I have a CloudSearch domain with a filename text field. My issue is that a text query won't match (some) documents with filenames I think it (logically) should. If I have documents with these filenames:
'cars'
'Cars Movie.jpg'
'cars.pdf'
'cars#.jpg'
and I perform a simple text query of 'cars', I get back files #1, #2, and #4 but not #3. If I search 'cars*' (or do a structured query using prefix) I can match #3. This doesn't make sense to me, especially that #4 matches but #3 does not.
TL;DR It's because of the way the tokenization algorithm handles periods.
When you perform a text search, you're performing a search against processed data, not the literal field. (Maybe that should've been obvious, but it wasn't how I was thinking about it before.)
The documentation gives an overview of how text is processed:
During indexing, Amazon CloudSearch processes text and text-array fields according to the analysis scheme configured for the field to determine what terms to add to the index. Before the analysis options are applied, the text is tokenized and normalized.
The part of the process that's ultimately causing this behavior is the tokenization:
During tokenization, the stream of text in a field is split into separate tokens on detectable boundaries using the word break rules defined in the Unicode Text Segmentation algorithm.
According to the word break rules, strings separated by whitespace such as spaces and tabs are treated as separate tokens. In many cases, punctuation is dropped and treated as whitespace. For example, strings are split at hyphens (-) and the at symbol (#). However, periods that are not followed by whitespace are considered part of the token.
The reason I was seeing the matches described in the question is because the file extensions are being included with whatever precedes them as a single token. If we look back at the example, and build an index according to these rules, it makes sense why a search of 'cars' returns documents #1, #2, and #4 but not #3.
# Text Index
1 'cars' ['cars']
2 'Cars Movie.jpg' ['cars', 'movie.jpg']
3 'cars.pdf'. ['cars.pdf']
4 'cars#.jpg' ['cars', '.jpg']
Possible Solutions
It might seem like setting a custom analysis scheme could fix this, but none of the options there (stopwords, stemming, synonyms) help you overcome the tokenization problem. I think the only possible solution, to get the desired behavior, is to tokenize the filename (using a custom algorithm) before upload, and then store the tokens in a text array field. Although devising a custom tokenization algorithm that supports multiple languages is a large problem.
Related
I have a string variable containing school names and I need to find all the possible combination of each word in this string variable in stata:
For example variation of a word "Academy" would be:
Academy,
Academy,
acdamey,
aacdemy,
dmcaamy,
aacedmy,
and so on.
I need this to standardize the raw data of school names, which has many typos of each word due to data entry issues, like the ones given above for "academy".
Depending whether your data is already in the Excel sheets or a file, you can either use regex trying to match all possible combinations (and probably fix them when found) or parse the strings first before bringing them into Excel. In either case you could make a file (or Excel list/table/area/etc.) that includes all the common typos and pick each typo as regex match to use when comparing to your actual input.
Making regexp that would actually find all possible cases is next to impossible, especially if there are cases where very similar (but correct) names for schools exist. In any case direct regexps would be very messy and complex, so I would advice you to parse the data by finding first the correct form, excluding it and then using (greedy) search/regex to find the typoed versions. You can then save the typos to use them as a filter/match/pattern.
To get some sort of starting ideas, check this links:
Regex: Search for verb roots
Read text file and extract string into Excel sheet using regex
P.s You should keep the count of all strings/school names and finally get a list of all names that did not match correct form or any of your regexp filters, so you can manually insert/correct them.
I have a text file in Notepad++ that contains about 66,000 words all in 1 line, and it is a set of 200 "lines" of output that are all unique and placed in 1 line in the basic JSON form {output:[{output1},{output2},...}]}.
There is a set of characters matching the RegEx expression "id":.........,"kind":"track" that occurs about 285 times in total, and I am trying to either single them out, or copy all of them at once.
Basically, without some super complicated RegEx terms, I am stuck because I can't figure out how to highlight all of them at once, and also the Remove Unbookmarked Lines feature does not apply because this is all in one line. I have only managed to be able to Mark every single occurrence.
So does this require a large number of steps to get the file into multiple lines and work from there, or is there something else I am missing?
Edit: I have come up with a set of Macro schemes that make the process of doing this manually work much faster. It's another alternative but still takes a few steps and quite some time.
Edit 2: I intended there to be an answer for actually just highlighting the different sections all at once, but I guess that it not possible. The answer here turns out to be more useful in my case, allowing me to have a list of IDs without everything else.
You seem to already have a regex which matches single instances of your pattern, so assuming it works and that we must use Notepad++ for this:
Replace .*?("id":.........,"kind":"track").*?(?="id".........,"kind":"track"|$) with \1.
