osrm-backend/include/osmium/io/detail/pbf_decoder.hpp

905 lines
45 KiB
C++
Raw Normal View History

#ifndef OSMIUM_IO_DETAIL_PBF_DECODER_HPP
#define OSMIUM_IO_DETAIL_PBF_DECODER_HPP
/*
This file is part of Osmium (http://osmcode.org/libosmium).
Copyright 2013-2017 Jochen Topf <jochen@topf.org> and others (see README).
Boost Software License - Version 1.0 - August 17th, 2003
Permission is hereby granted, free of charge, to any person or organization
obtaining a copy of the software and accompanying documentation covered by
this license (the "Software") to use, reproduce, display, distribute,
execute, and transmit the Software, and to prepare derivative works of the
Software, and to permit third-parties to whom the Software is furnished to
do so, all subject to the following:
The copyright notices in the Software and this entire statement, including
the above license grant, this restriction and the following disclaimer,
must be included in all copies of the Software, in whole or in part, and
all derivative works of the Software, unless such copies or derivative
works are solely in the form of machine-executable object code generated by
a source language processor.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE, TITLE AND NON-INFRINGEMENT. IN NO EVENT
SHALL THE COPYRIGHT HOLDERS OR ANYONE DISTRIBUTING THE SOFTWARE BE LIABLE
FOR ANY DAMAGES OR OTHER LIABILITY, WHETHER IN CONTRACT, TORT OR OTHERWISE,
ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
DEALINGS IN THE SOFTWARE.
*/
#include <cstdint>
#include <cstring>
#include <limits>
#include <memory>
#include <stdexcept>
#include <string>
#include <utility>
#include <vector>
#include <protozero/iterators.hpp>
#include <protozero/pbf_message.hpp>
#include <protozero/types.hpp>
#include <osmium/builder/osm_object_builder.hpp>
#include <osmium/io/detail/pbf.hpp> // IWYU pragma: export
#include <osmium/io/detail/protobuf_tags.hpp>
#include <osmium/io/detail/zlib.hpp>
#include <osmium/io/file_format.hpp>
#include <osmium/io/header.hpp>
#include <osmium/memory/buffer.hpp>
#include <osmium/osm/box.hpp>
#include <osmium/osm/entity_bits.hpp>
#include <osmium/osm/item_type.hpp>
#include <osmium/osm/location.hpp>
#include <osmium/osm/node.hpp>
#include <osmium/osm/object.hpp>
#include <osmium/osm/relation.hpp>
#include <osmium/osm/timestamp.hpp>
#include <osmium/osm/types.hpp>
#include <osmium/osm/way.hpp>
#include <osmium/util/cast.hpp>
#include <osmium/util/delta.hpp>
namespace osmium {
namespace builder {
class Builder;
} // namespace builder
namespace io {
namespace detail {
using protozero::data_view;
using osm_string_len_type = std::pair<const char*, osmium::string_size_type>;
class PBFPrimitiveBlockDecoder {
static constexpr const size_t initial_buffer_size = 2 * 1024 * 1024;
data_view m_data;
std::vector<osm_string_len_type> m_stringtable;
int64_t m_lon_offset = 0;
int64_t m_lat_offset = 0;
int64_t m_date_factor = 1000;
int32_t m_granularity = 100;
osmium::osm_entity_bits::type m_read_types;
osmium::memory::Buffer m_buffer { initial_buffer_size };
osmium::io::read_meta m_read_metadata;
void decode_stringtable(const data_view& data) {
if (!m_stringtable.empty()) {
throw osmium::pbf_error{"more than one stringtable in pbf file"};
}
protozero::pbf_message<OSMFormat::StringTable> pbf_string_table{data};
