Merge remote-tracking branch 'upstream/master'

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

diff --git a/nominatim/tools/refresh.py b/nominatim/tools/refresh.py
index 885caca51273438dce4377fb9b86e5d8f06b1838..1fcb1577302d1fcca188683426ce6a1dfa48efc1 100644 (file)
--- a/nominatim/tools/refresh.py
+++ b/nominatim/tools/refresh.py
@@ -2,22 +2,23 @@
 Functions for bringing auxiliary data in the database up-to-date.
 """
 import json
+import re
 
 from psycopg2.extras import execute_values
 
 from ..db.utils import execute_file
 
-def update_postcodes(conn, datadir):
+def update_postcodes(conn, sql_dir):
     """ Recalculate postcode centroids and add, remove and update entries in the
         location_postcode table. `conn` is an opne connection to the database.
     """
-    execute_file(conn, datadir / 'sql' / 'update-postcodes.sql')
+    execute_file(conn, sql_dir / 'update-postcodes.sql')
 
 
-def recompute_word_counts(conn, datadir):
+def recompute_word_counts(conn, sql_dir):
     """ Compute the frequency of full-word search terms.
     """
-    execute_file(conn, datadir / 'sql' / 'words_from_search_name.sql')
+    execute_file(conn, sql_dir / 'words_from_search_name.sql')
 
 
 def _add_address_level_rows_from_entry(rows, entry):
@@ -69,3 +70,98 @@ def load_address_levels_from_file(conn, config_file):
     """
     with config_file.open('r') as fdesc:
         load_address_levels(conn, 'address_levels', json.load(fdesc))
+
+PLPGSQL_BASE_MODULES = (
+    'utils.sql',
+    'normalization.sql',
+    'ranking.sql',
+    'importance.sql',
+    'address_lookup.sql',
+    'interpolation.sql'
+)
+
+PLPGSQL_TABLE_MODULES = (
+    ('place', 'place_triggers.sql'),
+    ('placex', 'placex_triggers.sql'),
+    ('location_postcode', 'postcode_triggers.sql')
+)
+
+def _get_standard_function_sql(conn, config, sql_dir, enable_diff_updates, enable_debug):
+    """ Read all applicable SQLs containing PL/pgSQL functions, replace
+        placefolders and execute them.
+    """
+    sql_func_dir = sql_dir / 'functions'
+    sql = ''
+
+    # Get the basic set of functions that is always imported.
+    for sql_file in PLPGSQL_BASE_MODULES:
+        with (sql_func_dir / sql_file).open('r') as fdesc:
+            sql += fdesc.read()
+
+    # Some files require the presence of a certain table
+    for table, fname in PLPGSQL_TABLE_MODULES:
+        if conn.table_exists(table):
+            with (sql_func_dir / fname).open('r') as fdesc:
+                sql += fdesc.read()
+
+    # Replace placeholders.
+    sql = sql.replace('{modulepath}',
+                      config.DATABASE_MODULE_PATH or str((config.project_dir / 'module').resolve()))
+
+    if enable_diff_updates:
+        sql = sql.replace('RETURN NEW; -- %DIFFUPDATES%', '--')
+
+    if enable_debug:
+        sql = sql.replace('--DEBUG:', '')
+
+    if config.get_bool('LIMIT_REINDEXING'):
+        sql = sql.replace('--LIMIT INDEXING:', '')
+
+    if not config.get_bool('USE_US_TIGER_DATA'):
+        sql = sql.replace('-- %NOTIGERDATA% ', '')
+
+    if not config.get_bool('USE_AUX_LOCATION_DATA'):
+        sql = sql.replace('-- %NOAUXDATA% ', '')
+
+    reverse_only = 'false' if conn.table_exists('search_name') else 'true'
+
+    return sql.replace('%REVERSE-ONLY%', reverse_only)
+
+
+def replace_partition_string(sql, partitions):
+    """ Replace a partition template with the actual partition code.
+    """
+    for match in re.findall('^-- start(.*?)^-- end', sql, re.M | re.S):
+        repl = ''
+        for part in partitions:
+            repl += match.replace('-partition-', str(part))
+        sql = sql.replace(match, repl)
+
+    return sql
+
+def _get_partition_function_sql(conn, sql_dir):
+    """ Create functions that work on partition tables.
+    """
+    with conn.cursor() as cur:
+        cur.execute('SELECT distinct partition FROM country_name')
+        partitions = set([0])
+        for row in cur:
+            partitions.add(row[0])
+
+    with (sql_dir / 'partition-functions.src.sql').open('r') as fdesc:
+        sql = fdesc.read()
+
+    return replace_partition_string(sql, sorted(partitions))
+
+def create_functions(conn, config, sql_dir,
+                     enable_diff_updates=True, enable_debug=False):
+    """ (Re)create the PL/pgSQL functions.
+    """
+    sql = _get_standard_function_sql(conn, config, sql_dir,
+                                     enable_diff_updates, enable_debug)
+    sql += _get_partition_function_sql(conn, sql_dir)
+
+    with conn.cursor() as cur:
+        cur.execute(sql)
+
+    conn.commit()