get('import_data'),
+                   env=env, check=True)
+
+
+def get_url(url):
+    """ Get the contents from the given URL and return it as a UTF-8 string.
+    """
+    headers = {"User-Agent": "Nominatim/{0[0]}.{0[1]}.{0[2]}-{0[3]}".format(NOMINATIM_VERSION)}
+
+    try:
+        with urlrequest.urlopen(urlrequest.Request(url, headers=headers)) as response:
+            return response.read().decode('utf-8')
+    except Exception:
+        LOG.fatal('Failed to load URL: %s', url)
+        raise