Merge pull request #390 from AsherGlick/simplifying_reading_functions

Adding Zip file crawling
AsherGlick · Feb 10, 2025 · 13ed5c1 · 13ed5c1
2 parents ed94351 + 3f754fc
commit 13ed5c1
Show file tree

Hide file tree

Showing 2 changed files with 42 additions and 21 deletions.
diff --git a/xml_converter/src/file_helper.cpp b/xml_converter/src/file_helper.cpp
@@ -71,7 +71,7 @@ vector<MarkerPackFile> get_files_by_suffix(
         cout << "Error: " << base << " does not exist" << endl;
         return vector<MarkerPackFile>();
     }
-    // If it is a directory, call open_directory_file_to_read
+    // If it is a directory and parse the directory contents.
     else if (filesystem::is_directory(base)) {
         DIR* dir = opendir(join_file_paths(base, subpath).c_str());
         struct dirent* entry;
@@ -97,8 +97,29 @@ vector<MarkerPackFile> get_files_by_suffix(
         }
         closedir(dir);
     }
-    // else if (filesystem::is_regular_file(file.base)) {
-    // }
+    // If it is a file, assume it is a zip file and read the zip contents.
+    else if (filesystem::is_regular_file(base)) {
+        int error = 0;
+        zip* zip_archive = zip_open(base.c_str(), ZIP_RDONLY, &error);
+
+        if (zip_archive == nullptr) {
+            cerr << "Error: could not open the zip archive " << base << ". Got Error code " << error << endl;
+            return files;
+        }
+
+        int64_t num_entries = zip_get_num_entries(zip_archive, 0x00);
+        for (int64_t i = 0; i < num_entries; i++) {
+            const char* filename = zip_get_name(zip_archive, i, 0x00);
+
+            if (has_suffix(filename, suffix)) {
+                MarkerPackFile new_file(base, filename);
+                files.push_back(new_file);
+            }
+        }
+    }
+    else {
+        cerr << "Error: " << base << " is not a file or directory" << endl;
+    }
 
     sort(files.begin(), files.end(), marker_pack_file_comp);
     return files;

diff --git a/xml_converter/src/xml_converter.cpp b/xml_converter/src/xml_converter.cpp
@@ -41,17 +41,17 @@ map<string, vector<string>> read_taco_directory(
     if (!filesystem::exists(input_path)) {
         cout << "Error: " << input_path << " is not an existing directory or file" << endl;
     }
-    else if (filesystem::is_directory(input_path)) {
-        string directory_name = filesystem::path(input_path).filename();
-        vector<MarkerPackFile> xml_files = get_files_by_suffix(input_path, ".xml");
-        for (const MarkerPackFile& path : xml_files) {
-            set<string> top_level_category_names = parse_xml_file(path, marker_categories, parsed_pois);
-            string relative_path = join_file_paths(directory_name, path.relative_filepath);
-            for (set<string>::iterator it = top_level_category_names.begin(); it != top_level_category_names.end(); it++) {
-                top_level_category_file_locations[*it].push_back(relative_path);
-            }
+
+    string directory_name = filesystem::path(input_path).filename();
+    vector<MarkerPackFile> xml_files = get_files_by_suffix(input_path, ".xml");
+    for (const MarkerPackFile& path : xml_files) {
+        set<string> top_level_category_names = parse_xml_file(path, marker_categories, parsed_pois);
+        string relative_path = join_file_paths(directory_name, path.relative_filepath);
+        for (set<string>::iterator it = top_level_category_names.begin(); it != top_level_category_names.end(); it++) {
+            top_level_category_file_locations[*it].push_back(relative_path);
         }
     }
+
     return top_level_category_file_locations;
 }
 
@@ -63,17 +63,17 @@ map<string, vector<string>> read_burrito_directory(
     if (!filesystem::exists(input_path)) {
         cout << "Error: " << input_path << " is not an existing directory or file" << endl;
     }
-    else if (filesystem::is_directory(input_path)) {
-        string directory_name = filesystem::path(input_path).filename();
-        vector<MarkerPackFile> burrito_files = get_files_by_suffix(input_path, ".guildpoint");
-        for (const MarkerPackFile& path : burrito_files) {
-            set<string> top_level_category_names = read_protobuf_file(path, marker_categories, parsed_pois);
-            string relative_path = join_file_paths(directory_name, path.relative_filepath);
-            for (set<string>::iterator it = top_level_category_names.begin(); it != top_level_category_names.end(); it++) {
-                top_level_category_file_locations[*it].push_back(relative_path);
-            }
+
+    string directory_name = filesystem::path(input_path).filename();
+    vector<MarkerPackFile> burrito_files = get_files_by_suffix(input_path, ".guildpoint");
+    for (const MarkerPackFile& path : burrito_files) {
+        set<string> top_level_category_names = read_protobuf_file(path, marker_categories, parsed_pois);
+        string relative_path = join_file_paths(directory_name, path.relative_filepath);
+        for (set<string>::iterator it = top_level_category_names.begin(); it != top_level_category_names.end(); it++) {
+            top_level_category_file_locations[*it].push_back(relative_path);
         }
     }
+
     return top_level_category_file_locations;
 }