/// <summary>
/// Returns the encoding of the input file
/// </summary>
/// <param name="srcFile"></param>
/// <returns></returns>
public static Encoding GetFileEncoding(string srcFile)
{
var encoding = Encoding.Default;
if (string.IsNullOrEmpty(srcFile) || !File.Exists(srcFile))
{
return(encoding);
}
// Read in the file in binary
byte[] buffer;
try {
using (var file = new FileStream(srcFile, FileMode.Open, FileAccess.Read, FileShare.ReadWrite)) {
using (var memoryStream = new MemoryStream()) {
file.CopyTo(memoryStream);
buffer = memoryStream.ToArray();
}
}
} catch (Exception) {
return(encoding);
}
// Detect encoding
var textDetect = new TextEncodingDetect();
var textEnc = textDetect.DetectEncoding(buffer, buffer.Length);
switch (textEnc)
{
case EncodingEnum.Ascii:
// ASCII (chars in the 0-127 range)
encoding = Encoding.ASCII;
break;
case EncodingEnum.Ansi:
// ANSI (chars in the range 0-255 range)
encoding = Encoding.Default;
break;
case EncodingEnum.Utf8Bom:
case EncodingEnum.Utf8Nobom:
// UTF-8
encoding = Encoding.UTF8;
break;
case EncodingEnum.Utf16LeBom:
case EncodingEnum.Utf16LeNobom:
// UTF-16 Little Endian
encoding = Encoding.Unicode;
break;
case EncodingEnum.Utf16BeBom:
case EncodingEnum.Utf16BeNobom:
// UTF-16 Big Endian
encoding = Encoding.BigEndianUnicode;
break;
}
return(encoding);
}
/// <summary>
/// Read all the text of a file in one go, same as File.ReadAllText expect it's truly a read only function
/// </summary>
public static string ReadAllText(string path, Encoding encoding = null)
{
try {
using (var fileStream = new FileStream(path, FileMode.Open, FileAccess.Read, FileShare.ReadWrite)) {
using (var textReader = new StreamReader(fileStream, encoding ?? TextEncodingDetect.GetFileEncoding(path))) {
return(textReader.ReadToEnd());
}
}
} catch (Exception e) {
throw new Exception($"Can not read the file {path.PrettyQuote()}.", e);
}
}
/// <summary>
/// Reads all the line of either the filePath (if the file exists) or from byte array dataResources,
/// Apply the action toApplyOnEachLine to each line
/// Uses encoding as the Encoding to read the file or convert the byte array to a string
/// Uses the char # as a comment in the file
/// </summary>
public static bool ForEachLine(string filePath, byte[] dataResources, Action <int, string> toApplyOnEachLine, Encoding encoding = null, Action <Exception> onException = null)
{
encoding = encoding ?? TextEncodingDetect.GetFileEncoding(filePath);
var wentOk = true;
try {
SubForEachLine(filePath, dataResources, toApplyOnEachLine, encoding);
} catch (Exception e) {
wentOk = false;
onException?.Invoke(e);
// read default file, if it fails then we can't do much but to throw an exception anyway...
if (dataResources != null)
{
SubForEachLine(null, dataResources, toApplyOnEachLine, encoding);
}
}
return(wentOk);
}
