Merge pull request #3362 from lonvia/find-postcode-areas

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

diff --git a/nominatim/tools/exec_utils.py b/nominatim/tools/exec_utils.py
index b476c1231d130099bf1ca3ee321bcceaf52d382e..db89c38947d35a17f6e75a0964ea3b48a5aaa4ce 100644 (file)
--- a/nominatim/tools/exec_utils.py
+++ b/nominatim/tools/exec_utils.py
@@ -1,117 +1,51 @@
+# 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.
 """
 """
 Helper functions for executing external programs.
 """

+from typing import Any, Mapping, IO

 import logging
 import logging

+import os

 import subprocess
 import urllib.request as urlrequest
 import subprocess
 import urllib.request as urlrequest

-from urllib.parse import urlencode

 
 

-from ..version import NOMINATIM_VERSION
-from ..db.connection import get_pg_env
+from nominatim.typing import StrPath
+from nominatim.version import NOMINATIM_VERSION
+from nominatim.db.connection import get_pg_env

 
 LOG = logging.getLogger()
 
 
 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.
-
-        Returns the exit code of the script. If `throw_on_fail` is True
-        then throw a `CalledProcessError` on a non-zero exit.
-    """
-    cmd = ['/usr/bin/env', 'php', '-Cq',
-           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_SQLDIR'] = str(nominatim_env.sqllib_dir)
-    env['NOMINATIM_CONFIGDIR'] = str(nominatim_env.config_dir)
-    env['NOMINATIM_DATABASE_MODULE_SRC_PATH'] = nominatim_env.module_dir
-    if not env['NOMINATIM_OSM2PGSQL_BINARY']:
-        env['NOMINATIM_OSM2PGSQL_BINARY'] = nominatim_env.osm2pgsql_path
-
-    proc = subprocess.run(cmd, cwd=str(nominatim_env.project_dir), env=env,
-                          check=throw_on_fail)
-
-    return proc.returncode
-
-def run_api_script(endpoint, project_dir, extra_env=None, phpcgi_bin=None,
-                   params=None):
-    """ Execute a Nominiatim API function.
-
-        The function needs a project directory that contains the website
-        directory with the scripts to be executed. The scripts will be run
-        using php_cgi. Query parameters can be added as named arguments.
-
-        Returns the exit code of the script.
-    """
-    log = logging.getLogger()
-    webdir = str(project_dir / 'website')
-    query_string = urlencode(params or {})
-
-    env = dict(QUERY_STRING=query_string,
-               SCRIPT_NAME='/{}.php'.format(endpoint),
-               REQUEST_URI='/{}.php?{}'.format(endpoint, query_string),
-               CONTEXT_DOCUMENT_ROOT=webdir,
-               SCRIPT_FILENAME='{}/{}.php'.format(webdir, endpoint),
-               HTTP_HOST='localhost',
-               HTTP_USER_AGENT='nominatim-tool',
-               REMOTE_ADDR='0.0.0.0',
-               DOCUMENT_ROOT=webdir,
-               REQUEST_METHOD='GET',
-               SERVER_PROTOCOL='HTTP/1.1',
-               GATEWAY_INTERFACE='CGI/1.1',
-               REDIRECT_STATUS='CGI')
-
-    if extra_env:
-        env.update(extra_env)
-
-    if phpcgi_bin is None:
-        cmd = ['/usr/bin/env', 'php-cgi']
-    else:
-        cmd = [str(phpcgi_bin)]
-
-    proc = subprocess.run(cmd, cwd=str(project_dir), env=env, capture_output=True,
-                          check=False)
-
-    if proc.returncode != 0 or proc.stderr:
-        if proc.stderr:
-            log.error(proc.stderr.decode('utf-8').replace('\\n', '\n'))
-        else:
-            log.error(proc.stdout.decode('utf-8').replace('\\n', '\n'))
-        return proc.returncode or 1
-
-    result = proc.stdout.decode('utf-8')
-    content_start = result.find('\r\n\r\n')
-
-    print(result[content_start + 4:].replace('\\n', '\n'))
-
-    return 0
-
-
-def run_php_server(server_address, base_dir):
+def run_php_server(server_address: str, base_dir: StrPath) -> None:

