The code below is designed to take a string in and remove any of a set of arbitrary words that are considered non-essential to a search phrase.
I didn't write the code, but need to incorporate it into something else. It works, and that's good, but it just feels wrong to me. However, I can't seem to get my head outside the box that this method has created to think of another approach.
Maybe I'm just making it more complicated than it needs to be, but I feel like this might be cleaner with a different technique, perhaps by using LINQ.
I would welcome any suggestions; including the suggestion that I'm over thinking it and that the existing code is perfectly clear, concise and performant.
So, here's the code:
private string RemoveNonEssentialWords(string phrase)
{
//This array is being created manually for demo purposes. In production code it's passed in from elsewhere.
string[] nonessentials = {"left", "right", "acute", "chronic", "excessive", "extensive",
"upper", "lower", "complete", "partial", "subacute", "severe",
"moderate", "total", "small", "large", "minor", "multiple", "early",
"major", "bilateral", "progressive"};
int index = -1;
for (int i = 0; i < nonessentials.Length; i++)
{
index = phrase.ToLower().IndexOf(nonessentials[i]);
while (index >= 0)
{
phrase = phrase.Remove(index, nonessentials[i].Length);
phrase = phrase.Trim().Replace(" ", " ");
index = phrase.IndexOf(nonessentials[i]);
}
}
return phrase;
}
Thanks in advance for your help.
Cheers,
Steve
This appears to be an algorithm for removing stop words from a search phrase.
Here's one thought: If this is in fact being used for a search, do you need the resulting phrase to be a perfect representation of the original (with all original whitespace intact), but with stop words removed, or can it be "close enough" so that the results are still effectively the same?
One approach would be to tokenize the phrase (using the approach of your choice - could be a regex, I'll use a simple split) and then reassemble it with the stop words removed. Example:
public static string RemoveStopWords(string phrase, IEnumerable<string> stop)
{
var tokens = Tokenize(phrase);
var filteredTokens = tokens.Where(s => !stop.Contains(s));
return string.Join(" ", filteredTokens.ToArray());
}
public static IEnumerable<string> Tokenize(string phrase)
{
return string.Split(phrase, ' ');
// Or use a regex, such as:
// return Regex.Split(phrase, #"\W+");
}
This won't give you exactly the same result, but I'll bet that it's close enough and it will definitely run a lot more efficiently. Actual search engines use an approach similar to this, since everything is indexed and searched at the word level, not the character level.
I guess your code is not doing what you want it to do anyway. "moderated" would be converted to "d" if I'm right. To get a good solution you have to specify your requirements a bit more detailed. I would probably use Replace or regular expressions.
I would use a regular expression (created inside the function) for this task. I think it would be capable of doing all the processing at once without having to make multiple passes through the string or having to create multiple intermediate strings.
private string RemoveNonEssentialWords(string phrase)
{
return Regex.Replace(phrase, // input
#"\b(" + String.Join("|", nonessentials) + #")\b", // pattern
"", // replacement
RegexOptions.IgnoreCase)
.Replace(" ", " ");
}
The \b at the beginning and end of the pattern makes sure that the match is on a boundary between alphanumeric and non-alphanumeric characters. In other words, it will not match just part of the word, like your sample code does.
Yeah, that smells.
I like little state machines for parsing, they can be self-contained inside a method using lists of delegates, looping through the characters in the input and sending each one through the state functions (which I have return the next state function based on the examined character).
For performance I would flush out whole words to a string builder after I've hit a separating character and checked the word against the list (might use a hash set for that)
I would create A Hash table of Removed words parse each word if in the hash remove it only one time through the array and I believe that creating a has table is O(n).
How does this look?
foreach (string nonEssent in nonessentials)
{
phrase.Replace(nonEssent, String.Empty);
}
phrase.Replace(" ", " ");
If you want to go the Regex route, you could do it like this. If you're going for speed it's worth a try and you can compare/contrast with other methods:
Start by creating a Regex from the array input. Something like:
var regexString = "\\b(" + string.Join("|", nonessentials) + ")\\b";
That will result in something like:
\b(left|right|chronic)\b
Then create a Regex object to do the find/replace:
System.Text.RegularExpressions.Regex regex = new System.Text.RegularExpressions.Regex(regexString, System.Text.RegularExpressions.RegexOptions.IgnoreCase);
Then you can just do a Replace like so:
string fixedPhrase = regex.Replace(phrase, "");
Related
What is the best way to achieve the following result?
input string: this is a {test} for {performance}.
output string[]: ["this is a ", "{test}", " for ", "{performance}", "."]
