Merge remote-tracking branch 'upstream/master'

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

diff --git a/nominatim/tools/tiger_data.py b/nominatim/tools/tiger_data.py
index 521d11c4b1f28a40303e3b183227d11b3656fe70..70cecae58648997f0bc6a81a863fee7b420432f5 100644 (file)
--- a/nominatim/tools/tiger_data.py
+++ b/nominatim/tools/tiger_data.py
@@ -1,96 +1,149 @@
+# 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.

 """
 """

-Functions for setting up and importing a new Nominatim database.
+Functions for importing tiger data and handling tarbar and directory files

 """
 """

+from typing import Any, TextIO, List, Union, cast
+import csv
+import io

 import logging
 import os
 import logging
 import os

-import time

 import tarfile
 import tarfile

-import selectors

 
 

-from ..db.connection import connect
-from ..db.async_connection import DBConnection
-from ..db.sql_preprocessor import SQLPreprocessor
+from psycopg2.extras import Json
+
+from nominatim.config import Configuration
+from nominatim.db.connection import connect
+from nominatim.db.async_connection import WorkerPool
+from nominatim.db.sql_preprocessor import SQLPreprocessor
+from nominatim.errors import UsageError
+from nominatim.data.place_info import PlaceInfo
+from nominatim.tokenizer.base import AbstractAnalyzer, AbstractTokenizer
+from nominatim.tools import freeze

 
 LOG = logging.getLogger()
 
 
 LOG = logging.getLogger()
 

+class TigerInput:
+    """ Context manager that goes through Tiger input files which may
+        either be in a directory or gzipped together in a tar file.
+    """
+
+    def __init__(self, data_dir: str) -> None:
+        self.tar_handle = None
+        self.files: List[Union[str, tarfile.TarInfo]] = []
+
+        if data_dir.endswith('.tar.gz'):
+            try:
+                self.tar_handle = tarfile.open(data_dir) # pylint: disable=consider-using-with
+            except tarfile.ReadError as err:
+                LOG.fatal("Cannot open '%s'. Is this a tar file?", data_dir)
+                raise UsageError("Cannot open Tiger data file.") from err
+
+            self.files = [i for i in self.tar_handle.getmembers() if i.name.endswith('.csv')]
+            LOG.warning("Found %d CSV files in tarfile with path %s", len(self.files), data_dir)
+        else:
+            files = os.listdir(data_dir)
+            self.files = [os.path.join(data_dir, i) for i in files if i.endswith('.csv')]
+            LOG.warning("Found %d CSV files in path %s", len(self.files), data_dir)
+
+        if not self.files:
+            LOG.warning("Tiger data import selected but no files found at %s", data_dir)
+
+
+    def __enter__(self) -> 'TigerInput':
+        return self
+
+
+    def __exit__(self, exc_type: Any, exc_val: Any, exc_tb: Any) -> None:
+        if self.tar_handle:
+            self.tar_handle.close()
+            self.tar_handle = None
+
+
+    def next_file(self) -> TextIO:
+        """ Return a file handle to the next file to be processed.
+            Raises an IndexError if there is no file left.
+        """
+        fname = self.files.pop(0)
+
+        if self.tar_handle is not None:
+            extracted = self.tar_handle.extractfile(fname)
+            assert extracted is not None
+            return io.TextIOWrapper(extracted)
+
+        return open(cast(str, fname), encoding='utf-8')
+
+
+    def __len__(self) -> int:
+        return len(self.files)

 
 

-def add_tiger_data(dsn, data_dir, threads, config, sqllib_dir):
-    """ Import tiger data from directory or tar file
+
+def handle_threaded_sql_statements(pool: WorkerPool, fd: TextIO,
+                                   analyzer: AbstractAnalyzer) -> None:
+    """ Handles sql statement with multiplexing

     """
     """

