A simple free text search engine service.

1. index - given a string, tokenize it and store it along with its tokens
2. search - given a string, tokenize it and use a similarity function to rank how similar it is to previously indexed strings

1. This is a very basic implementation, where searching is likely not to scale well
2. Moreover, as the store becomes larger it may become impossible to store it in-memory. This basic implementation does not address this issue

1. WhitespaceTokenizer
2. CharacterTokenizer

1. SharedTokensSimilarity
2. LengthPercentSimilarity

• Maven - Dependency Management
• Spark Java - A lightweight http framework for java
• slf4j - A logging facade, allows easier transition between logging implementations
• slf4j-log4j12 - A binding of slf4j to log4j
• Gson - A Java serialization/deserialization library to convert Java Objects into JSON and back

1. Running this project requires maven and java (8+)
2. Clone (or download) the project
3. Run mvn dependency:copy-dependencies install in the projects folder
4. Run java -cp "target\FreeTextSearch-0.0.1-SNAPSHOT.jar;target\dependency\*" app.FreeTextSearch
   1. This may be slightly different on different operating systems
5. By default, step 4 will run the server with a character tokenizer and length percent similarity function
   1. You may choose to use a different tokenizer / similarity function
   2. To do that follow this example: java -cp "target\FreeTextSearch-0.0.1-SNAPSHOT.jar;target\dependency\*" app.FreeTextSearch word sharedtokens
   3. "word" (case doesn't matter) will start up the service with a word tokenizer (other values will start up the service with a character tokenizer)
   4. "sharedtokens" (case doesn't matter) will start up the service with a shared tokens similarity function (other values will start up the service with a length percent similarity function)
6. The service should be up and ready to receive http requests now

Assuming you've started the service with a word tokenizer and a shared tokens similarity function, the following series of http requests:

curl -d "the cow says moo" -X POST http://127.0.0.1:5000/index
curl -d "the cat and the hat" -X POST http://127.0.0.1:5000/index
curl -d " the dish ran away with the spoon" -X POST http://127.0.0.1:5000/index
curl -d " a cat ran away" -X POST http://127.0.0.1:5000/search

should yield the following output:

[
  {
    "matches": [
      {
        "matchedAgainst": "the dish ran away with the spoon",
        "similarity": 2
      },
      {
        "matchedAgainst": "the cat and the hat",
        "similarity": 1
      },
      {
        "matchedAgainst": "the cow says moo",
        "similarity": 0
      }
    ]
  }
]