Moved statistics handler over from Osmium

The statistics handler was kept in the Osmium distribution, but it makes more
sense here.

4 files changed, 283 insertions, 8 deletions

diff --git a/tagstats/Makefile b/tagstats/Makefile
index 1543335..be6cfb1 100644
--- a/tagstats/Makefile
+++ b/tagstats/Makefile
@@ -33,7 +33,7 @@ LIB_PROTOBUF = -lz -lprotobuf-lite -losmpbf
 
 all: tagstats
 
-tagstats: tagstats.cpp tagstats_handler.hpp string_store.hpp geodistribution.hpp
+tagstats: tagstats.cpp tagstats_handler.hpp statistics_handler.hpp string_store.hpp geodistribution.hpp
 	$(CXX) $(CXXFLAGS) -o $@ $< $(LDFLAGS) $(LIB_PROTOBUF) $(LIB_GD) $(LIB_SQLITE)
 
 check:
diff --git a/tagstats/statistics_handler.hpp b/tagstats/statistics_handler.hpp
new file mode 100644
index 0000000..7f5587e
--- /dev/null
+++ b/tagstats/statistics_handler.hpp
@@ -0,0 +1,232 @@
+#ifndef TAGSTATS_HANDLER_STATISTICS_HPP
+#define TAGSTATS_HANDLER_STATISTICS_HPP
+
+/*
+
+Copyright 2011 Jochen Topf <jochen@topf.org> and others (see README).
+
+This file is part of Taginfo (https://github.com/joto/taginfo).
+
+Osmium is free software: you can redistribute it and/or modify it under the
+terms of the GNU Lesser General Public License or (at your option) the GNU
+General Public License as published by the Free Software Foundation, either
+version 3 of the Licenses, or (at your option) any later version.
+
+Osmium is distributed in the hope that it will be useful, but WITHOUT ANY
+WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A
+PARTICULAR PURPOSE. See the GNU Lesser General Public License and the GNU
+General Public License for more details.
+
+You should have received a copy of the Licenses along with Osmium. If not, see
+<http://www.gnu.org/licenses/>.
+
+*/
+
+#include <osmium/utils/sqlite.hpp>
+
+/**
+ * Osmium handler that collects basic statistics from OSM data and
+ * writes it to a Sqlite database.
+ */
+class StatisticsHandler : public Osmium::Handler::Base {
+
+public:
+
+    StatisticsHandler() : Base() {
+        // if you change anything in this array, also change the corresponding struct below
+        static const char *sn[] = {
+            "nodes",
+            "nodes_without_tags",
+            "node_tags",
+            "max_node_id",
+            "max_tags_on_node",
+            "ways",
+            "way_tags",
+            "way_nodes",
+            "max_way_id",
+            "max_tags_on_way",
+            "max_nodes_on_way",
+            "closed_ways",
+            "relations",
+            "relation_tags",
+            "relation_members",
+            "max_relation_id",
+            "max_tags_on_relation",
+            "max_members_on_relation",
+            "max_user_id",
+            "anon_user_objects",
+            "max_node_version",
+            "max_way_version",
+            "max_relation_version",
+            "sum_node_version",
+            "sum_way_version",
+            "sum_relation_version",
+            "max_changeset_id",
+            0    // last element (sentinel) must always be 0
+        };
+        m_stat_names = sn;
+
+        // initialize all statistics to zero
+        for (int i=0; m_stat_names[i]; ++i) {
+            ((uint64_t *) &m_stats)[i] = 0;
+        }
+    }
+
+    void node(const Osmium::OSM::Node* node) {
+        update_common_stats(node);
+        m_stats.nodes++;
+        if (m_tag_count == 0) {
+            m_stats.nodes_without_tags++;
+        }
+        if (m_id > (int64_t) m_stats.max_node_id) {
+            m_stats.max_node_id = m_id;
+        }
+        m_stats.node_tags += m_tag_count;
+        if (m_tag_count > (int64_t) m_stats.max_tags_on_node) {
+            m_stats.max_tags_on_node = m_tag_count;
+        }
+        if (m_version > (int64_t) m_stats.max_node_version) {
+            m_stats.max_node_version = m_version;
+        }
+        m_stats.sum_node_version += m_version;
+    }
+
+    void way(const Osmium::OSM::Way* way) {
+        update_common_stats(way);
+        m_stats.ways++;
+        if (way->is_closed()) {
+            m_stats.closed_ways++;
+        }
+        if (m_id > (int64_t) m_stats.max_way_id) {
+            m_stats.max_way_id = m_id;
+        }
+        m_stats.way_tags += m_tag_count;
+        m_stats.way_nodes += way->node_count();
+        if (m_tag_count > (int64_t) m_stats.max_tags_on_way) {
+            m_stats.max_tags_on_way = m_tag_count;
+        }
+        if (way->node_count() > (int64_t) m_stats.max_nodes_on_way) {