while (pbf_string_table.next(OSMFormat::StringTable::repeated_bytes_s)) {
const auto str_view = pbf_string_table.get_view();
if (str_view.size() > osmium::max_osm_string_length) {
throw osmium::pbf_error{"overlong string in string table"};
}
m_stringtable.emplace_back(str_view.data(), osmium::string_size_type(str_view.size()));
}
}
void decode_primitive_block_metadata() {
protozero::pbf_message<OSMFormat::PrimitiveBlock> pbf_primitive_block{m_data};
while (pbf_primitive_block.next()) {
switch (pbf_primitive_block.tag()) {
case OSMFormat::PrimitiveBlock::required_StringTable_stringtable:
decode_stringtable(pbf_primitive_block.get_view());
break;
case OSMFormat::PrimitiveBlock::optional_int32_granularity:
m_granularity = pbf_primitive_block.get_int32();
break;
case OSMFormat::PrimitiveBlock::optional_int32_date_granularity:
m_date_factor = pbf_primitive_block.get_int32();
break;
case OSMFormat::PrimitiveBlock::optional_int64_lat_offset:
m_lat_offset = pbf_primitive_block.get_int64();
break;
case OSMFormat::PrimitiveBlock::optional_int64_lon_offset:
m_lon_offset = pbf_primitive_block.get_int64();
break;
default:
pbf_primitive_block.skip();
}
}
}
void decode_primitive_block_data() {
protozero::pbf_message<OSMFormat::PrimitiveBlock> pbf_primitive_block{m_data};
while (pbf_primitive_block.next(OSMFormat::PrimitiveBlock::repeated_PrimitiveGroup_primitivegroup)) {
protozero::pbf_message<OSMFormat::PrimitiveGroup> pbf_primitive_group = pbf_primitive_block.get_message();
while (pbf_primitive_group.next()) {
switch (pbf_primitive_group.tag()) {
case OSMFormat::PrimitiveGroup::repeated_Node_nodes:
if (m_read_types & osmium::osm_entity_bits::node) {
decode_node(pbf_primitive_group.get_view());
m_buffer.commit();
} else {
pbf_primitive_group.skip();
}
break;
case OSMFormat::PrimitiveGroup::optional_DenseNodes_dense:
if (m_read_types & osmium::osm_entity_bits::node) {
if (m_read_metadata == osmium::io::read_meta::yes) {
decode_dense_nodes(pbf_primitive_group.get_view());
} else {
decode_dense_nodes_without_metadata(pbf_primitive_group.get_view());
}
m_buffer.commit();
} else {
pbf_primitive_group.skip();
}
break;
case OSMFormat::PrimitiveGroup::repeated_Way_ways:
if (m_read_types & osmium::osm_entity_bits::way) {
decode_way(pbf_primitive_group.get_view());
m_buffer.commit();
} else {
pbf_primitive_group.skip();
}
break;
case OSMFormat::PrimitiveGroup::repeated_Relation_relations:
if (m_read_types & osmium::osm_entity_bits::relation) {
decode_relation(pbf_primitive_group.get_view());
m_buffer.commit();
} else {
pbf_primitive_group.skip();
}
break;
default:
pbf_primitive_group.skip();
}
}
}
}
osm_string_len_type decode_info(const data_view& data, osmium::OSMObject& object) {
osm_string_len_type user{"", 0};
protozero::pbf_message<OSMFormat::Info> pbf_info{data};
while (pbf_info.next()) {
switch (pbf_info.tag()) {
case OSMFormat::Info::optional_int32_version:
{
const auto version = pbf_info.get_int32();
if (version < 0) {
throw osmium::pbf_error{"object version must not be negative"};
}
object.set_version(static_cast_with_assert<object_version_type>(version));
}
break;
case OSMFormat::Info::optional_int64_timestamp:
object.set_timestamp(pbf_info.get_int64() * m_date_factor / 1000);
break;
case OSMFormat::Info::optional_int64_changeset:
{
const auto changeset_id = pbf_info.get_int64();
if (changeset_id < 0) {
