Merge pull request #201 from p-ranav/feature/parse_known_args

parse_known_args
p-ranav · Sep 21, 2022 · 4dbc910 · 4dbc910
2 parents 4f10f37 + 6a3c6e0
commit 4dbc910
Show file tree

Hide file tree

Showing 4 changed files with 207 additions and 0 deletions.
diff --git a/README.md b/README.md
@@ -37,6 +37,7 @@
      *    [Gathering Remaining Arguments](#gathering-remaining-arguments)
      *    [Parent Parsers](#parent-parsers)
      *    [Subcommands](#subcommands)
+     *    [Parse Known Args](#parse-known-args)
 *    [Further Examples](#further-examples)
      *    [Construct a JSON object from a filename argument](#construct-a-json-object-from-a-filename-argument)
      *    [Positional Arguments with Compound Toggle Arguments](#positional-arguments-with-compound-toggle-arguments)
@@ -799,6 +800,28 @@ When a help message is requested from a subparser, only the help for that partic
 
 Additionally, every parser has a `.is_subcommand_used("<command_name>")` member function to check if a subcommand was used. 
 
+### Parse Known Args
+
+Sometimes a program may only parse a few of the command-line arguments, passing the remaining arguments on to another script or program. In these cases, the `parse_known_args()` function can be useful. It works much like `parse_args()` except that it does not produce an error when extra arguments are present. Instead, it returns a list of remaining argument strings.
+
+```cpp
+#include <argparse/argparse.hpp>
+#include <cassert>
+
+int main(int argc, char *argv[]) {
+  argparse::ArgumentParser program("test");
+  program.add_argument("--foo").implicit_value(true).default_value(false);
+  program.add_argument("bar");
+
+  auto unknown_args =
+    program.parse_known_args({"test", "--foo", "--badger", "BAR", "spam"});
+
+  assert(program.get<bool>("--foo") == true);
+  assert(program.get<std::string>("bar") == std::string{"BAR"});
+  assert((unknown_args == std::vector<std::string>{"--badger", "spam"}));
+}
+```
+
 ## Further Examples
 
 ### Construct a JSON object from a filename argument

diff --git a/include/argparse/argparse.hpp b/include/argparse/argparse.hpp
@@ -1041,6 +1041,21 @@ class ArgumentParser {
     }
   }
 
+  /* Call parse_known_args_internal - which does all the work
+   * Then, validate the parsed arguments
+   * This variant is used mainly for testing
+   * @throws std::runtime_error in case of any invalid argument
+   */
+  std::vector<std::string>
+  parse_known_args(const std::vector<std::string> &arguments) {
+    auto unknown_arguments = parse_known_args_internal(arguments);
+    // Check if all arguments are parsed
+    for ([[maybe_unused]] const auto &[unused, argument] : m_argument_map) {
+      argument->validate();
+    }
+    return unknown_arguments;
+  }
+
   /* Main entry point for parsing command-line arguments using this
    * ArgumentParser
    * @throws std::runtime_error in case of any invalid argument
@@ -1050,6 +1065,15 @@ class ArgumentParser {
     parse_args({argv, argv + argc});
   }
 
+  /* Main entry point for parsing command-line arguments using this
+   * ArgumentParser
+   * @throws std::runtime_error in case of any invalid argument
+   */
+  // NOLINTNEXTLINE(cppcoreguidelines-avoid-c-arrays)
+  void parse_known_args(int argc, const char *const argv[]) {
+    parse_known_args({argv, argv + argc});
+  }
+
   /* Getter for options with default values.
    * @throws std::logic_error if parse_args() has not been previously called
    * @throws std::logic_error if there is no such option
@@ -1262,6 +1286,83 @@ class ArgumentParser {
     m_is_parsed = true;
   }
 
+  /*
+   * Like parse_args_internal but collects unused args into a vector<string>
+   */
+  std::vector<std::string>
+  parse_known_args_internal(const std::vector<std::string> &arguments) {
+
+    std::vector<std::string> unknown_arguments{};
+
+    if (m_program_name.empty() && !arguments.empty()) {
+      m_program_name = arguments.front();
+    }
+    auto end = std::end(arguments);
+    auto positional_argument_it = std::begin(m_positional_arguments);
+    for (auto it = std::next(std::begin(arguments)); it != end;) {
+      const auto &current_argument = *it;
+      if (Argument::is_positional(current_argument)) {
+        if (positional_argument_it == std::end(m_positional_arguments)) {
+
+          std::string_view maybe_command = current_argument;
+
+          // Check sub-parsers
+          auto subparser_it = m_subparser_map.find(maybe_command);
+          if (subparser_it != m_subparser_map.end()) {
+
+            // build list of remaining args
+            const auto unprocessed_arguments =
+                std::vector<std::string>(it, end);
+
+            // invoke subparser
+            m_is_parsed = true;
+            m_subparser_used[maybe_command] = true;
+            return subparser_it->second->get().parse_known_args_internal(
+                unprocessed_arguments);
+          }
+
+          // save current argument as unknown and go to next argument
+          unknown_arguments.push_back(current_argument);
+          ++it;
+        } else {
+          // current argument is the value of a positional argument
+          // consume it
+          auto argument = positional_argument_it++;
+          it = argument->consume(it, end);
+        }
+        continue;
+      }
+
+      auto arg_map_it = m_argument_map.find(current_argument);
+      if (arg_map_it != m_argument_map.end()) {
+        auto argument = arg_map_it->second;
+        it = argument->consume(std::next(it), end, arg_map_it->first);
+      } else if (const auto &compound_arg = current_argument;
+                 compound_arg.size() > 1 && compound_arg[0] == '-' &&
+                 compound_arg[1] != '-') {
+        ++it;
+        for (std::size_t j = 1; j < compound_arg.size(); j++) {
+          auto hypothetical_arg = std::string{'-', compound_arg[j]};
+          auto arg_map_it2 = m_argument_map.find(hypothetical_arg);
+          if (arg_map_it2 != m_argument_map.end()) {
+            auto argument = arg_map_it2->second;
+            it = argument->consume(it, end, arg_map_it2->first);
+          } else {
+            unknown_arguments.push_back(current_argument);
+            break;
+          }
+        }
+      } else {
+        // current argument is an optional-like argument that is unknown
+        // save it and move to next argument
+        unknown_arguments.push_back(current_argument);
+        ++it;
+      }
+    }
+    m_is_parsed = true;
+    return unknown_arguments;
+  }
+
   // Used by print_help.
   std::size_t get_length_of_longest_argument() const {
     if (m_argument_map.empty()) {

