/// <summary>
/// Returns all possible matches from a given list of files
/// </summary>
/// <param name="files">A list containing the information for each file</param>
/// <returns>A dictionary where the key is the file hash and the value is the list of files that are identical></returns>
private static Dictionary<string, List<FileHashPair>> GetMatchesFromFiles(List<FileInfo> files)
{
Dictionary<string, List<FileHashPair>> possibleMatches = new Dictionary<string, List<FileHashPair>>();
List<FileHashPair> fileHashPairs = new List<FileHashPair>();
ProgressInfo info = new ProgressInfo();
// We can eliminate a lot of possibilities by comparing file size.
// Each file that does not have a matching file with the same size cannot possibly have
// the same file hash as another file and can be safely discarded.
if (files.Count > 2)
{
files = new List<FileInfo>(files.OrderBy(p => p.Length));
List<int> indexesToRemove = new List<int>();
for (int i = 0; i < files.Count; i++)
{
if (i == 0 && files[i].Length != files[i + 1].Length)
{
indexesToRemove.Add(i);
}
else if (i == files.Count - 1 && files[i].Length != files[i - 1].Length)
{
indexesToRemove.Add(i);
}
else if ((i != 0 && i != files.Count - 1) &&
files[i].Length != files[i - 1].Length && files[i].Length != files[i + 1].Length)
{
indexesToRemove.Add(i);
}
}
int indicesRemoved = 0;
foreach (var index in indexesToRemove)
{
files.RemoveAt(index - indicesRemoved);
indicesRemoved++;
}
}
// Calculate the hash value for each file
foreach (FileInfo file in files)
{
FileHashPair fileHashPair = new FileHashPair(file.FullName, Cryptography.GetMD5Hash(file), file.Length);
fileHashPairs.Add(fileHashPair);
if (HashProgressUpdate != null)
{
HashProgressUpdate(fileHashPairs.Count, files.Count);
}
if (progressWorker != null && progressWorker.IsBusy && progressWorker.WorkerReportsProgress)
{
info.FilesProcessed = fileHashPairs.Count;
info.TotalNumberOfFiles = files.Count;
progressWorker.ReportProgress((int)((fileHashPairs.Count / (double)files.Count) * 100), info);
}
if (progressWorker != null && progressWorker.WorkerSupportsCancellation && progressWorker.CancellationPending)
{
possibleMatches.Clear();
return possibleMatches;
}
}
foreach (FileHashPair fileHashPair in fileHashPairs)
{
if (possibleMatches.ContainsKey(fileHashPair.FileHash))
{
(possibleMatches[fileHashPair.FileHash]).Add(fileHashPair);
}
else
{
possibleMatches.Add(fileHashPair.FileHash, new List<FileHashPair>());
possibleMatches[fileHashPair.FileHash].Add(fileHashPair);
}
}
var orphanKeys = new List<string>();
foreach (var possibleMatch in possibleMatches.Keys)
{
if (possibleMatches[possibleMatch].Count == 1)
{
orphanKeys.Add(possibleMatch);
}
}
foreach (var orphanKey in orphanKeys)
{
possibleMatches.Remove(orphanKey);
}
return possibleMatches;
}
public bool Equals(FileHashPair fileHashPair)
{
if ((object)fileHashPair == null)
{
return false;
}
return FileHash != null && (FileHash == fileHashPair.FileHash);
}