throw osmium::pbf_error{"object changeset_id must not be negative"};
}
object.set_changeset(static_cast_with_assert<changeset_id_type>(changeset_id));
}
break;
case OSMFormat::Info::optional_int32_uid:
object.set_uid_from_signed(pbf_info.get_int32());
break;
case OSMFormat::Info::optional_uint32_user_sid:
user = m_stringtable.at(pbf_info.get_uint32());
break;
case OSMFormat::Info::optional_bool_visible:
object.set_visible(pbf_info.get_bool());
break;
default:
pbf_info.skip();
}
}
return user;
}
using kv_type = protozero::iterator_range<protozero::pbf_reader::const_uint32_iterator>;
void build_tag_list(osmium::builder::Builder& parent, const kv_type& keys, const kv_type& vals) {
if (!keys.empty()) {
osmium::builder::TagListBuilder builder{parent};
auto kit = keys.begin();
auto vit = vals.begin();
while (kit != keys.end()) {
if (vit == vals.end()) {
// this is against the spec, must have same number of elements
throw osmium::pbf_error{"PBF format error"};
}
const auto& k = m_stringtable.at(*kit++);
const auto& v = m_stringtable.at(*vit++);
builder.add_tag(k.first, k.second, v.first, v.second);
}
}
}
int32_t convert_pbf_coordinate(int64_t c) const {
return int32_t((c * m_granularity + m_lon_offset) / resolution_convert);
}
void decode_node(const data_view& data) {
osmium::builder::NodeBuilder builder{m_buffer};
osmium::Node& node = builder.object();
kv_type keys;
kv_type vals;
int64_t lon = std::numeric_limits<int64_t>::max();
int64_t lat = std::numeric_limits<int64_t>::max();
osm_string_len_type user{"", 0};
protozero::pbf_message<OSMFormat::Node> pbf_node{data};
while (pbf_node.next()) {
switch (pbf_node.tag()) {
case OSMFormat::Node::required_sint64_id:
node.set_id(pbf_node.get_sint64());
break;
case OSMFormat::Node::packed_uint32_keys:
keys = pbf_node.get_packed_uint32();
break;
case OSMFormat::Node::packed_uint32_vals:
vals = pbf_node.get_packed_uint32();
break;
case OSMFormat::Node::optional_Info_info:
if (m_read_metadata == osmium::io::read_meta::yes) {
user = decode_info(pbf_node.get_view(), builder.object());
} else {
pbf_node.skip();
}
break;
case OSMFormat::Node::required_sint64_lat:
lat = pbf_node.get_sint64();
break;
case OSMFormat::Node::required_sint64_lon:
lon = pbf_node.get_sint64();
break;
default:
pbf_node.skip();
}
}
if (node.visible()) {
if (lon == std::numeric_limits<int64_t>::max() ||
lat == std::numeric_limits<int64_t>::max()) {
throw osmium::pbf_error{"illegal coordinate format"};
}
node.set_location(osmium::Location{
convert_pbf_coordinate(lon),
convert_pbf_coordinate(lat)
});
}
builder.set_user(user.first, user.second);
build_tag_list(builder, keys, vals);
}
void decode_way(const data_view& data) {
osmium::builder::WayBuilder builder{m_buffer};
kv_type keys;
kv_type vals;
protozero::iterator_range<protozero::pbf_reader::const_sint64_iterator> refs;
protozero::iterator_range<protozero::pbf_reader::const_sint64_iterator> lats;
protozero::iterator_range<protozero::pbf_reader::const_sint64_iterator> lons;
osm_string_len_type user{"", 0};
protozero::pbf_message<OSMFormat::Way> pbf_way{data};
while (pbf_way.next()) {
switch (pbf_way.tag()) {
case OSMFormat::Way::required_int64_id:
builder.object().set_id(pbf_way.get_int64());
break;
case OSMFormat::Way::packed_uint32_keys:
keys = pbf_way.get_packed_uint32();
break;
case OSMFormat::Way::packed_uint32_vals:
