Add --output option to osrm-extract for custom output base path

Author: afarber

.gitignore

@@ -43,7 +43,7 @@ Thumbs.db
 #######################
 /_build*
 /build/
-/test/data/monaco.osrm*
+/test/data/monaco*.osrm*
 /test/data/ch
 /test/data/mld
 /cmake/postinst

CHANGELOG.md

@@ -17,6 +17,7 @@
       - FIXED: Work around compilation error due to a false-positive of array-bounds check in sol2 [#7317](https://github.com/Project-OSRM/osrm-backend/pull/7317)
       - FIXED: Fix compilation with gcc >14 in release with LTO. [#7268](https://github.com/Project-OSRM/osrm-backend/issues/7268)
     - Misc:
+      - ADDED: `--output` / `-o` option to osrm-extract for specifying custom output base path [#4930](https://github.com/Project-OSRM/osrm-backend/issues/4930)
       - ADDED: `--max-header-size` to override the (automatically) configured maximum header size for osrm-routed [#7336](https://github.com/Project-OSRM/osrm-backend/pull/7336)
       - CHANGED: Use boost::beast instead of own HTTP code for osrm-routed [#7328](https://github.com/Project-OSRM/osrm-backend/pull/7328)
       - ADDED: `SHM_LOCK_DIR` environment variable for shared memory lock file directory [#7312](https://github.com/Project-OSRM/osrm-backend/pull/7312)

docs/profiles.md

@@ -12,6 +12,18 @@ When running OSRM preprocessing commands you specify the profile with the --prof
 
 `osrm-extract --profile ../profiles/car.lua planet-latest.osm.pbf`
 
+### Using Multiple Profiles with the Same Input
+
+You can extract the same OSM file with different profiles by specifying an output path:
+
+```bash
+osrm-extract --profile profiles/car.lua planet.osm.pbf --output /data/car
+osrm-extract --profile profiles/bicycle.lua planet.osm.pbf --output /data/bicycle
+osrm-extract --profile profiles/foot.lua planet.osm.pbf --output /data/foot
+```
+
+This avoids the need to create symbolic links to the input file.
+
 ## Processing flow
 It's important to understand that profiles are used when preprocessing the OSM data, NOT at query time when routes are computed.
 

include/extractor/extractor_config.hpp

@@ -77,6 +77,7 @@ struct ExtractorConfig final : storage::IOConfig
     std::filesystem::path input_path;
     std::filesystem::path profile_path;
     std::vector<std::filesystem::path> location_dependent_data_paths;
+    std::filesystem::path output_path;
     std::string data_version;
 
     unsigned requested_num_threads = 0;

src/tools/extract.cpp

@@ -78,7 +78,10 @@ return_code parseArguments(int argc,
         boost::program_options::bool_switch(&extractor_config.dump_nbg_graph)
             ->implicit_value(true)
             ->default_value(false),
-        "Dump raw node-based graph to *.osrm file for debug purposes.");
+        "Dump raw node-based graph to *.osrm file for debug purposes.")(
+        "output,o",
+        boost::program_options::value<std::filesystem::path>(&extractor_config.output_path),
+        "Output base path for generated files (default: derived from input file name)");
 
     bool dummy;
     // hidden options, will be allowed on command line, but will not be
@@ -166,7 +169,14 @@ try
 
     util::LogPolicy::GetInstance().SetLevel(verbosity);
 
-    extractor_config.UseDefaultOutputNames(extractor_config.input_path);
+    if (!extractor_config.output_path.empty())
+    {
+        extractor_config.UseDefaultOutputNames(extractor_config.output_path);
+    }
+    else
+    {
+        extractor_config.UseDefaultOutputNames(extractor_config.input_path);
+    }
 
     if (1 > extractor_config.requested_num_threads)
     {

unit_tests/library/extract.cpp

@@ -5,6 +5,7 @@
 #include "osrm/extractor_config.hpp"
 
 #include <boost/algorithm/string.hpp>
+#include <filesystem>
 #include <thread>
 
 // utility class to redirect stderr so we can test it
@@ -48,6 +49,21 @@ BOOST_AUTO_TEST_CASE(test_extract_with_valid_config)
     BOOST_CHECK_NO_THROW(osrm::extract(config));
 }
 
+BOOST_AUTO_TEST_CASE(test_extract_with_custom_output_path)
+{
+    osrm::ExtractorConfig config;
+    config.input_path = OSRM_TEST_DATA_DIR "/monaco.osm.pbf";
+    // Use custom output path instead of deriving from input
+    config.UseDefaultOutputNames(OSRM_TEST_DATA_DIR "/monaco-custom-output");
+    config.profile_path = OSRM_TEST_DATA_DIR "/../../profiles/car.lua";
+    config.small_component_size = 1000;
+    config.requested_num_threads = std::thread::hardware_concurrency();
+    BOOST_CHECK_NO_THROW(osrm::extract(config));
+
+    // Verify output files exist at custom path
+    BOOST_CHECK(std::filesystem::exists(OSRM_TEST_DATA_DIR "/monaco-custom-output.osrm.names"));
+}
+
 BOOST_AUTO_TEST_CASE(test_setup_runtime_error)
 {
     osrm::ExtractorConfig config;