This is what I have so far:
private void StringToArray(string text)
{
var firstSplit = text.Split('{');
var list = new List<string>();
foreach(var s in firstSplit)
{
if (s.Contains("}"))
{
var secondSplit = s.Split('}');
list.Add("{" + secondSplit[0] + "}");
if(secondSplit.Count() > 1)
{
list.Add(secondSplit[1]);
}
}
else
{
list.Add(s);
}
}
Console.WriteLine(string.Join(",", list.ToArray()));
}
If you want to get all computer-science-y, you could build a state machine that walks the string one character at a time, using a stack to track the start position for your next string section. When you find a { or } character, you pop the prior stack value and use it to create a substring you can add to your array. Then you push the current index onto the stack.
This can be more or less complex based on how carefully you want to handle nesting (ie: "This is { a {test}.}"), escape characters (ie: {{ or \{), unbalanced strings (ie: This {is} a {Test), whitespace within brace values (ie: This {is some} text) or excluding things like quoted strings (ie: "This is the text, "A {person} lived there."") But a character-by-character state machine will tend to be the best-performing option.
At a higher level, you can look at a domain specific language, or using a lexxer/parser tool. Unfortunately, there's very little in the middle. RegEx can be made to work, but it's generally awful for this because your input can't be said to be formally "regular". Splitting on word boundaries will be error-prone. Basic string manipulation (ie: IndexOf()) is slower and no simpler or easier to write than the state machine would be.
I'm currently trying to strip a string of data that is may contain the hyphen symbol.
E.g. Basic logic:
string stringin = "test - 9894"; OR Data could be == "test";
if (string contains a hyphen "-"){
Strip stringin;
output would be "test" deleting from the hyphen.
}
Console.WriteLine(stringin);
The current C# code i'm trying to get to work is shown below:
string Details = "hsh4a - 8989";
var regexItem = new Regex("^[^-]*-?[^-]*$");
string stringin;
stringin = Details.ToString();
if (regexItem.IsMatch(stringin)) {
stringin = stringin.Substring(0, stringin.IndexOf("-") - 1); //Strip from the ending chars and - once - is hit.
}
Details = stringin;
Console.WriteLine(Details);
But pulls in an Error when the string does not contain any hyphen's.
How about just doing this?
stringin.Split('-')[0].Trim();
You could even specify the maximum number of substrings using overloaded Split constructor.
stringin.Split('-', 1)[0].Trim();
Your regex is asking for "zero or one repetition of -", which means that it matches even if your input does NOT contain a hyphen. Thereafter you do this
stringin.Substring(0, stringin.IndexOf("-") - 1)
Which gives an index out of range exception (There is no hyphen to find).
Make a simple change to your regex and it works with or without - ask for "one or more hyphens":
var regexItem = new Regex("^[^-]*-+[^-]*$");
here -------------------------^
It seems that you want the (sub)string starting from the dash ('-') if original one contains '-' or the original string if doesn't have dash.
If it's your case:
String Details = "hsh4a - 8989";
Details = Details.Substring(Details.IndexOf('-') + 1);
I wouldn't use regex for this case if I were you, it makes the solution much more complex than it can be.
For string I am sure will have no more than a couple of dashes I would use this code, because it is one liner and very simple:
string str= entryString.Split(new [] {'-'}, StringSplitOptions.RemoveEmptyEntries)[0];
If you know that a string might contain high amount of dashes, it is not recommended to use this approach - it will create high amount of different strings, although you are looking just for the first one. So, the solution would look like something like this code:
int firstDashIndex = entryString.IndexOf("-");
string str = firstDashIndex > -1? entryString.Substring(0, firstDashIndex) : entryString;
you don't need a regex for this. A simple IndexOf function will give you the index of the hyphen, then you can clean it up from there.
This is also a great place to start writing unit tests as well. They are very good for stuff like this.
Here's what the code could look like :
string inputString = "ho-something";
string outPutString = inputString;
var hyphenIndex = inputString.IndexOf('-');
if (hyphenIndex > -1)
{
outPutString = inputString.Substring(0, hyphenIndex);
}
return outPutString;
I am trying to create a bad word filter method that I can call before every insert and update to check the string for any bad words and replace with "[Censored]".
I have an SQL table with has a list of bad words, I want to bring them back and add them to a List or string array and check through the string of text that has been passed in and if any bad words are found replace them and return a filtered string back.
I am using C# for this.