vals = pbf_way.get_packed_uint32();
break;
case OSMFormat::Way::optional_Info_info:
if (m_read_metadata == osmium::io::read_meta::yes) {
user = decode_info(pbf_way.get_view(), builder.object());
} else {
pbf_way.skip();
}
break;
case OSMFormat::Way::packed_sint64_refs:
refs = pbf_way.get_packed_sint64();
break;
case OSMFormat::Way::packed_sint64_lat:
lats = pbf_way.get_packed_sint64();
break;
case OSMFormat::Way::packed_sint64_lon:
lons = pbf_way.get_packed_sint64();
break;
default:
pbf_way.skip();
}
}
builder.set_user(user.first, user.second);
if (!refs.empty()) {
osmium::builder::WayNodeListBuilder wnl_builder{builder};
osmium::util::DeltaDecode<int64_t> ref;
if (lats.empty()) {
for (const auto& ref_value : refs) {
wnl_builder.add_node_ref(ref.update(ref_value));
}
} else {
osmium::util::DeltaDecode<int64_t> lon;
osmium::util::DeltaDecode<int64_t> lat;
while (!refs.empty() && !lons.empty() && !lats.empty()) {
wnl_builder.add_node_ref(
ref.update(refs.front()),
osmium::Location{convert_pbf_coordinate(lon.update(lons.front())),
convert_pbf_coordinate(lat.update(lats.front()))}
);
refs.drop_front();
lons.drop_front();
lats.drop_front();
}
}
}
build_tag_list(builder, keys, vals);
}
void decode_relation(const data_view& data) {
osmium::builder::RelationBuilder builder{m_buffer};
kv_type keys;
kv_type vals;
protozero::iterator_range<protozero::pbf_reader::const_int32_iterator> roles;
protozero::iterator_range<protozero::pbf_reader::const_sint64_iterator> refs;
protozero::iterator_range<protozero::pbf_reader::const_int32_iterator> types;
osm_string_len_type user{"", 0};
protozero::pbf_message<OSMFormat::Relation> pbf_relation{data};
while (pbf_relation.next()) {
switch (pbf_relation.tag()) {
case OSMFormat::Relation::required_int64_id:
builder.object().set_id(pbf_relation.get_int64());
break;
case OSMFormat::Relation::packed_uint32_keys:
keys = pbf_relation.get_packed_uint32();
break;
case OSMFormat::Relation::packed_uint32_vals:
vals = pbf_relation.get_packed_uint32();
break;
case OSMFormat::Relation::optional_Info_info:
if (m_read_metadata == osmium::io::read_meta::yes) {
user = decode_info(pbf_relation.get_view(), builder.object());
} else {
pbf_relation.skip();
}
break;
case OSMFormat::Relation::packed_int32_roles_sid:
roles = pbf_relation.get_packed_int32();
break;
case OSMFormat::Relation::packed_sint64_memids:
refs = pbf_relation.get_packed_sint64();
break;
case OSMFormat::Relation::packed_MemberType_types:
types = pbf_relation.get_packed_enum();
break;
default:
pbf_relation.skip();
}
}
builder.set_user(user.first, user.second);
if (!refs.empty()) {
osmium::builder::RelationMemberListBuilder rml_builder{builder};
osmium::util::DeltaDecode<int64_t> ref;
while (!roles.empty() && !refs.empty() && !types.empty()) {
const auto& r = m_stringtable.at(roles.front());
const int type = types.front();
if (type < 0 || type > 2) {
throw osmium::pbf_error{"unknown relation member type"};
}
rml_builder.add_member(
osmium::item_type(type + 1),
ref.update(refs.front()),
r.first,
r.second
);
roles.drop_front();
refs.drop_front();
types.drop_front();
}
}
build_tag_list(builder, keys, vals);
}
void build_tag_list_from_dense_nodes(osmium::builder::NodeBuilder& builder, protozero::pbf_reader::const_int32_iterator& it, protozero::pbf_reader::const_int32_iterator last) {
