[python] Add booking hotels count test

 - Review fixes
 - Rename steps
 - Fix format

2 files changed, 81 insertions, 5 deletions

diff --git a/tools/python/mwm/mwm_feature_compare.py b/tools/python/mwm/mwm_feature_compare.py
new file mode 100755
index 0000000000..82787bdadf
--- /dev/null
+++ b/tools/python/mwm/mwm_feature_compare.py
@@ -0,0 +1,68 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+from __future__ import print_function
+from mwm import MWM
+
+import argparse
+import os
+import multiprocessing
+
+
+OMIM_ROOT = os.path.join(os.path.dirname(os.path.abspath(__file__)), '..', '..', '..')
+
+
+def count_feature(mwm_path, feature_name):
+    mwm = MWM(open(mwm_path, 'rb'))
+    mwm.read_header()
+    mwm.read_types(os.path.join(OMIM_ROOT, 'data', 'types.txt'))
+    counter = 0
+    for feature in mwm.iter_features():
+        if feature_name in feature['header']['types']:
+            counter += 1
+    return counter
+
+
+def compare_feature_num(args_tuple):
+    old_mwm, new_mwm, feature_name, threshold = args_tuple
+    old_feature_count = count_feature(old_mwm, feature_name)
+    new_feature_count = count_feature(new_mwm, feature_name)
+    delta = new_feature_count - old_feature_count
+
+    if delta < 0:
+        p_change = float(abs(delta)) / old_feature_count * 100
+
+        if p_change > threshold:
+            print('In \'{0}\' number of \'{1}\' decreased by {2:.0f}% ({3} → {4})'.format(
+                os.path.basename(new_mwm), feature_name, round(p_change), old_feature_count, new_feature_count))
+            return False
+    return True
+
+
+def compare_mwm(old_mwm_path, new_mwm_path, feature_name, threshold):
+    def valid_mwm(mwm_name):
+        return mwm_name.endswith('.mwm') and not mwm_name.startswith('World')
+
+    def generate_names_dict(path):
+        return dict((file_name, os.path.abspath(os.path.join(path, file_name)))
+                    for file_name in os.listdir(path) if valid_mwm(file_name))
+
+    old_mwm_list = generate_names_dict(old_mwm_path)
+    new_mwm_list = generate_names_dict(new_mwm_path)
+
+    same_mwm_names = set(new_mwm_list).intersection(set(old_mwm_list))
+    args = ((old_mwm_list[mwm], new_mwm_list[mwm], feature_name, threshold) for mwm in same_mwm_names)
+
+    pool = multiprocessing.Pool()
+    return all(pool.imap(compare_feature_num, args))
+
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser(description='Script to compare feature count in \'.mwm\' files')
+    parser.add_argument('-n', '--new', help='New mwm files path', type=str, required=True)
+    parser.add_argument('-o', '--old', help='Old mwm files path', type=str, required=True)
+    parser.add_argument('-f', '--feature', help='Feature name to count', type=str, required=True)
+    parser.add_argument('-t', '--threshold', help='Threshold in percent to warn', type=int, default=20)
+
+    args = parser.parse_args()
+    if not compare_mwm(args.old, args.new, args.feature, args.threshold):
+        print('Warning: some .mwm files lost more than {}% booking hotels'.format(args.threshold))
diff --git a/tools/unix/test_planet.sh b/tools/unix/test_planet.sh
index 8e8150f24e..b9db599aa9 100755
--- a/tools/unix/test_planet.sh
+++ b/tools/unix/test_planet.sh
@@ -19,6 +19,7 @@ TARGET="$(cd "${TARGET:-$1}"; pwd)"
 LOG_PATH="${LOG_PATH:-$TARGET/logs}"
 PLANET_LOG="$LOG_PATH/generate_planet.log"
 DELTA_WITH=
+BOOKING_THRESHOLD=20
 [ $# -gt 1 -a -d "${2-}" ] && DELTA_WITH="$2"
 
 source "$SCRIPT_PATH/find_generator_tool.sh"
@@ -39,7 +40,7 @@ for log in "$LOG_PATH"/*.log; do
   fi
 done
 
-# Step 2: test if mwms and routing were made
+# Step 2.1: test if mwms and routing were made
 echo
 echo '### MISSING FILES'
 # Missing MWM files can be derived only from intermediate borders
@@ -50,7 +51,7 @@ if [ -d "$TARGET/borders" ]; then
   done
 fi
 
-# Step 2.5: compare new files sizes with old
+# Step 2.2: compare new files sizes with old
 if [ -n "$DELTA_WITH" ]; then
   echo
   echo "### SIZE DIFFERENCE WITH $DELTA_WITH"
@@ -69,7 +70,7 @@ for file in "$TARGET"/*.mwm*; do
   ln -s "$TARGET/$BASENAME" "$FTARGET/$BASENAME"
 done
 
-# Step 3: run calc_statistics and check for sections
+# Step 3.1: run calc_statistics and check for sections
 echo
 echo '### MISSING MWM SECTIONS'
 FOUND_COASTS=
@@ -83,7 +84,7 @@ done
 
 [ -z "$FOUND_COASTS" ] && echo && echo 'WARNING: Did not find any coastlines in MWM files'
 
-# Step 3.5: run type_statistics for old and new files to compare
+# Step 3.2: run type_statistics for old and new files to compare
 if [ -n "$DELTA_WITH" ]; then
   echo
   echo '### FEATURE DIFFERENCE'
@@ -104,7 +105,14 @@ if [ -n "$DELTA_WITH" ]; then
   rm "$TMPBASE"_*
 fi
 
-# Step 4: run intergation tests
+# Step 3.3: check booking hotels count in new .mwm files
+if [ -n "$DELTA_WITH" ]; then
+  echo
+  echo '### BOOKING HOTELS COUNT DIFFERENCE'
+  python "$OMIM_PATH/tools/python/mwm/mwm_feature_compare.py" -n "$TARGET" -o "$DELTA_WITH" -f "sponsored-booking" -t $BOOKING_THRESHOLD
+fi
+
+# Step 4: run integration tests
 echo
 echo '### INTEGRATION TESTS'
 "$(dirname "$GENERATOR_TOOL")/routing_integration_tests" "--data_path=$FTARGET/../" "--user_resource_path=$OMIM_PATH/data/" "--suppress=online_cross_tests.*" 2>&1