Please see this "clbuttic" (or for your case cl[Censored]ic) article before doing a string replace without considering word boundaries:
http://www.codinghorror.com/blog/2008/10/obscenity-filters-bad-idea-or-incredibly-intercoursing-bad-idea.html
Update
Obviously not foolproof (see article above - this approach is so easy to get around or produce false positives...) or optimized (the regular expressions should be cached and compiled), but the following will filter out whole words (no "clbuttics") and simple plurals of words:
const string CensoredText = "[Censored]";
const string PatternTemplate = #"\b({0})(s?)\b";
const RegexOptions Options = RegexOptions.IgnoreCase;
string[] badWords = new[] { "cranberrying", "chuffing", "ass" };
IEnumerable<Regex> badWordMatchers = badWords.
Select(x => new Regex(string.Format(PatternTemplate, x), Options));
string input = "I've had no cranberrying sleep for chuffing chuffings days -
the next door neighbour is playing classical music at full tilt!";
string output = badWordMatchers.
Aggregate(input, (current, matcher) => matcher.Replace(current, CensoredText));
Console.WriteLine(output);
Gives the output:
I've had no [Censored] sleep for [Censored] [Censored] days - the next door neighbour is playing classical music at full tilt!
Note that "classical" does not become "cl[Censored]ical", as whole words are matched with the regular expression.
Update 2
And to demonstrate a flavour of how this (and in general basic string\pattern matching techniques) can be easily subverted, see the following string:
"I've had no cranberryıng sleep for chuffıng chuffıngs days - the next door neighbour is playing classical music at full tilt!"
I have replaced the "i"'s with Turkish lower case undottted "ı"'s. Still looks pretty offensive!
Although I'm a big fan of Regex, I think it won't help you here. You should fetch your bad word into a string List or string Array and use System.String.Replace on your incoming message.
Maybe better, use System.String.Split and .Join methods:
string mayContainBadWords = "... bla bla ...";
string[] badWords = new string[]{"bad", "worse", "worst"};
string[] temp = string.Split(badWords, StringSplitOptions.RemoveEmptyEntries);
string cleanString = string.Join("[Censored]", temp);
In the sample, mayContainBadWords is the string you want to check; badWords is a string array, you load from your bad word sql table and cleanString is your result.
you can use string.replace() method or RegEx class
There is also a nice article about it which can e found here
With a little html-parsing skills, you can get a large list with swear words from noswear
Is there any .Net library to remove all problematic characters of a string and only leave alphanumeric, hyphen and underscore (or similar subset) in an intelligent way? This is for using in URLs, file names, etc.
I'm looking for something similar to stringex which can do the following:
A simple prelude
"simple English".to_url =>
"simple-english"
"it's nothing at all".to_url =>
"its-nothing-at-all"
"rock & roll".to_url =>
"rock-and-roll"
Let's show off
"$12 worth of Ruby power".to_url =>
"12-dollars-worth-of-ruby-power"
"10% off if you act now".to_url =>
"10-percent-off-if-you-act-now"
You don't even wanna trust Iconv for this next part
"kick it en Français".to_url =>
"kick-it-en-francais"
"rock it Español style".to_url =>
"rock-it-espanol-style"
"tell your readers 你好".to_url =>
"tell-your-readers-ni-hao"
You can try this
string str = phrase.ToLower(); //optional
str = str.Trim();
str = Regex.Replace(str, #"[^a-z0-9\s_]", ""); // invalid chars
str = Regex.Replace(str, #"\s+", " ").Trim(); // convert multiple spaces into one space
str = str.Substring(0, str.Length <= 400 ? str.Length : 400).Trim(); // cut and trim it
str = Regex.Replace(str, #"\s", "-");
Perhaps this question here can help you on your way. It gives you code on how Stackoverflow generates its url's (more specifically, how question names are turned into nice urls.
Link to Question here, where Jeff Atwood shows their code
From your examples, the closest thing I've found (although I don't think it does everything that you're after) is:
My Favorite String Extension Methods in C#
and also:
ÜberUtils - Part 3 : Strings
Since neither of these solutions will give you exactly what you're after (going from the examples in your question) and assuming that the goal here is to make your string "safe", I'd second Hogan's advice and go with Microsoft's Anti Cross Site Scripting Library, or at least use that as a basis for something that you create yourself, perhaps deriving from the library.
Here's a link to a class that builds a number of string extension methods (like the first two examples) but leverages Microsoft's AntiXSS Library:
Extension Methods for AntiXss
Of course, you can always combine the algorithms (or similar ones) used within the AntiXSS library with the kind of algorithms that are often used in websites to generate "slug" URL's (much like Stack Overflow and many blog platforms do).