osmium::builder::TagListBuilder tl_builder{builder};
while (it != last && *it != 0) {
const auto& k = m_stringtable.at(*it++);
if (it == last) {
throw osmium::pbf_error{"PBF format error"}; // this is against the spec, keys/vals must come in pairs
}
const auto& v = m_stringtable.at(*it++);
tl_builder.add_tag(k.first, k.second, v.first, v.second);
}
if (it != last) {
++it;
}
}
void decode_dense_nodes_without_metadata(const data_view& data) {
protozero::iterator_range<protozero::pbf_reader::const_sint64_iterator> ids;
protozero::iterator_range<protozero::pbf_reader::const_sint64_iterator> lats;
protozero::iterator_range<protozero::pbf_reader::const_sint64_iterator> lons;
protozero::iterator_range<protozero::pbf_reader::const_int32_iterator> tags;
protozero::pbf_message<OSMFormat::DenseNodes> pbf_dense_nodes{data};
while (pbf_dense_nodes.next()) {
switch (pbf_dense_nodes.tag()) {
case OSMFormat::DenseNodes::packed_sint64_id:
ids = pbf_dense_nodes.get_packed_sint64();
break;
case OSMFormat::DenseNodes::packed_sint64_lat:
lats = pbf_dense_nodes.get_packed_sint64();
break;
case OSMFormat::DenseNodes::packed_sint64_lon:
lons = pbf_dense_nodes.get_packed_sint64();
break;
case OSMFormat::DenseNodes::packed_int32_keys_vals:
tags = pbf_dense_nodes.get_packed_int32();
break;
default:
pbf_dense_nodes.skip();
}
}
osmium::util::DeltaDecode<int64_t> dense_id;
osmium::util::DeltaDecode<int64_t> dense_latitude;
osmium::util::DeltaDecode<int64_t> dense_longitude;
auto tag_it = tags.begin();
while (!ids.empty()) {
if (lons.empty() ||
lats.empty()) {
// this is against the spec, must have same number of elements
throw osmium::pbf_error{"PBF format error"};
}
osmium::builder::NodeBuilder builder{m_buffer};
osmium::Node& node = builder.object();
node.set_id(dense_id.update(ids.front()));
ids.drop_front();
const auto lon = dense_longitude.update(lons.front());
lons.drop_front();
const auto lat = dense_latitude.update(lats.front());
lats.drop_front();
builder.object().set_location(osmium::Location(
convert_pbf_coordinate(lon),
convert_pbf_coordinate(lat)
));
if (tag_it != tags.end()) {
build_tag_list_from_dense_nodes(builder, tag_it, tags.end());
}
}
}
void decode_dense_nodes(const data_view& data) {
bool has_info = false;
bool has_visibles = false;
protozero::iterator_range<protozero::pbf_reader::const_sint64_iterator> ids;
protozero::iterator_range<protozero::pbf_reader::const_sint64_iterator> lats;
protozero::iterator_range<protozero::pbf_reader::const_sint64_iterator> lons;
protozero::iterator_range<protozero::pbf_reader::const_int32_iterator> tags;
protozero::iterator_range<protozero::pbf_reader::const_int32_iterator> versions;
protozero::iterator_range<protozero::pbf_reader::const_sint64_iterator> timestamps;
protozero::iterator_range<protozero::pbf_reader::const_sint64_iterator> changesets;
protozero::iterator_range<protozero::pbf_reader::const_sint32_iterator> uids;
protozero::iterator_range<protozero::pbf_reader::const_sint32_iterator> user_sids;
protozero::iterator_range<protozero::pbf_reader::const_int32_iterator> visibles;
protozero::pbf_message<OSMFormat::DenseNodes> pbf_dense_nodes{data};
while (pbf_dense_nodes.next()) {
switch (pbf_dense_nodes.tag()) {
case OSMFormat::DenseNodes::packed_sint64_id:
ids = pbf_dense_nodes.get_packed_sint64();
break;
case OSMFormat::DenseNodes::optional_DenseInfo_denseinfo:
