[testing] Add delta tests for map data

docs/MAPS.md

@@ -136,6 +136,8 @@ it can fail due to low memory.
 * `MERGE_INTERVAL`: delay in minutes between attempts to merge a coast line.
 Default is 40.
 * `REGIONS`: a list of `.poly` files for regions to be built. One for each line.
+* `DELTA_WITH`: a path to an older map directory, to compare with the freshly
+generated data in the testing step.
 Can be empty. Example: `$(ls ../../data/borders/{UK*,Ireland}.poly)`.
 
 ### Testing

tools/unix/diff_features.py

@@ -0,0 +1,34 @@
+#!/usr/bin/python
+import sys, re
+
+RE_STAT = re.compile(r'(?:\d+\. )?([\w:|-]+?)\|: size = (\d+); count = (\d+); length = ([0-9.e+-]+) m; area = ([0-9.e+-]+) m.\s*')
+def parse_and_add(data, line):
+  m = RE_STAT.match(line)
+  if m:
+    data[m.group(1)] = int(m.group(3))
+
+if len(sys.argv) < 3:
+  print 'This tool compares type_statistics output for feature sizes'
+  print 'Usage: {0} <output_new> <output_old> [threshold_in_%]'.format(sys.argv[0])
+  sys.exit(0)
+
+data1 = {}
+with open(sys.argv[2], 'r') as f:
+  for line in f:
+    parse_and_add(data1, line)
+data2 = {}
+with open(sys.argv[1], 'r') as f:
+  for line in f:
+    parse_and_add(data2, line)
+
+threshold = (int(sys.argv[3]) if len(sys.argv) > 3 else 50) / 100.0 + 1
+min_diff = 40
+
+for k in data1:
+  v1 = int(data1[k])
+  if k in data2:
+    v2 = int(data2[k])
+    if v1 == 0 or v2 == 0 or max(v1, v2) / float(min(v1, v2)) > threshold and abs(v1 - v2) > min_diff:
+      print '{0}: {1} to {2}'.format(k, v1, v2)
+  elif v1 > min_diff:
+    print '- not found: {0}, {1}'.format(k, v1)

tools/unix/diff_size.py

@@ -0,0 +1,24 @@
+#!/usr/bin/python
+import os, sys
+
+if len(sys.argv) < 3:
+  print 'This tool shows very different file sizes'
+  print 'Usage: {0} <newdir> <olddir> [threshold_in_%]'.format(sys.argv[0])
+  sys.exit(0)
+
+new_path = sys.argv[1]
+old_path = sys.argv[2]
+threshold = (int(sys.argv[3]) if len(sys.argv) > 3 else 10) / 100.0 + 1
+min_diff = 1024 * 1024
+
+for f in os.listdir(new_path):
+  new_file = os.path.join(new_path, f)
+  old_file = os.path.join(old_path, f)
+  if os.path.isfile(new_file) and os.path.isfile(old_file):
+    new_size = os.path.getsize(new_file)
+    old_size = os.path.getsize(old_file)
+    if new_size + old_size > 0:
+      if new_size == 0 or old_size == 0 or max(new_size, old_size) / float(min(new_size, old_size)) > threshold and abs(new_size - old_size) > min_diff:
+          print '{0}: {1} to {2} MB'.format(f, old_size / 1024 / 1024, new_size / 1024 / 1024)
+  else:
+    print 'Not found a mirror for {0}'.format(f)

tools/unix/generate_planet.sh

@@ -424,7 +424,7 @@ fi
 if [ "$MODE" == "test" ]; then
   putmode "Step 8: Testing data"
   TEST_LOG="$LOG_PATH/test_planet.log"
-  bash "$TESTING_SCRIPT" "$TARGET" > "$TEST_LOG"
+  bash "$TESTING_SCRIPT" "$TARGET" "${DELTA_WITH-}" > "$TEST_LOG"
   # Send both log files via e-mail
   if [ -n "${MAIL-}" ]; then
     cat <(grep STATUS "$PLANET_LOG") <(echo ---------------) "$TEST_LOG" | mailx -s "Generate_planet: build completed at $(hostname)" "$MAIL"

tools/unix/test_planet.sh

@@ -6,7 +6,7 @@
 if [ $# -eq 0 ]; then
   echo
   echo "This script analyzes a generate_planet.sh run and prints all issues."
-  echo "Usage: $0 <target_dir>"
+  echo "Usage: $0 <target_dir> [<old_maps_dir>]"
   echo
   exit 1
 fi
@@ -17,6 +17,8 @@ OMIM_PATH="${OMIM_PATH:-$(cd "$(dirname "$0")/../.."; pwd)}"
 TARGET="${TARGET:-$1}"
 LOG_PATH="${LOG_PATH:-$TARGET/logs}"
 PLANET_LOG="$LOG_PATH/generate_planet.log"
+DELTA_WITH=
+[ $# -gt 1 -a -d "${2-}" ] && DELTA_WITH="$2"
 
 source "$(dirname "$0")/find_generator_tool.sh"
 
@@ -54,6 +56,13 @@ if [ -n "$(ls "$TARGET" | grep '\.mwm\.routing')" ]; then
   done
 fi
 
+# Step 2.5: compare new files sizes with old
+if [ -n "$DELTA_WITH" ]; then
+  echo
+  echo "### SIZE DIFFERENCE WITH $DELTA_WITH"
+  python "$(dirname "$0")/diff_size.py" "$TARGET" "$DELTA_WITH" 5
+fi
+
 # For generator_tool, we create a temporary directory with symlinks to all maps
 # That way, it can be easily cleaned after routing engine creates a lot of temporary directories in it
 FTARGET="$TARGET/symlinked_copy"
@@ -78,6 +87,27 @@ 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
+if [ -n "$DELTA_WITH" ]; then
+  echo
+  echo '### FEATURE DIFFERENCE'
+  TMPBASE="$HOME/test_planet_tmp"
+  for mwm in "$FTARGET"/*.mwm; do
+    BASENAME="$(basename "$mwm" .mwm)"
+    if [ -f "$DELTA_WITH/$BASENAME.mwm" ]; then
+      "$GENERATOR_TOOL" --data_path="$FTARGET"    --user_resource_path="$OMIM_PATH/data/" --output="$BASENAME" --type_statistics >"${TMPBASE}_new" 2>/dev/null
+      "$GENERATOR_TOOL" --data_path="$DELTA_WITH" --user_resource_path="$OMIM_PATH/data/" --output="$BASENAME" --type_statistics >"${TMPBASE}_old" 2>/dev/null
+      DIFFERENCE="$(python "$(dirname "$0")/diff_features.py" "${TMPBASE}_new" "${TMPBASE}_old" 50)"
+      if [ -n "$DIFFERENCE" ]; then
+        echo
+        echo "$BASENAME"
+        echo "$DIFFERENCE"
+      fi
+    fi
+  done
+  rm "$TMPBASE"_*
+fi
+
 # Step 4: run intergation tests
 echo
 echo '### INTEGRATION TESTS'