/// <summary>
/// Combines the parsed tags with the current dictionary. Record any differences found, or existing
/// dictionary entries that are not found in the new list.
/// </summary>
static void Combine()
{
StreamWriter differences = new StreamWriter("differences.txt");
StreamWriter missing = new StreamWriter("missing.txt");
// tags contains what we were able to successfully parse this time
// compare that to the dictionary, to see what may be missing or different
// go through the known dictionary
foreach (Tag tag in EK.Capture.Dicom.DicomToolKit.Dictionary.Instance)
{
string key = tag.ToString().ToUpper();
// if it is already known, but was not found this time, add it in and record it as missing
if (!tags.ContainsKey(key))
{
missing.WriteLine(String.Format("Tag {0} {1}", key, tag.Description));
tags.Add(tag.ToString().ToUpper(), new Tag2(tag));
}
else
{
Tag2 temp = tags[key];
string check = temp.Check ? "*" : String.Empty;
// if we find any differences, record the differences
string difference = String.Empty;
if (tag.Description.Trim() != temp.Description.Trim())
{
}
if (tag.VR != temp.VR)
{
if (difference != String.Empty)
{
difference += " ";
}
difference += String.Format("vr {0} >> {1}", tag.VR, temp.VR);
}
if (tag.VM != temp.VM)
{
if (difference != String.Empty)
{
difference += " ";
}
difference += String.Format("vm {0} >> {1}", tag.VM, temp.VM);
}
if (difference.Length > 0)
{
differences.WriteLine(String.Format("{0}Tag {1} {2}", check, key, difference));
}
}
}
differences.Close();
differences.Dispose();
missing.Close();
missing.Dispose();
}
/// <summary>
/// Parse a combined line into a Tag
/// </summary>
/// <param name="line">The line to parse</param>
/// <returns>A Tag</returns>
static Tag2 Parse(string line, bool check)
{
// we break each line into words based on whitespace
string[] strings = line.Split(" \t".ToCharArray());
Tag2 tag = null;
try
{
tag = Tag2.Parse(strings[0].ToUpper().Replace('X', '0'));
// retired tags are at the end of the line
tag.Retired = (strings[strings.Length - 1] == "RET");
tag.VM = String.Empty;
tag.VR = String.Empty;
tag.Description = String.Empty;
try
{
// the next part of the line looks somethign like this
// Related General SOP Class UID RelatedGeneralSOPClassUI D UI 1-n
// the spaces are generally there because of the lines being appended
// we have the name with spaces, the name without spaces, VR and VM
int index = 2;
tag.Description = EK.Capture.Dicom.DicomToolKit.Dictionary.ModifyWordForDescription(strings[1]);
string first = EK.Capture.Dicom.DicomToolKit.Dictionary.ModifyWordForEnumeration(strings[1]).ToLower();
// we keep adding up the description until we encounter a word that
// starts with the first word, in the example above, the first word is Related
// and we want to stop adding up the description at RelatedGeneralSOPClassUI
while (!strings[index].ToLower().Contains(first))
{
tag.Description += " " + EK.Capture.Dicom.DicomToolKit.Dictionary.ModifyWordForDescription(strings[index++]);
}
tag.Check = false;
}
catch
{
// if we encounter any problems, i.e. never match the first word,
// we have to check the result
tag.Check = true;;
tag.Reason = "never matched first word";
}
// we have the tag, we have the description, now we just need to get the VM and VR
// there are four things to parse, tag, description, VR, and VM
if (strings.Length >= 4)
{
// VM is the last word, or second to the last if RET is present
tag.VM = (tag.Retired) ? strings[strings.Length - 2] : strings[strings.Length - 1];
// if we do not recognize the multiplicity, check it
if (!multiplicity.Contains(tag.VM))
{
tag.Check = true;
tag.Reason = "unknown value multiplicity";
tag.VM = "?";
}
// VR is the second or third to last string
tag.VR = (tag.Retired) ? strings[strings.Length - 3] : strings[strings.Length - 2];
// if we do recognize it, check it
if (!EK.Capture.Dicom.DicomToolKit.Dictionary.DicomTypes.Contains(tag.VR))
{
tag.Check = true;
tag.Reason = "unknown value representation";
tag.VR = "?";
}
if (tag.VM != "1" && !multipletypes.Contains(tag.VR))
{
multipletypes.Add(tag.VR);
}
// if the tag can have multiple VRs, we find the word "or", and we need to check
for (int n = 3; n < strings.Length; n++)
{
if (strings[n] == "or")
{
tag.Check = true;
tag.Reason = "possible multiple VRs";
tag.VR = "?";
}
}
}
// if we are going to check the tag, we may as well give it all the text we have.
if (tag.Check)
{
tag.Description = line.Substring(12);
}
}
catch
{
tag = null;
}
if (tag == null || tag.VM == String.Empty || tag.VR == String.Empty || tag.Description == String.Empty)
{
tag = null;
}
return(tag);
}
/// <summary>
/// Scans the input file and creates two sets of tags, those that are accepted and
/// those that are rejected. The results are written to two text files, and the
/// accepted tags are also added to a collection of tags.
/// </summary>
/// <param name="filename">The filename to parse</param>
static void Extract(string filename)
{
string line, next;
tags = new SortedDictionary <string, Tag2>();
StreamWriter cleaned = new StreamWriter("cleaned.txt");
StreamWriter rejected = new StreamWriter("rejected.txt");
StreamReader input = new StreamReader(filename);
line = input.ReadLine();
//as long as there are lines to read in the input file
while ((next = input.ReadLine()) != null)
{
// assume that we will get a clean parse of a tag
bool check = false;
// we can automatically reject some common lines
if (IsEasilyRejected(line))
{
rejected.WriteLine(line);
}
else
{
// some lines are split into several lines
// we need to combine partial lines until they fit the proper format
// the proper format is a dicom tag followed by a description and keyword, a VR, a VM, and then maybe a "RET"
// the tag is strings[0]
// "RET" may be at strings[strings.Length - 1]
// if the tag is not retired the VR is at strings[strings.Length - 2]
// the description is everything after the tag and before strings[strings.Length - 3]
// append each next line from the input as long as
// they do not start with a DICOM tag
while (!Matches(DicomTagPattern, next))
{
if (next != null)
{
// ignore the next line if it is a page number or other recognizable string
if (IsEasilyRejected(next))
{
rejected.WriteLine(next);
next = input.ReadLine();
continue;
}
}
// we have had to append lines to get a complete line
// we need to check the output to make sure.
check = true;
// append it with a space
line += " ";
line += next;
next = input.ReadLine();
if (next == null)
{
break;
}
}
// when we reach here, we may have a properly combined line
string key = line.Substring(0, 11).ToUpper();
// if the line begins with a DICOM tag
if (Matches(DicomTagPattern, key))
{
// try and parse what we have put together
Tag2 tag = Parse(line, check);
if (tag != null)
{
cleaned.WriteLine(line);
tags.Add(tag.Name.ToUpper(), tag);
}
else
{
rejected.WriteLine(line);
}
}
// reject everything else
else
{
rejected.WriteLine(line);
}
}
line = next;
}
// the above loop drops the last line.
input.Close();
input.Dispose();
cleaned.Close();
cleaned.Dispose();
rejected.Close();
rejected.Dispose();
}