List files and some statistics in directory for huge directories.

dirlist.cpp

/**
 * dirlist
 * A program to list files in a directory with some information as fast as possible.
 * Useful for directories with 500,000 files.
 *
 * @date 2023-10-23
 * @author relipse
 */ 
#include <sys/stat.h>
#include <time.h>
#include <stdio.h>
#include <cstring>
#include <dirent.h>
#include <locale.h>
#include <string>
#include <fstream>
#include <iostream>
#include <map>
#include <vector>
using namespace std;


void printHelp(int argc, char *argv[]){
	printf("%s [Options] [directory]\n", argv[0]);
	printf("List number of files and each file (if -f,--hide-files is not set)\n"
	       "in a directory as fast as possible,\n"
	       "also list alphabetical min and max (by filename)\n");
	printf("\nOptions:\n");
	printf("-f,--hide-files    Do not show each file\n");
	printf("-d,--dots          Use 1 period (dot) to represent each file\n");
	printf("-s,--file-size     Show file sizes and min/max (much slower)\n");
	printf("-c,--created       Show created/modified dates and min/max\n");
	printf("-g,--grep <match>  Only show file names matching <match>\n");
	printf("-h,--help          Show this page\n");
	printf("\nExamples:\n");
	printf("# ./dirlist\n"
	       "1. dirlist\n"
               "2. dirlist.cpp\n"
               "3. Makefile"
               "\n"
               "Directory: . contains 3 files\n"
               "Alphabetical Max: dirlist.cpp\n"
               "Alphabetical Min: Makefile\n");
	
	printf("\n# ./dirlist -d\n"
               "...\n"
	       "Directory: . contains 3 files\n"
               "Alphabetical Max: dirlist.cpp\n"
	       "Alphabetical Min: Makefile\n"
);
}


void printExtensionMap(std::map<string,unsigned long>& m){
	cout << "Extensions List (extension: count):" << endl;
	for(map<string,unsigned long>::const_iterator it = m.begin();
    it != m.end(); ++it)
	{	
    	    std::cout << "\"" << it->first << "\": " << it->second << endl;
	}
}

int main(int argc, char *argv[]) {
    DIR *dir;
    struct dirent *ent;
    unsigned long count = 0;
    string theDir = ".";
    string max = "";
    string min = "";
    string file;
    string fileTypes = "";
    bool printAllFiles = true;
    bool printDots = false;
    bool printSize = false;
    unsigned long fileSizeSum = 0;
    unsigned long maxSize = 0;
    unsigned long minSize = 0;
    unsigned long fileSize = 0;
    string maxSizeFileName = "";
    string minSizeFileName = "";
    string grep = "";
    std::vector<string> aryGrep;
    std::map<string,unsigned long> foundExts;
    setlocale(LC_NUMERIC, "");
    bool printMoreFileStats = false;
    string earliestCreatedDate = "";
    string latestCreatedDate = "";
    string earliestCreatedFile = "";
    string latestCreatedFile = "";
    if (argc > 1){
	    string arg;
	    for (int i = 1; i < argc; ++i){
		    arg = argv[i];
		    if (arg == "--hide-files" || arg == "-f"){
			printAllFiles = false;
		    }else if (arg == "-c" || arg == "--created"){
			printMoreFileStats = true;
		    }else if (arg == "-s" || arg == "--file-size"){
			printSize = true;
		    }else if (arg == "-h" || arg == "--help"){
			printHelp(argc, argv);	 
		        return 0;	
		    }else if (arg == "--dots" || arg == "-d"){
			printDots = true;
			printAllFiles = false;
		    }else if (arg == "--grep" || arg == "-g"){
			    if (i+1 >= argc){	
				    cout << "--grep|-g needs next argument (missing)." << endl;	    
				    printHelp(argc, argv);
				    return 0;
			    }else{
				    grep = argv[i+1];
				    aryGrep.push_back(grep);
				    i++;
			    }
		    }else{
			theDir = arg;
			if (theDir == ""){
				theDir = ".";
			}
		    }
	    }
    }
    if (aryGrep.size() > 0){
	    cout << "Grep on, -g,--grep matching all terms: ";
	    for(unsigned int i = 0; i < aryGrep.size(); ++i){
		    if (i > 0){ cout << ", "; }
		    cout << aryGrep[i];
	    }
	    cout << endl;
    }
    dir = opendir(theDir.c_str());
    while((ent = readdir(dir))){
	    file = ent->d_name;
            std::string::size_type idx;
	    idx = file.rfind('.');

