The sphinxesc package

[Tags:library, mit, program]

Transform queries for sphinx input


[Skip to Readme]

Properties

Versions 0.1.0.0, 0.1.0.1
Dependencies base (>=4.8 && <5.0), MissingH (<1.5), optparse-applicative (<0.14), parsec (<3.2), sphinxesc, split (<0.3) [details]
License MIT
Author Daniel Choi
Maintainer Mackey RMS
Category Text
Home page https://github.com/mackeyrms/sphinxesc#readme
Bug tracker https://github.com/mackeyrms/sphinxesc/issues
Source repository head: git clone https://github.com/mackeyrms/sphinxesc
Uploaded Thu Mar 23 17:11:28 UTC 2017 by mackeyrms
Updated Thu Mar 23 17:28:08 UTC 2017 by HerbertValerioRiedel to revision 1
Distributions NixOS:0.1.0.1
Downloads 84 total (7 in the last 30 days)
Votes
0 []
Status Docs available [build log]
Last success reported on 2017-03-23 [all 1 reports]
Hackage Matrix CI

Modules

[Index]

Downloads

Maintainer's Corner

For package maintainers and hackage trustees

Readme for sphinxesc

Readme for sphinxesc-0.1.0.1

sphinxesc

A small module to prevent user-submitted search expressions from being mis-parsed into invalid Sphinx Extended Query Expressions.

The module provides a function

module SphinxEscape where
escapeSphinxQueryString :: String -> String

that sanitizes the Sphinx query expression in a way that can be safely submitted to the Sphinx API.

Synopsis

Example from ghci:

ghci> :m SphinxEscape 
ghci> putStrLn $ escapeSphinxQueryString "@tag_list hello OR quick brown fox 7/11"
@tag_list hello | quick brown fox 7 11
ghci> 
ghci> putStrLn $ escapeSphinxQueryString "hello AND quick brown fox 7/11"
hello & quick brown fox 7 11
ghci> 

Explanation

escapeSphinxQueryString performs very simple escaping with the help of a simplified abtract syntax tree. The abstract syntax tree it builds is:

data Expression = 
        TagFieldSearch String 
      | Literal String
      | Phrase String
      | AndOrExpr Conj Expression Expression 
  deriving Show

The escaping does not parse more advanced Sphinx query expressions such as NEAR/n, quorum, etc., nor does it recognize arbitrary @field expressions. The only special expressions recognized are & (AND), | (OR) and @tag_list WORDS. Except for quoted phrases, non-alphanumeric characters that do not form part of these specific expressions are simply turned into whitespace.

See the Testing section below for examples of conversions.

Obviously these rules are quite domain specific. The rules can be made more configurable later.

Testing

The command line executable sphinxesc can be used to test the expression parser and escaping of the input to the final sphinx search expression.

$ sphinxesc "test OR hello"
test | hello

# -p option shows the parsing result

$ sphinxesc -p "test OR hello"
AndOrExpr Or (Literal "test") (Literal "hello")

There is a suite of Bash-based regression tests in tests.txt, where the input is on the left, followed by :: surrounded by any whitespace, followed by the expected escaped output result. To run the tests, execute the script ./test.sh

NOTE This test output may be outdated. Please look at the tests.txt for the current tests.

./test.sh

INPUT                         EXPECTED                      RESULT                        PASS      
7/11                          7 11                          7 11                          PASS      
hello 7/11                    hello 7 11                    hello 7 11                    PASS      
hello OR 7/11                 hello | 7 11                  hello | 7 11                  PASS      
hello or 7/11                 hello | 7 11                  hello | 7 11                  PASS      
hello | 7/11                  hello | 7 11                  hello | 7 11                  PASS      
hello AND 7/11                hello & 7 11                  hello & 7 11                  PASS      
@tag_list fox tango 7/11      @tag_list fox tango 7 11      @tag_list fox tango 7 11      PASS      
@(tag_list) fox tango 7/11    @tag_list fox tango 7 11      @tag_list fox tango 7 11      PASS      
@(tag_list) AND               @tag_list AND                 @tag_list AND                 PASS      
@other_field AND              other field AND               other field AND               PASS      
hello & @other_field AND      hello &  other field AND      hello &  other field AND      PASS      
hello &                       hello                         hello                         PASS      
& hello &                     hello                         hello                         PASS      
& & hello &                   hello                         hello                         PASS      
| | hello |                   hello                         hello                         PASS      
"hello" hello                 hello  hello                  hello  hello                  PASS      
hello" hello                  hello  hello                  hello  hello                  PASS      
hello' hello                  hello  hello                  hello  hello                  PASS      
hello' @tag_list fox          hello   @tag_list fox         hello   @tag_list fox         PASS      
hello' @tag_list fox &        hello   @tag_list fox         hello   @tag_list fox         PASS      
                                                                                          PASS      

(The last case is hard to see, but the input is a blank string "" and the output is a blank string "".)

Future directions

The escaping function can be made more configurable. The parser and AST data structure can also be made more sophisticated, so that the AST can cover more of the Sphinx Extended Query syntax.

Reference