     """ Run the built-in server from the given directory.
     """
     subprocess.run(['/usr/bin/env', 'php', '-S', server_address],
                    cwd=str(base_dir), check=True)
 
 
     """ 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):
+def run_osm2pgsql(options: Mapping[str, Any]) -> None:

     """ Run osm2pgsql with the given options.
     """
     env = get_pg_env(options['dsn'])
     """ Run osm2pgsql with the given options.
     """
     env = get_pg_env(options['dsn'])

-    cmd = [options['osm2pgsql'],
-           '--hstore', '--latlon', '--slim',
-           '--with-forward-dependencies', 'false',
+    cmd = [str(options['osm2pgsql']),
+           '--slim',

            '--log-progress', 'true',
            '--log-progress', 'true',

-           '--number-processes', str(options['threads']),
+           '--number-processes', '1' if options['append'] else str(options['threads']),

            '--cache', str(options['osm2pgsql_cache']),
            '--cache', str(options['osm2pgsql_cache']),

-           '--output', 'gazetteer',

            '--style', str(options['osm2pgsql_style'])
           ]
            '--style', str(options['osm2pgsql_style'])
           ]

-    if options['append']:
-        cmd.append('--append')
+
+    if str(options['osm2pgsql_style']).endswith('.lua'):
+        env['LUA_PATH'] = ';'.join((str(options['osm2pgsql_style_path'] / '?.lua'),
+                                    os.environ.get('LUAPATH', ';')))
+        cmd.extend(('--output', 'flex'))

     else:
     else:

-        cmd.append('--create')
+        cmd.extend(('--output', 'gazetteer', '--hstore', '--latlon'))
+
+    cmd.append('--append' if options['append'] else '--create')

 
     if options['flatnode_file']:
         cmd.extend(('--flat-nodes', options['flatnode_file']))
 
     if options['flatnode_file']:
         cmd.extend(('--flat-nodes', options['flatnode_file']))


@@ -123,22 +57,36 @@ def run_osm2pgsql(options):
         if options['tablespaces'][key]:
             cmd.extend((param, options['tablespaces'][key]))
 
         if options['tablespaces'][key]:
             cmd.extend((param, options['tablespaces'][key]))
 

+    if options['tablespaces']['main_data']:
+        env['NOMINATIM_TABLESPACE_PLACE_DATA'] = options['tablespaces']['main_data']
+    if options['tablespaces']['main_index']:
+        env['NOMINATIM_TABLESPACE_PLACE_INDEX'] = options['tablespaces']['main_index']
+

     if options.get('disable_jit', False):
         env['PGOPTIONS'] = '-c jit=off -c max_parallel_workers_per_gather=0'
 
     if options.get('disable_jit', False):
         env['PGOPTIONS'] = '-c jit=off -c max_parallel_workers_per_gather=0'
 

-    cmd.append(str(options['import_file']))
+    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', '.'), env=env, check=True)
+    subprocess.run(cmd, cwd=options.get('cwd', '.'),
+                   input=options.get('import_data'),
+                   env=env, check=True)

 
 
 
 

-def get_url(url):
+def get_url(url: str) -> str:

     """ Get the contents from the given URL and return it as a UTF-8 string.
     """
     """ Get the contents from the given URL and return it as a UTF-8 string.
     """

-    headers = {"User-Agent" : "Nominatim/" + NOMINATIM_VERSION}
+    headers = {"User-Agent": f"Nominatim/{NOMINATIM_VERSION!s}"}

 
     try:
 
     try:

-        with urlrequest.urlopen(urlrequest.Request(url, headers=headers)) as response:
+        request = urlrequest.Request(url, headers=headers)
+        with urlrequest.urlopen(request) as response: # type: IO[bytes]

             return response.read().decode('utf-8')
             return response.read().decode('utf-8')

-    except:
+    except Exception:

         LOG.fatal('Failed to load URL: %s', url)
         raise
         LOG.fatal('Failed to load URL: %s', url)
         raise