	    if (file == ".." || file == "."){
		continue;
	    }
	    if (aryGrep.size() > 0){
		    unsigned int matches = 0;
		    for (unsigned int i = 0; i < aryGrep.size(); ++i){
			grep = aryGrep[i];
		 	if (file.find(grep) != string::npos){
		    		matches++;
		 	}
		    }
		    //all greps must match or we go to next file
		    if (matches != aryGrep.size()){
			//no grep matches found, go to next file
			continue;
		    }
	    }
            string extension = "";
            if(idx != std::string::npos)
            {
                extension = file.substr(idx+1);
            }
           
            if (!foundExts.count(extension)){
                       foundExts[extension] = 0;
            }
            foundExts[extension]++;
            ++count;
	    
	    if (max == "" || ent->d_name > max){
		max = ent->d_name;
	    }
	    if (min == "" || ent->d_name < min){
		    min = ent->d_name;
	    }
	    if (printAllFiles){

	    	printf("%'lu. %s", count, ent->d_name);
	        if (printMoreFileStats){
			string modtime = "";
			string createdtime = "";
			struct stat t_stat;
			stat(ent->d_name, &t_stat);
			struct tm * timeinfo = localtime(&t_stat.st_ctime);
			struct tm * timemod = localtime(&(t_stat.st_mtime));
			char buffer [80];
			strftime(buffer, 80, "%Y-%m-%d %H:%M:%S",timeinfo);
			string ct = buffer;
			strftime(buffer, 80, "%Y-%m-%d %H:%M:%S",timemod);
			string mt = buffer;
			if (earliestCreatedDate == "" || ct < earliestCreatedDate){
				earliestCreatedDate = ct;
				earliestCreatedFile = ent->d_name;
			}
			if (latestCreatedDate == "" || ct > latestCreatedDate){
				latestCreatedDate = ct;
				latestCreatedFile = ent->d_name;
			}


			createdtime = asctime(timeinfo);
			createdtime.erase (strcspn (createdtime.c_str(), "\r\n"));
			modtime = asctime(timemod);
			modtime.erase(strcspn(modtime.c_str(), "\r\n"));

			if (createdtime == modtime){
				cout << " (created/modified: " << createdtime;
			}else{
				cout << " (created: " << createdtime << ", modified: " << modtime;
			}
			cout << ")";
		}
		if (printSize){ // -s, --file-size
			ifstream in_file(ent->d_name, ios::binary);
			in_file.seekg(0, ios::end);
			fileSize = (unsigned long) in_file.tellg();
			fileSizeSum += fileSize;
			if (fileSize > maxSize){
				maxSize = fileSize;
				maxSizeFileName = ent->d_name;
			}
			if (minSize == 0 || fileSize < minSize){
				minSize = fileSize;
				minSizeFileName = ent->d_name;
			}

			printf(" %'lu bytes", fileSize);
		}
		printf("\n");
	    }
	    if (printDots){
		printf(".");
	    }
    }
    closedir(dir);
    if (printDots){
	printf("\n");
    }
    if (aryGrep.size() > 0){
	cout << "Directory " << theDir << " contains " << count << " matching files." << endl;
    }else{
    	printf("Directory: %s contains %'lu files\n", theDir.c_str(), count);
    }
    printf("Alphabetical Max: %s\n", max.c_str());
    printf("Alphabetical Min: %s\n", min.c_str());
    if (printSize){
	    printf("Largest File: %s %'lu bytes\n", maxSizeFileName.c_str(), maxSize);
	    printf("Smallest File: %s %'lu bytes\n", minSizeFileName.c_str(), minSize); 
	    printf("Total File Size Summation: %'lu bytes\n", fileSizeSum);
    }
    if (printMoreFileStats){
	    cout << "Earliest File: " << earliestCreatedFile << " (created: " << earliestCreatedDate << ")" << endl;
	    cout << "Latest File: " << latestCreatedFile << " (created: " << latestCreatedDate << ")" << endl;
    }
    printExtensionMap(foundExts);
    return 0;
}