+            m_stats.max_nodes_on_way = way->node_count();
+        }
+        if (m_version > (int64_t) m_stats.max_way_version) {
+            m_stats.max_way_version = m_version;
+        }
+        m_stats.sum_way_version += m_version;
+    }
+
+    void relation(const Osmium::OSM::Relation* relation) {
+        update_common_stats(relation);
+        m_stats.relations++;
+        if (m_id > (int64_t) m_stats.max_relation_id) {
+            m_stats.max_relation_id = m_id;
+        }
+        m_stats.relation_tags += m_tag_count;
+        osm_sequence_id_t member_count = relation->members().size();
+        m_stats.relation_members += member_count;
+        if (m_tag_count > (int64_t) m_stats.max_tags_on_relation) {
+            m_stats.max_tags_on_relation = m_tag_count;
+        }
+        if (member_count > (int64_t) m_stats.max_members_on_relation) {
+            m_stats.max_members_on_relation = member_count;
+        }
+        if (m_version > (int64_t) m_stats.max_relation_version) {
+            m_stats.max_relation_version = m_version;
+        }
+        m_stats.sum_relation_version += m_version;
+    }
+
+    void final() {
+        unlink("count.db");
+        Osmium::Sqlite::Database db("count.db");
+
+        sqlite3 *sqlite_db = db.get_sqlite3();
+        if (SQLITE_OK != sqlite3_exec(sqlite_db, \
+                                        "CREATE TABLE stats (" \
+                                        "  key    TEXT, " \
+                                        "  value  INT64 " \
+                                        ");", 0, 0, 0)) {
+            std::cerr << "Database error: " << sqlite3_errmsg(sqlite_db) << "\n";
+            sqlite3_close(sqlite_db);
+            exit(1);
+        }
+
+        Osmium::Sqlite::Statement *statement_insert_into_main_stats = db.prepare("INSERT INTO stats (key, value) VALUES (?, ?);");
+        db.begin_transaction();
+
+        for (int i=0; m_stat_names[i]; ++i) {
+            statement_insert_into_main_stats
+            ->bind_text(m_stat_names[i])
+            ->bind_int64( ((uint64_t *) &m_stats)[i] )
+            ->execute();
+        }
+        statement_insert_into_main_stats
+        ->bind_text("nodes_with_tags")
+        ->bind_int64( ((uint64_t *) &m_stats)[0] - ((uint64_t *) &m_stats)[1] )
+        ->execute();
+
+        db.commit();
+
+        delete statement_insert_into_main_stats;
+    }
+
+private:
+
+    // if you change anything in this struct, also change the corresponding array above
+    struct statistics {
+        uint64_t nodes;
+        uint64_t nodes_without_tags;
+        uint64_t node_tags;
+        uint64_t max_node_id;
+        uint64_t max_tags_on_node;
+        uint64_t ways;
+        uint64_t way_tags;
+        uint64_t way_nodes;
+        uint64_t max_way_id;
+        uint64_t max_tags_on_way;
+        uint64_t max_nodes_on_way;
+        uint64_t closed_ways;
+        uint64_t relations;
+        uint64_t relation_tags;
+        uint64_t relation_members;
+        uint64_t max_relation_id;
+        uint64_t max_tags_on_relation;
+        uint64_t max_members_on_relation;
+        uint64_t max_user_id;
+        uint64_t anon_user_objects;
+        uint64_t max_node_version;
+        uint64_t max_way_version;
+        uint64_t max_relation_version;
+        uint64_t sum_node_version;
+        uint64_t sum_way_version;
+        uint64_t sum_relation_version;
+        uint64_t max_changeset_id;
+    } m_stats;
+
+    const char **m_stat_names;
+
+    osm_object_id_t m_id;
+    osm_version_t   m_version;
+    int             m_tag_count;
+
+    void update_common_stats(const Osmium::OSM::Object* object) {
+        m_id        = object->id();
+        m_version   = object->version();
+        m_tag_count = object->tags().size();
+
+        osm_user_id_t uid = object->uid();
+        if (uid == 0) {
+            m_stats.anon_user_objects++;
+        }
+        if (uid > (int64_t) m_stats.max_user_id) {
+            m_stats.max_user_id = uid;
+        }
+
+        osm_changeset_id_t changeset = object->changeset();
+        if (changeset > (int64_t) m_stats.max_changeset_id) {
+            m_stats.max_changeset_id = changeset;
+        }
+    }
+
+}; // class StatisticsHandler
+
+#endif // TAGSTATS_HANDLER_STATISTICS_HPP
diff --git a/tagstats/tagstats.cpp b/tagstats/tagstats.cpp
index 8a6add7..95394a7 100644
--- a/tagstats/tagstats.cpp
+++ b/tagstats/tagstats.cpp
@@ -1,9 +1,30 @@
 