{
has_info = true;
protozero::pbf_message<OSMFormat::DenseInfo> pbf_dense_info{pbf_dense_nodes.get_message()};
while (pbf_dense_info.next()) {
switch (pbf_dense_info.tag()) {
case OSMFormat::DenseInfo::packed_int32_version:
versions = pbf_dense_info.get_packed_int32();
break;
case OSMFormat::DenseInfo::packed_sint64_timestamp:
timestamps = pbf_dense_info.get_packed_sint64();
break;
case OSMFormat::DenseInfo::packed_sint64_changeset:
changesets = pbf_dense_info.get_packed_sint64();
break;
case OSMFormat::DenseInfo::packed_sint32_uid:
uids = pbf_dense_info.get_packed_sint32();
break;
case OSMFormat::DenseInfo::packed_sint32_user_sid:
user_sids = pbf_dense_info.get_packed_sint32();
break;
case OSMFormat::DenseInfo::packed_bool_visible:
has_visibles = true;
visibles = pbf_dense_info.get_packed_bool();
break;
default:
pbf_dense_info.skip();
}
}
}
break;
case OSMFormat::DenseNodes::packed_sint64_lat:
lats = pbf_dense_nodes.get_packed_sint64();
break;
case OSMFormat::DenseNodes::packed_sint64_lon:
lons = pbf_dense_nodes.get_packed_sint64();
break;
case OSMFormat::DenseNodes::packed_int32_keys_vals:
tags = pbf_dense_nodes.get_packed_int32();
break;
default:
pbf_dense_nodes.skip();
}
}
osmium::util::DeltaDecode<int64_t> dense_id;
osmium::util::DeltaDecode<int64_t> dense_latitude;
osmium::util::DeltaDecode<int64_t> dense_longitude;
osmium::util::DeltaDecode<int64_t> dense_uid;
osmium::util::DeltaDecode<int64_t> dense_user_sid;
osmium::util::DeltaDecode<int64_t> dense_changeset;
osmium::util::DeltaDecode<int64_t> dense_timestamp;
auto tag_it = tags.begin();
while (!ids.empty()) {
if (lons.empty() ||
lats.empty()) {
// this is against the spec, must have same number of elements
throw osmium::pbf_error{"PBF format error"};
}
bool visible = true;
osmium::builder::NodeBuilder builder{m_buffer};
osmium::Node& node = builder.object();
node.set_id(dense_id.update(ids.front()));
ids.drop_front();
if (has_info) {
if (versions.empty() ||
changesets.empty() ||
timestamps.empty() ||
uids.empty() ||
user_sids.empty()) {
// this is against the spec, must have same number of elements
throw osmium::pbf_error{"PBF format error"};
}
const auto version = versions.front();
versions.drop_front();
if (version < 0) {
throw osmium::pbf_error{"object version must not be negative"};
}
node.set_version(static_cast<osmium::object_version_type>(version));
const auto changeset_id = dense_changeset.update(changesets.front());
changesets.drop_front();
if (changeset_id < 0) {
throw osmium::pbf_error{"object changeset_id must not be negative"};
}
node.set_changeset(static_cast<osmium::changeset_id_type>(changeset_id));
node.set_timestamp(dense_timestamp.update(timestamps.front()) * m_date_factor / 1000);
timestamps.drop_front();
node.set_uid_from_signed(static_cast<osmium::signed_user_id_type>(dense_uid.update(uids.front())));
uids.drop_front();
if (has_visibles) {
if (visibles.empty()) {
// this is against the spec, must have same number of elements
throw osmium::pbf_error{"PBF format error"};
}
visible = (visibles.front() != 0);
visibles.drop_front();
}
node.set_visible(visible);
const auto& u = m_stringtable.at(dense_user_sid.update(user_sids.front()));
user_sids.drop_front();
builder.set_user(u.first, u.second);
}
// even if the node isn't visible, there's still a record
// of its lat/lon in the dense arrays.
