feat(clp-s): Add command line options for stubbed out kv-pair-IR inge…

…stion. (y-scope#618)
Bill-hbrhbr · Dec 19, 2024 · 604bd75 · 604bd75
1 parent 36892c1
commit 604bd75
Show file tree

Hide file tree

Showing 4 changed files with 44 additions and 2 deletions.
diff --git a/components/core/src/clp_s/CommandLineArguments.cpp b/components/core/src/clp_s/CommandLineArguments.cpp
@@ -148,6 +148,9 @@ CommandLineArguments::parse_arguments(int argc, char const** argv) {
             po::options_description compression_options("Compression options");
             std::string metadata_db_config_file_path;
             std::string input_path_list_file_path;
+            constexpr std::string_view cJsonFileType{"json"};
+            constexpr std::string_view cKeyValueIrFileType{"kv-ir"};
+            std::string file_type{cJsonFileType};
             // clang-format off
             compression_options.add_options()(
                     "compression-level",
@@ -202,6 +205,10 @@ CommandLineArguments::parse_arguments(int argc, char const** argv) {
                     "disable-log-order",
                     po::bool_switch(&m_disable_log_order),
                     "Do not record log order at ingestion time."
+            )(
+                    "file-type",
+                    po::value<std::string>(&file_type)->value_name("FILE_TYPE")->default_value(file_type),
+                    "The type of file being compressed (json or kv-ir)"
             );
             // clang-format on
 
@@ -255,6 +262,22 @@ CommandLineArguments::parse_arguments(int argc, char const** argv) {
                 throw std::invalid_argument("No input paths specified.");
             }
 
+            if (cJsonFileType == file_type) {
+                m_file_type = FileType::Json;
+            } else if (cKeyValueIrFileType == file_type) {
+                m_file_type = FileType::KeyValueIr;
+                if (m_structurize_arrays) {
+                    SPDLOG_ERROR(
+                            "Invalid combination of arguments; --file-type {} and "
+                            "--structurize-arrays can't be used together",
+                            cKeyValueIrFileType
+                    );
+                    return ParsingResult::Failure;
+                }
+            } else {
+                throw std::invalid_argument("Unknown FILE_TYPE: " + file_type);
+            }
+
             // Parse and validate global metadata DB config
             if (false == metadata_db_config_file_path.empty()) {
                 clp::GlobalMetadataDBConfig metadata_db_config;

diff --git a/components/core/src/clp_s/CommandLineArguments.hpp b/components/core/src/clp_s/CommandLineArguments.hpp
@@ -36,6 +36,11 @@ class CommandLineArguments {
         Stdout,
     };
 
+    enum class FileType : uint8_t {
+        Json = 0,
+        KeyValueIr
+    };
+
     // Constructors
     explicit CommandLineArguments(std::string const& program_name) : m_program_name(program_name) {}
 
@@ -116,6 +121,8 @@ class CommandLineArguments {
 
     bool get_record_log_order() const { return false == m_disable_log_order; }
 
+    [[nodiscard]] auto get_file_type() const -> FileType { return m_file_type; }
+
 private:
     // Methods
     /**
@@ -184,6 +191,7 @@ class CommandLineArguments {
     size_t m_target_ordered_chunk_size{};
     size_t m_minimum_table_size{1ULL * 1024 * 1024};  // 1 MB
     bool m_disable_log_order{false};
+    FileType m_file_type{FileType::Json};
 
     // Metadata db variables
     std::optional<clp::GlobalMetadataDBConfig> m_metadata_db_config;

diff --git a/components/core/src/clp_s/JsonParser.hpp b/components/core/src/clp_s/JsonParser.hpp
@@ -12,6 +12,7 @@
 
 #include "../clp/GlobalMySQLMetadataDB.hpp"
 #include "ArchiveWriter.hpp"
+#include "CommandLineArguments.hpp"
 #include "DictionaryWriter.hpp"
 #include "FileReader.hpp"
 #include "FileWriter.hpp"
@@ -29,6 +30,7 @@ using namespace simdjson;
 namespace clp_s {
 struct JsonParserOption {
     std::vector<std::string> file_paths;
+    CommandLineArguments::FileType input_file_type{CommandLineArguments::FileType::Json};
     std::string timestamp_key;
     std::string archives_dir;
     size_t target_encoded_size{};

diff --git a/components/core/src/clp_s/clp-s.cpp b/components/core/src/clp_s/clp-s.cpp
@@ -88,6 +88,7 @@ bool compress(CommandLineArguments const& command_line_arguments) {
 
     clp_s::JsonParserOption option{};
     option.file_paths = command_line_arguments.get_file_paths();
+    option.input_file_type = command_line_arguments.get_file_type();
     option.archives_dir = archives_dir.string();
     option.target_encoded_size = command_line_arguments.get_target_encoded_size();
     option.max_document_size = command_line_arguments.get_max_document_size();
@@ -113,9 +114,17 @@ bool compress(CommandLineArguments const& command_line_arguments) {
     }
 
     clp_s::JsonParser parser(option);
-    if (false == parser.parse()) {
-        SPDLOG_ERROR("Encountered error while parsing input");
+    if (CommandLineArguments::FileType::KeyValueIr == option.input_file_type) {
+        // Functionality Coming in later PR
+        //  -->Call new parsing function in Json Parser to parse IRv2 to archive
+        //  -->Check for error from parsing function
+        SPDLOG_ERROR("Compressing Key Value IR Files is not yet supported");
         return false;
+    } else {
+        if (false == parser.parse()) {
+            SPDLOG_ERROR("Encountered error while parsing input");
+            return false;
+        }
     }
     parser.store();
     return true;