-    # Handling directory or tarball file.
-    is_tarfile = False
-    if(data_dir.endswith('.tar.gz')):
-        is_tarfile = True
-        tar = tarfile.open(data_dir)
-        sql_files = [i for i in tar.getmembers() if i.name.endswith('.sql')]
-        LOG.warning(f'Found {len(sql_files)} SQL files in tarfile with path {data_dir}')
-        if(not len(sql_files)):
-            LOG.warning(f'Tiger data import selected but no files found in tarfile with path {data_dir}')
-            return
-    else:
-        files = os.listdir(data_dir)
-        sql_files = [i for i in files if i.endswith('.sql')]
-        LOG.warning(f'Found {len(sql_files)} SQL files in path {data_dir}')
-        if(not len(sql_files)):
-            LOG.warning(f'Tiger data import selected but no files found in path {data_dir}')
-            return
-    
+    lines = 0
+    # Using pool of database connections to execute sql statements
+
+    sql = "SELECT tiger_line_import(%s, %s, %s, %s, %s, %s)"
+
+    for row in csv.DictReader(fd, delimiter=';'):
+        try:
+            address = dict(street=row['street'], postcode=row['postcode'])
+            args = ('SRID=4326;' + row['geometry'],
+                    int(row['from']), int(row['to']), row['interpolation'],
+                    Json(analyzer.process_place(PlaceInfo({'address': address}))),
+                    analyzer.normalize_postcode(row['postcode']))
+        except ValueError:
+            continue
+        pool.next_free_worker().perform(sql, args=args)
+
+        lines += 1
+        if lines == 1000:
+            print('.', end='', flush=True)
+            lines = 0
+
+
+def add_tiger_data(data_dir: str, config: Configuration, threads: int,
+                   tokenizer: AbstractTokenizer) -> int:
+    """ Import tiger data from directory or tar file `data dir`.
+    """
+    dsn = config.get_libpq_dsn()
+

     with connect(dsn) as conn:
     with connect(dsn) as conn:

-        sql = SQLPreprocessor(conn, config, sqllib_dir)
-        sql.run_sql_file(conn, 'tiger_import_start.sql')
-
-    # Reading sql_files and then for each file line handling
-    # sql_query in <threads - 1> chunks.
-    sel = selectors.DefaultSelector()
-    place_threads = max(1, threads - 1)
-    for sql_file in sql_files:
-        if(not is_tarfile):
-            file_path = os.path.join(data_dir, sql_file)
-            file = open(file_path)
-        else:
-            file = tar.extractfile(sql_file)
-        lines = 0
-        end_of_file = False
-        total_used_threads = place_threads
-        while(True):
-            if(end_of_file):
-                break
-            for imod in range(place_threads):
-                conn = DBConnection(dsn)
-                conn.connect()
-
-                sql_query = file.readline()
-                lines+=1
-
-                if(not sql_query):
-                    end_of_file = True
-                    total_used_threads = imod
-                    break
-
-                conn.perform(sql_query)
-                sel.register(conn, selectors.EVENT_READ, conn)
-
-                if(lines==1000):
-                    print('. ', end='', flush=True)
-                    lines=0
-
-            todo = min(place_threads,total_used_threads)
-            while todo > 0:
-                for key, _ in sel.select(1):
-                    try:
-                        conn = key.data
-                        sel.unregister(conn)
-                        conn.wait()
-                        conn.close()
-                        todo -= 1
-                    except:
-                        todo -=1
-
-    if(is_tarfile):
-        tar.close()
-    print('\n')
+        is_frozen = freeze.is_frozen(conn)
+        conn.close()
+
+        if is_frozen:
+            raise UsageError("Tiger cannot be imported when database frozen (Github issue #3048)")
+
+    with TigerInput(data_dir) as tar:
+        if not tar:
+            return 1
+
+        with connect(dsn) as conn:
+            sql = SQLPreprocessor(conn, config)
+            sql.run_sql_file(conn, 'tiger_import_start.sql')
+
+        # Reading files and then for each file line handling
+        # sql_query in <threads - 1> chunks.
+        place_threads = max(1, threads - 1)
+
+        with WorkerPool(dsn, place_threads, ignore_sql_errors=True) as pool:
+            with tokenizer.name_analyzer() as analyzer:
+                while tar:
+                    with tar.next_file() as fd:
+                        handle_threaded_sql_statements(pool, fd, analyzer)
+
+        print('\n')
+

     LOG.warning("Creating indexes on Tiger data")
     with connect(dsn) as conn:
     LOG.warning("Creating indexes on Tiger data")
     with connect(dsn) as conn:

-        sql = SQLPreprocessor(conn, config, sqllib_dir)
+        sql = SQLPreprocessor(conn, config)

         sql.run_sql_file(conn, 'tiger_import_finish.sql')
         sql.run_sql_file(conn, 'tiger_import_finish.sql')

-    
\ No newline at end of file
+
+    return 0