If this textfile is valid JSON, this opens you up to other, non-notepad++ options, like using Python with the json module.
Edited to remove unnecessary steps
I've recently upgraded a CloudSearch instance from the 2011 to the 2013 API. Both instances have a field called sid, which is a text field containing a two-letter code followed by some digits e.g. LC12345. With the 2011 API, if I run a search like this:
q=12345*&return-fields=sid,name,desc
...I get back 1 result, which is great. But the sid of the result is LC12345 and that's the way it was indexed. The number 12345 does not appear anywhere else in any of the resulting document fields. I don't understand why it works. I can only assume that this type of query is looking for any terms in any fields that even contain the number 12345.
The reason I'm asking is because this functionality is now broken when I query using the 2013 API. I need to use the structured query parser, but even a comparable wildcard query using the simple parser is not working e.g.
q.parser=simple&q=12345*&return=sid,name,desc
...returns nothing, although the document is definitely there i.e. if I query for LC12345* it finds the document.
If I could figure out how to get the simple query working like it was before, that would at least get me started on how to do the same with the structured syntax.
Why it's not working
CloudSearch v1 (2011) had a different way of tokenizing mixed alpha+numeric strings. Here's the logic as described in the archived docs (emphasis mine).
If a string contains both alphabetic and numeric characters and is at
least three and no more than nine characters long, the alphabetic and
numeric portions of the string are treated as separate tokens. For
example, the string DOC298 is tokenized into two terms: doc 298
CloudSearch v2 (2013) text processing follows Unicode Text Segmentation, which does not specify that behavior:
Do not break within sequences of digits, or digits adjacent to letters (“3a”, or “A3”).
Solution
You should just be able to search *12345 to get back results with any prefix. There may be some edge cases like getting back results you don't want (things with more preceding digits like AB99912345); I don't know enough about your data to say whether those are real concerns.
Another option would would be to index the numeric prefix separately from the alphabetical suffix but that's additional work that may be unnecessary.
I'm guessing you are using Cloudsearch in English, so maybe this isn't your specific problem, but also watch out for Stopwords in your search queries:
https://docs.aws.amazon.com/cloudsearch/latest/developerguide/configuring-analysis-schemes.html#stopwords
In your example, the word "jo" is a stop word in Danish and another languages, and of course, supported languages, have a dictionary of stop words that has very common ones. If you don't specify a language in your text field, it will be English. You can see them here: https://docs.aws.amazon.com/cloudsearch/latest/developerguide/text-processing.html#text-processing-settings
I often grab quotes from articles that include citations that include superscripted footnotes, which when copied are a pain in the ass. They show up as actual letters in the text as they are pasted in plaintext and not in html.
Is there a way I could run this through a regex to take out these superscripts?
For example
In the abeginning bGod ccreated the dheaven and the eearth.
Should become
In the beginning God created the heaven and the earth.
I can't think of a way to have regex search for misspellings and a corresponding sequential set of numbers and letters.
Any thoughts? I'm also using Sublime Text 3 for the majority of my writing, but I wouldn't mind outsourcing this to an AppleScript, or text replacement app (aText, textExpander, etc.).
Matching Code vs. Matching a Screen
It's hard to tell without seeing an example, but this should be doable if you copy the text from code view, as opposed to the regular browser view. (Ctrl or Cmd-J is your friend). Since writing the rules will take time, this will only be worthwhile for large chunks of text.
In code view, your superscript will be marked up in a way that can be targetted by regex. For instance:
and therefore bananas make you smartera
in the browser view (where the a at the end is a citation note) may look like this in code view:
and therefore bananas make you smarter<span class="mycitations">a</span>
In your editor, using regex, you can process the text to remove all tags, or just certain tags. The rules may not always be easy to write, and of course there are many disclaimers about using regex to parse html.
However, if your source is always the same (Wikipedia for instance), then you can create and save rules that should work across many pages.
I recently discussed editors with a co-worker. He uses one of the less popular editors and I use another (I won't say which ones since it's not relevant and I want to avoid an editor flame war). I was saying that I didn't like his editor as much because it doesn't let you do find/replace with regular expressions.
He said he's never wanted to do that, which was surprising since it's something I find myself doing all the time. However, off the top of my head I wasn't able to come up with more than one or two examples. Can anyone here offer some examples of times when they've found regex find/replace useful in their editor? Here's what I've been able to come up with since then as examples of things that I've actually had to do:
Strip the beginning of a line off of every line in a file that looks like:
Line 25634 :
Line 632157 :
Taking a few dozen files with a standard header which is slightly different for each file and stripping the first 19 lines from all of them all at once.