+/*
+
+Copyright 2011 Jochen Topf <jochen@topf.org> and others (see README).
+
+This file is part of Taginfo (https://github.com/joto/taginfo).
+
+Osmium is free software: you can redistribute it and/or modify it under the
+terms of the GNU Lesser General Public License or (at your option) the GNU
+General Public License as published by the Free Software Foundation, either
+version 3 of the Licenses, or (at your option) any later version.
+
+Osmium is distributed in the hope that it will be useful, but WITHOUT ANY
+WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A
+PARTICULAR PURPOSE. See the GNU Lesser General Public License and the GNU
+General Public License for more details.
+
+You should have received a copy of the Licenses along with Osmium. If not, see
+<http://www.gnu.org/licenses/>.
+
+*/
+
 #include <getopt.h>
 
 #include <osmium.hpp>
 
-#include <osmium/handler/statistics.hpp>
+#include "statistics_handler.hpp"
 
 /**
  * Positions are stored in this type of integer for the distribution images.
diff --git a/tagstats/tagstats_handler.hpp b/tagstats/tagstats_handler.hpp
index 75b4e65..2e66853 100644
--- a/tagstats/tagstats_handler.hpp
+++ b/tagstats/tagstats_handler.hpp
@@ -1,6 +1,28 @@
 #ifndef TAGSTATS_HANDLER_HPP
 #define TAGSTATS_HANDLER_HPP
 
+/*
+
+Copyright 2011 Jochen Topf <jochen@topf.org> and others (see README).
+
+This file is part of Taginfo (https://github.com/joto/taginfo).
+
+Osmium is free software: you can redistribute it and/or modify it under the
+terms of the GNU Lesser General Public License or (at your option) the GNU
+General Public License as published by the Free Software Foundation, either
+version 3 of the Licenses, or (at your option) any later version.
+
+Osmium is distributed in the hope that it will be useful, but WITHOUT ANY
+WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A
+PARTICULAR PURPOSE. See the GNU Lesser General Public License and the GNU
+General Public License for more details.
+
+You should have received a copy of the Licenses along with Osmium. If not, see
+<http://www.gnu.org/licenses/>.
+
+*/
+
+
 #include <google/sparse_hash_map>
 #include <string>
 #include <fstream>
@@ -282,7 +304,7 @@ class TagStatsHandler : public Osmium::Handler::Base {
 #endif // TAGSTATS_COUNT_KEY_COMBINATIONS
     }
 
-    Osmium::Handler::Statistics osmium_handler_stats;
+    StatisticsHandler statistics_handler;
 
     MapToInt<rough_position_t> m_map_to_int;
 
@@ -295,7 +317,7 @@ public:
     TagStatsHandler(MapToInt<rough_position_t>& map_to_int) :
         Base(),
         max_timestamp(0),
-        osmium_handler_stats(),
+        statistics_handler(),
         m_map_to_int(map_to_int)
 #ifdef TAGSTATS_GEODISTRIBUTION_FOR_WAYS
         , m_storage()
@@ -311,7 +333,7 @@ public:
     }
 
     void node(Osmium::OSM::Node* node) {
-        osmium_handler_stats.node(node);
+        statistics_handler.node(node);
         collect_tag_stats(node);
 #ifdef TAGSTATS_GEODISTRIBUTION_FOR_WAYS
         try {
@@ -323,12 +345,12 @@ public:
     }
 
     void way(Osmium::OSM::Way* way) {
-        osmium_handler_stats.way(way);
+        statistics_handler.way(way);
         collect_tag_stats(way);
     }
 
     void relation(Osmium::OSM::Relation* relation) {
-        osmium_handler_stats.relation(relation);
+        statistics_handler.relation(relation);
         collect_tag_stats(relation);
     }
 
@@ -396,7 +418,7 @@ public:
     }
 
     void final() {
-        osmium_handler_stats.final();
+        statistics_handler.final();
         _print_memory_usage();
         timer = time(0);