diff --git a/test/CMakeLists.txt b/test/CMakeLists.txt
@@ -47,6 +47,7 @@ file(GLOB ARGPARSE_TEST_SOURCES
     test_value_semantics.cpp
     test_version.cpp
     test_subparsers.cpp
+    test_parse_known_args.cpp
 )
 set_source_files_properties(main.cpp
     PROPERTIES

diff --git a/test/test_parse_known_args.cpp b/test/test_parse_known_args.cpp
@@ -0,0 +1,82 @@
+#include <argparse/argparse.hpp>
+#include <doctest.hpp>
+
+using doctest::test_suite;
+
+TEST_CASE("Parse unknown optional and positional arguments without exceptions" *
+          test_suite("parse_known_args")) {
+  argparse::ArgumentParser program("test");
+  program.add_argument("--foo").implicit_value(true).default_value(false);
+  program.add_argument("bar");
+
+  SUBCASE("Parse unknown optional and positional arguments") {
+    auto unknown_args =
+        program.parse_known_args({"test", "--foo", "--badger", "BAR", "spam"});
+    REQUIRE((unknown_args == std::vector<std::string>{"--badger", "spam"}));
+    REQUIRE(program.get<bool>("--foo") == true);
+    REQUIRE(program.get<std::string>("bar") == std::string{"BAR"});
+  }
+
+  SUBCASE("Parse unknown compound arguments") {
+    auto unknown_args = program.parse_known_args({"test", "-jc", "BAR"});
+    REQUIRE((unknown_args == std::vector<std::string>{"-jc"}));
+    REQUIRE(program.get<bool>("--foo") == false);
+    REQUIRE(program.get<std::string>("bar") == std::string{"BAR"});
+  }
+}
+
+TEST_CASE("Parse unknown optional and positional arguments in subparsers "
+          "without exceptions" *
+          test_suite("parse_known_args")) {
+  argparse::ArgumentParser program("test");
+  program.add_argument("--output");
+
+  argparse::ArgumentParser command_1("add");
+  command_1.add_argument("file").nargs(2);
+
+  argparse::ArgumentParser command_2("clean");
+  command_2.add_argument("--fullclean")
+      .default_value(false)
+      .implicit_value(true);
+
+  program.add_subparser(command_1);
+  program.add_subparser(command_2);
+
+  SUBCASE("Parse unknown optional argument") {
+    auto unknown_args =
+        program.parse_known_args({"test", "add", "--badger", "BAR", "spam"});
+    REQUIRE(program.is_subcommand_used("add") == true);
+    REQUIRE((command_1.get<std::vector<std::string>>("file") ==
+             std::vector<std::string>{"BAR", "spam"}));
+    REQUIRE((unknown_args == std::vector<std::string>{"--badger"}));
+  }
+
+  SUBCASE("Parse unknown positional argument") {
+    auto unknown_args =
+        program.parse_known_args({"test", "add", "FOO", "BAR", "spam"});
+    REQUIRE(program.is_subcommand_used("add") == true);
+    REQUIRE((command_1.get<std::vector<std::string>>("file") ==
+             std::vector<std::string>{"FOO", "BAR"}));
+    REQUIRE((unknown_args == std::vector<std::string>{"spam"}));
+  }
+
+  SUBCASE("Parse unknown positional and optional arguments") {
+    auto unknown_args = program.parse_known_args(
+        {"test", "add", "--verbose", "FOO", "5", "BAR", "-jn", "spam"});
+    REQUIRE(program.is_subcommand_used("add") == true);
+    REQUIRE((command_1.get<std::vector<std::string>>("file") ==
+             std::vector<std::string>{"FOO", "5"}));
+    REQUIRE((unknown_args ==
+             std::vector<std::string>{"--verbose", "BAR", "-jn", "spam"}));
+  }
+
+  SUBCASE("Parse unknown positional and optional arguments 2") {
+    auto unknown_args =
+        program.parse_known_args({"test", "clean", "--verbose", "FOO", "5",
+                                  "BAR", "--fullclean", "-jn", "spam"});
+    REQUIRE(program.is_subcommand_used("clean") == true);
+    REQUIRE(command_2.get<bool>("--fullclean") == true);
+    REQUIRE((unknown_args == std::vector<std::string>{"--verbose", "FOO", "5",
+                                                      "BAR", "-jn", "spam"}));
+  }
+}