Piping the result of a MySQL select statement into a text file, then removing all of the formatting junk and reformatting it as a Python dictionary for use in a simple script.
In a CSV file with no escaped commas, replace the first character of the 8th column of each row with a capital A.
Given a bunch of GDB stack traces with lines like
#3 0x080a6d61 in _mvl_set_req_done (req=0x82624a4, result=27158) at ../../mvl/src/mvl_serv.c:850
strip out everything from each line except the function names.
Does anyone else have any real-life examples? The next time this comes up, I'd like to be more prepared to list good examples of why this feature is useful.
Just last week, I used regex find/replace to convert a CSV file to an XML file.
Simple enough to do really, just chop up each field (luckily it didn't have any escaped commas) and push it back out with the appropriate tags in place of the commas.
Regex make it easy to replace whole words using word boundaries.
(\b\w+\b)
So you can replace unwanted words in your file without disturbing words like Scunthorpe
Yesterday I took a create table statement I made for an Oracle table and converted the fields to setString() method calls using JDBC and PreparedStatements. The table's field names were mapped to my class properties, so regex search and replace was the perfect fit.
Create Table text:
...
field_1 VARCHAR2(100) NULL,
field_2 VARCHAR2(10) NULL,
field_3 NUMBER(8) NULL,
field_4 VARCHAR2(100) NULL,
....
My Regex Search:
/([a-z_])+ .*?,?/
My Replacement:
pstmt.setString(1, \1);
The result:
...
pstmt.setString(1, field_1);
pstmt.setString(1, field_2);
pstmt.setString(1, field_3);
pstmt.setString(1, field_4);
....
I then went through and manually set the position int for each call and changed the method to setInt() (and others) where necessary, but that worked handy for me. I actually used it three or four times for similar field to method call conversions.
I like to use regexps to reformat lists of items like this:
int item1
double item2
to
public void item1(int item1){
}
public void item2(double item2){
}
This can be a big time saver.
I use it all the time when someone sends me a list of patient visit numbers in a column (say 100-200) and I need them in a '0000000444','000000004445' format. works wonders for me!
I also use it to pull out email addresses in an email. I send out group emails often and all the bounced returns come back in one email. So, I regex to pull them all out and then drop them into a string var to remove from the database.
I even wrote a little dialog prog to apply regex to my clipboard. It grabs the contents applies the regex and then loads it back into the clipboard.
One thing I use it for in web development all the time is stripping some text of its HTML tags. This might need to be done to sanitize user input for security, or for displaying a preview of a news article. For example, if you have an article with lots of HTML tags for formatting, you can't just do LEFT(article_text,100) + '...' (plus a "read more" link) and render that on a page at the risk of breaking the page by splitting apart an HTML tag.
Also, I've had to strip img tags in database records that link to images that no longer exist. And let's not forget web form validation. If you want to make a user has entered a correct email address (syntactically speaking) into a web form this is about the only way of checking it thoroughly.
I've just pasted a long character sequence into a string literal, and now I want to break it up into a concatenation of shorter string literals so it doesn't wrap. I also want it to be readable, so I want to break only after spaces. I select the whole string (minus the quotation marks) and do an in-selection-only replace-all with this regex:
/.{20,60} /
...and this replacement:
/$0"¶ + "/
...where the pilcrow is an actual newline, and the number of spaces varies from one incident to the next. Result:
String s = "I recently discussed editors with a co-worker. He uses one "
+ "of the less popular editors and I use another (I won't say "
+ "which ones since it's not relevant and I want to avoid an "
+ "editor flame war). I was saying that I didn't like his "
+ "editor as much because it doesn't let you do find/replace "
+ "with regular expressions.";
The first thing I do with any editor is try to figure out it's Regex oddities. I use it all the time. Nothing really crazy, but it's handy when you've got to copy/paste stuff between different types of text - SQL <-> PHP is the one I do most often - and you don't want to fart around making the same change 500 times.
Regex is very handy any time I am trying to replace a value that spans multiple lines. Or when I want to replace a value with something that contains a line break.
I also like that you can match things in a regular expression and not replace the full match using the $# syntax to output the portion of the match you want to maintain.
I agree with you on points 3, 4, and 5 but not necessarily points 1 and 2.
In some cases 1 and 2 are easier to achieve using a anonymous keyboard macro.
By this I mean doing the following:
Position the cursor on the first line
Start a keyboard macro recording
Modify the first line
Position the cursor on the next line
Stop record.
Now all that is needed to modify the next line is to repeat the macro.
I could live with out support for regex but could not live without anonymous keyboard macros.