Here's an example of a good C# slug generator:
Improved C# Slug Generator
You could use HTTPUtility.UrlEncode, but that would encode everything, and not replace or remove problematic characters. So your spaces would be + and ' would be encoded as well. Not a solution, but maybe a starting point
If the goal is to make the string "safe" I recommend Mirosoft's anti-xss libary
There will be no library capable of what you want since you are stating specific rules that you want applied, e.g. $x => x-dollars, x% => x-percent. You will almost certainly have to write your own method to acheive this. It shouldn't be too difficult. A string extension method and use of one or more Regex's for making the replacements would probably be quite a nice concise way of doing it.
e.g.
public static string ToUrl(this string text)
{
return text.Trim().Regex.Replace(text, ..., ...);
}
Something the Ruby version doesn't make clear (but the original Perl version does) is that the algorithm it's using to transliterate non-Roman characters is deliberately simplistic -- "better than nothing" in both senses. For example, while it does have a limited capability to transliterate Chinese characters, this is entirely context-insensitive -- so if you feed it Japanese text then you get gibberish out.
The advantage of this simplistic nature is that it's pretty trivial to implement. You just have a big table of Unicode characters and their corresponding ASCII "equivalents". You could pull this straight from the Perl (or Ruby) source code if you decide to implement this functionality yourself.
I'm using something like this in my blog.
public class Post
{
public string Subject { get; set; }
public string ResolveSubjectForUrl()
{
return Regex.Replace(Regex.Replace(this.Subject.ToLower(), "[^\\w]", "-"), "[-]{2,}", "-");
}
}
I couldn't find any library that does it, like in Ruby, so I ended writing my own method. This is it in case anyone cares:
/// <summary>
/// Turn a string into something that's URL and Google friendly.
/// </summary>
/// <param name="str"></param>
/// <returns></returns>
public static string ForUrl(this string str) {
return str.ForUrl(true);
}
public static string ForUrl(this string str, bool MakeLowerCase) {
// Go to lowercase.
if (MakeLowerCase) {
str = str.ToLower();
}
// Replace accented characters for the closest ones:
char[] from = "ÂÃÄÀÁÅÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖØÙÚÛÜÝàáâãäåçèéêëìíîïðñòóôõöøùúûüýÿ".ToCharArray();
char[] to = "AAAAAACEEEEIIIIDNOOOOOOUUUUYaaaaaaceeeeiiiidnoooooouuuuyy".ToCharArray();
for (int i = 0; i < from.Length; i++) {
str = str.Replace(from[i], to[i]);
}
// Thorn http://en.wikipedia.org/wiki/%C3%9E
str = str.Replace("Þ", "TH");
str = str.Replace("þ", "th");
// Eszett http://en.wikipedia.org/wiki/%C3%9F
str = str.Replace("ß", "ss");
// AE http://en.wikipedia.org/wiki/%C3%86
str = str.Replace("Æ", "AE");
str = str.Replace("æ", "ae");
// Esperanto http://en.wikipedia.org/wiki/Esperanto_orthography
from = "ĈĜĤĴŜŬĉĝĥĵŝŭ".ToCharArray();
to = "CXGXHXJXSXUXcxgxhxjxsxux".ToCharArray();
for (int i = 0; i < from.Length; i++) {
str = str.Replace(from[i].ToString(), "{0}{1}".Args(to[i*2], to[i*2+1]));
}
// Currencies.
str = new Regex(#"([¢€£\$])([0-9\.,]+)").Replace(str, #"$2 $1");
str = str.Replace("¢", "cents");
str = str.Replace("€", "euros");
str = str.Replace("£", "pounds");
str = str.Replace("$", "dollars");
// Ands
str = str.Replace("&", " and ");
// More aesthetically pleasing contractions
str = str.Replace("'", "");
str = str.Replace("’", "");
// Except alphanumeric, everything else is a dash.
str = new Regex(#"[^A-Za-z0-9-]").Replace(str, "-");
// Remove dashes at the begining or end.
str = str.Trim("-".ToCharArray());
// Compact duplicated dashes.
str = new Regex("-+").Replace(str, "-");
// Let's url-encode just in case.
return str.UrlEncode();
}
I have some site content that contains abbreviations. I have a list of recognised abbreviations for the site, along with their explanations. I want to create a regular expression which will allow me to replace all of the recognised abbreviations found in the content with some markup.
For example:
content: This is just a little test of the memb to see if it gets picked up.
Deb of course should also be caught here.
abbreviations: memb = Member; deb = Debut;
result: This is just a little test of the [a title="Member"]memb[/a] to see if it gets picked up.
[a title="Debut"]Deb[/a] of course should also be caught here.
(This is just example markup for simplicity).
Thanks.
