Delete Git Duplicates (useful for the Obsdian Git Plugin)

deleteGitDuplicates.py

"terminal version"

import os
import platform
try:    
    from tqdm import tqdm
    USETQDM = True
except ImportError:
    USETQDM = False

HOSTNAME = platform.node()

def deleteDuplicateMdFiles(dir):
    duplicates = []
    for root, dirs, files in os.walk(dir):
            duplicates += [os.path.join(root, f) for f in files if f.endswith(HOSTNAME + ".md") and f.replace(f"-{HOSTNAME}", "") in files]
    
    for d in duplicates:
        print("- " + d)
        
    print(f"Found {len(duplicates)} duplicate files.")
    
    if USETQDM:
        for file in tqdm(duplicates)):
            deleteDuplicate(dir, file, False)
    else:
        for file in duplicates:
            deleteDuplicate(dir, file, True)

def deleteDuplicate(dir, file, printInfo):
    filePath = os.path.join(dir, file)
    os.remove(filePath)
    if printInfo: print(f"Deleted: {filePath}")

if __name__ == "__main__":
    target_directory = input("Enter the path of the directory (. for the current working directory): ")
    if target_directory == ".": target_directory = os.getcwd()
    
    if os.path.isdir(target_directory):
        deleteDuplicateMdFiles(target_directory)
        print("Finished process.")
    else:
        print("Invalid directory path.")

deleteGitDuplicatesAuto.py

"automatic version using the current working directory"

import os
import platform

HOSTNAME = platform.node()

def deleteDuplicateMdFiles(dir):
    duplicates = []
    for root, dirs, files in os.walk(dir):
            duplicates += [os.path.join(root, f) for f in files if f.endswith(HOSTNAME + ".md") and f.replace(f"-{HOSTNAME}", "") in files]
    
    for file in duplicates:
        deleteDuplicate(dir, file, False)

def deleteDuplicate(dir, file, printInfo):
    filePath = os.path.join(dir, file)
    os.remove(filePath)
    if printInfo: print(f"Deleted: {filePath}")

if __name__ == "__main__":
    target_directory = os.getcwd()
    deleteDuplicateMdFiles(target_directory)

deleteGitDuplicatesAuto.ts

// typescript-equivalent, not tested yet

import * as fs from 'fs';
import * as path from 'path';
import os from 'os';

const HOSTNAME = os.hostname();

function deleteDuplicate(dir: string, file: string): void {
    const filePath = path.join(dir, file);
    fs.unlinkSync(filePath);
    console.log(`Deleted: ${filePath}`);
}

if (require.main === module) {
    const targetDirectory = process.cwd();
    const files = fs.readdirSync(targetDirectory);

    files.forEach(file => {
        if (file.endsWith(`${HOSTNAME}.md`) && files.includes(file.replace(`-${HOSTNAME}`, ''))) {
            deleteDuplicate(targetDirectory, file);
        }
    });
}

deleteGitDuplicatesMin.py

"smaller version without tqdm"

import os
import sys
import platform

HOSTNAME = platform.node()

def deleteDuplicateMdFiles(dir):
    duplicates = []
    for root, dirs, files in os.walk(dir):
            duplicates += [os.path.join(root, f) for f in files if f.endswith(HOSTNAME + ".md") and f.replace(f"-{HOSTNAME}", "") in files]

    print(f"Found {len(duplicates)} duplicate files.")

    for file in duplicates:
        deleteDuplicate(dir, file, True)

def deleteDuplicate(dir, file, printInfo):
    filePath = os.path.join(dir, file)
    os.remove(filePath)
    if printInfo: print(f"Deleted: {filePath}")

if __name__ == "__main__":
    target_directory = input("Enter the path of the directory (. for the current working directory): ")
    if target_directory == ".": target_directory = os.getcwd()
    
    if os.path.isdir(target_directory):
        deleteDuplicateMdFiles(target_directory)
        print("Finished process.")
    else:
        print("Invalid directory path.")