const auto lon = dense_longitude.update(lons.front());
lons.drop_front();
const auto lat = dense_latitude.update(lats.front());
lats.drop_front();
if (visible) {
builder.object().set_location(osmium::Location{
convert_pbf_coordinate(lon),
convert_pbf_coordinate(lat)
});
}
if (tag_it != tags.end()) {
build_tag_list_from_dense_nodes(builder, tag_it, tags.end());
}
}
}
public:
PBFPrimitiveBlockDecoder(const data_view& data, osmium::osm_entity_bits::type read_types, osmium::io::read_meta read_metadata) :
m_data(data),
m_read_types(read_types),
m_read_metadata(read_metadata) {
}
PBFPrimitiveBlockDecoder(const PBFPrimitiveBlockDecoder&) = delete;
PBFPrimitiveBlockDecoder& operator=(const PBFPrimitiveBlockDecoder&) = delete;
PBFPrimitiveBlockDecoder(PBFPrimitiveBlockDecoder&&) = delete;
PBFPrimitiveBlockDecoder& operator=(PBFPrimitiveBlockDecoder&&) = delete;
~PBFPrimitiveBlockDecoder() noexcept = default;
osmium::memory::Buffer operator()() {
try {
decode_primitive_block_metadata();
decode_primitive_block_data();
} catch (const std::out_of_range&) {
throw osmium::pbf_error{"string id out of range"};
}
return std::move(m_buffer);
}
}; // class PBFPrimitiveBlockDecoder
inline data_view decode_blob(const std::string& blob_data, std::string& output) {
int32_t raw_size = 0;
protozero::data_view zlib_data;
protozero::pbf_message<FileFormat::Blob> pbf_blob{blob_data};
while (pbf_blob.next()) {
switch (pbf_blob.tag()) {
case FileFormat::Blob::optional_bytes_raw:
{
const auto data_len = pbf_blob.get_view();
if (data_len.size() > max_uncompressed_blob_size) {
throw osmium::pbf_error{"illegal blob size"};
}
return data_len;
}
case FileFormat::Blob::optional_int32_raw_size:
raw_size = pbf_blob.get_int32();
if (raw_size <= 0 || uint32_t(raw_size) > max_uncompressed_blob_size) {
throw osmium::pbf_error{"illegal blob size"};
}
break;
case FileFormat::Blob::optional_bytes_zlib_data:
zlib_data = pbf_blob.get_view();
break;
case FileFormat::Blob::optional_bytes_lzma_data:
throw osmium::pbf_error{"lzma blobs not implemented"};
default:
throw osmium::pbf_error{"unknown compression"};
}
}
if (zlib_data.size() != 0 && raw_size != 0) {
return osmium::io::detail::zlib_uncompress_string(
zlib_data.data(),
static_cast<unsigned long>(zlib_data.size()),
static_cast<unsigned long>(raw_size),
output
);
}
throw osmium::pbf_error{"blob contains no data"};
}
inline osmium::Box decode_header_bbox(const data_view& data) {
int64_t left = std::numeric_limits<int64_t>::max();
int64_t right = std::numeric_limits<int64_t>::max();
int64_t top = std::numeric_limits<int64_t>::max();
int64_t bottom = std::numeric_limits<int64_t>::max();
protozero::pbf_message<OSMFormat::HeaderBBox> pbf_header_bbox{data};
while (pbf_header_bbox.next()) {
switch (pbf_header_bbox.tag()) {
case OSMFormat::HeaderBBox::required_sint64_left:
left = pbf_header_bbox.get_sint64();
break;
case OSMFormat::HeaderBBox::required_sint64_right:
right = pbf_header_bbox.get_sint64();
break;
case OSMFormat::HeaderBBox::required_sint64_top:
top = pbf_header_bbox.get_sint64();
break;
case OSMFormat::HeaderBBox::required_sint64_bottom:
bottom = pbf_header_bbox.get_sint64();
break;
default:
pbf_header_bbox.skip();
}
}
if (left == std::numeric_limits<int64_t>::max() ||
right == std::numeric_limits<int64_t>::max() ||
top == std::numeric_limits<int64_t>::max() ||
bottom == std::numeric_limits<int64_t>::max()) {
throw osmium::pbf_error{"invalid bbox"};
}
osmium::Box box;