EDIT:
CraigD's answer is nearly there, but there are issues. I only want to match whole words. I also want to keep the correct capitalisation of each word replaced, so that deb is still deb, and Deb is still Deb as per the original text. For example, this input:
This is just a little test of the memb.
And another memb, but not amemba.
Deb of course should also be caught here.deb!
First you would need to Regex.Escape() all the input strings.
Then you can look for them in the string, and iteratively replace them by the markup you have in mind:
string abbr = "memb";
string word = "Member";
string pattern = String.Format("\b{0}\b", Regex.Escape(abbr));
string substitue = String.Format("[a title=\"{0}\"]{1}[/a]", word, abbr);
string output = Regex.Replace(input, pattern, substitue);
EDIT: I asked if a simple String.Replace() wouldn't be enough - but I can see why regex is desirable: you can use it to enforce "whole word" replacements only by making a pattern that uses word boundary anchors.
You can go as far as building a single pattern from all your escaped input strings, like this:
\b(?:{abbr_1}|{abbr_2}|{abbr_3}|{abbr_n})\b
and then using a match evaluator to find the right replacement. This way you can avoid iterating the input string more than once.
Not sure how well this will scale to a big word list, but I think it should give the output you want (although in your question the 'result' seems identical to 'content')?
Anyway, let me know if this is what you're after
using System;
using System.Collections.Generic;
using System.Linq;
using System.Text.RegularExpressions;
namespace ConsoleApplication1
{
class Program
{
static void Main(string[] args)
{
var input = #"This is just a little test of the memb to see if it gets picked up.
Deb of course should also be caught here.";
var dictionary = new Dictionary<string,string>
{
{"memb", "Member"}
,{"deb","Debut"}
};
var regex = "(" + String.Join(")|(", dictionary.Keys.ToArray()) + ")";
foreach (Match metamatch in Regex.Matches(input
, regex /*#"(memb)|(deb)"*/
, RegexOptions.IgnoreCase | RegexOptions.ExplicitCapture))
{
input = input.Replace(metamatch.Value, dictionary[metamatch.Value.ToLower()]);
}
Console.Write (input);
Console.ReadLine();
}
}
}
For anyone interested, here is my final solution. It is for a .NET user control. It uses a single pattern with a match evaluator, as suggested by Tomalak, so there is no foreach loop. It's an elegant solution, and it gives me the correct output for the sample input while preserving correct casing for matched strings.
public partial class Abbreviations : System.Web.UI.UserControl
{
private Dictionary<String, String> dictionary = DataHelper.GetAbbreviations();
protected void Page_Load(object sender, EventArgs e)
{
string input = "This is just a little test of the memb. And another memb, but not amemba to see if it gets picked up. Deb of course should also be caught here.deb!";
var regex = "\\b(?:" + String.Join("|", dictionary.Keys.ToArray()) + ")\\b";
MatchEvaluator myEvaluator = new MatchEvaluator(GetExplanationMarkup);
input = Regex.Replace(input, regex, myEvaluator, RegexOptions.IgnoreCase);
litContent.Text = input;
}
private string GetExplanationMarkup(Match m)
{
return string.Format("<b title='{0}'>{1}</b>", dictionary[m.Value.ToLower()], m.Value);
}
}
The output looks like this (below). Note that it only matches full words, and that the casing is preserved from the original string:
This is just a little test of the <b title='Member'>memb</b>. And another <b title='Member'>memb</b>, but not amemba to see if it gets picked up. <b title='Debut'>Deb</b> of course should also be caught here.<b title='Debut'>deb</b>!
I doubt it will perform better than just doing normal string.replace, so if performance is critical measure (refactoring a bit to use a compiled regex). You can do the regex version as:
var abbrsWithPipes = "(abbr1|abbr2)";
var regex = new Regex(abbrsWithPipes);
return regex.Replace(html, m => GetReplaceForAbbr(m.Value));
You need to implement GetReplaceForAbbr, which receives the specific abbr being matched.
I'm doing pretty exactly what you're looking for in my application and this works for me:
the parameter str is your content:
public static string GetGlossaryString(string str)
{
List<string> glossaryWords = GetGlossaryItems();//this collection would contain your abbreviations; you could just make it a Dictionary so you can have the abbreviation-full term pairs and use them in the loop below
str = string.Format(" {0} ", str);//quick and dirty way to also search the first and last word in the content.
foreach (string word in glossaryWords)
str = Regex.Replace(str, "([\\W])(" + word + ")([\\W])", "$1<span class='glossaryItem'>$2</span>$3", RegexOptions.IgnoreCase);
return str.Trim();
}