box.extend(osmium::Location(left / resolution_convert, bottom / resolution_convert));
box.extend(osmium::Location(right / resolution_convert, top / resolution_convert));
return box;
}
inline osmium::io::Header decode_header_block(const data_view& data) {
osmium::io::Header header;
int i = 0;
protozero::pbf_message<OSMFormat::HeaderBlock> pbf_header_block{data};
while (pbf_header_block.next()) {
switch (pbf_header_block.tag()) {
case OSMFormat::HeaderBlock::optional_HeaderBBox_bbox:
header.add_box(decode_header_bbox(pbf_header_block.get_view()));
break;
case OSMFormat::HeaderBlock::repeated_string_required_features:
{
auto feature = pbf_header_block.get_view();
if (!std::strncmp("OsmSchema-V0.6", feature.data(), feature.size())) {
// intentionally left blank
} else if (!std::strncmp("DenseNodes", feature.data(), feature.size())) {
header.set("pbf_dense_nodes", true);
} else if (!std::strncmp("HistoricalInformation", feature.data(), feature.size())) {
header.set_has_multiple_object_versions(true);
} else {
std::string msg{"required feature not supported: "};
msg.append(feature.data(), feature.size());
throw osmium::pbf_error{msg};
}
}
break;
case OSMFormat::HeaderBlock::repeated_string_optional_features:
header.set("pbf_optional_feature_" + std::to_string(i++), pbf_header_block.get_string());
break;
case OSMFormat::HeaderBlock::optional_string_writingprogram:
header.set("generator", pbf_header_block.get_string());
break;
case OSMFormat::HeaderBlock::optional_int64_osmosis_replication_timestamp:
{
const auto timestamp = osmium::Timestamp{pbf_header_block.get_int64()}.to_iso();
header.set("osmosis_replication_timestamp", timestamp);
header.set("timestamp", timestamp);
}
break;
case OSMFormat::HeaderBlock::optional_int64_osmosis_replication_sequence_number:
header.set("osmosis_replication_sequence_number", std::to_string(pbf_header_block.get_int64()));
break;
case OSMFormat::HeaderBlock::optional_string_osmosis_replication_base_url:
header.set("osmosis_replication_base_url", pbf_header_block.get_string());
break;
default:
pbf_header_block.skip();
}
}
return header;
}
/**
* Decode HeaderBlock.
*
* @param header_block_data Input data
* @returns Header object
* @throws osmium::pbf_error If there was a parsing error
*/
inline osmium::io::Header decode_header(const std::string& header_block_data) {
std::string output;
return decode_header_block(decode_blob(header_block_data, output));
}
class PBFDataBlobDecoder {
std::shared_ptr<std::string> m_input_buffer;
osmium::osm_entity_bits::type m_read_types;
osmium::io::read_meta m_read_metadata;
public:
PBFDataBlobDecoder(std::string&& input_buffer, osmium::osm_entity_bits::type read_types, osmium::io::read_meta read_metadata) :
m_input_buffer(std::make_shared<std::string>(std::move(input_buffer))),
m_read_types(read_types),
m_read_metadata(read_metadata) {
}
PBFDataBlobDecoder(const PBFDataBlobDecoder&) = default;
PBFDataBlobDecoder& operator=(const PBFDataBlobDecoder&) = default;
PBFDataBlobDecoder(PBFDataBlobDecoder&&) = default;
PBFDataBlobDecoder& operator=(PBFDataBlobDecoder&&) = default;
~PBFDataBlobDecoder() noexcept = default;
osmium::memory::Buffer operator()() {
std::string output;
PBFPrimitiveBlockDecoder decoder{decode_blob(*m_input_buffer, output), m_read_types, m_read_metadata};
return decoder();
}
}; // class PBFDataBlobDecoder
} // namespace detail
} // namespace io
} // namespace osmium
#endif // OSMIUM_IO_DETAIL_PBF_DECODER_HPP