Create new index.

Example

import ElmTextSearch


{-| Example document type.
-}
type alias ExampleDocType =
    { cid : String
    , title : String
    , author : String
    , body : String
    }


{-| Create an index with default configuration.
See ElmTextSearch.SimpleConfig documentation for parameter information.
-}
createNewIndexExample : ElmTextSearch.Index ExampleDocType
createNewIndexExample =
    ElmTextSearch.new
        { ref = .cid
        , fields =
            [ ( .title, 5.0 )
            , ( .body, 1.0 )
            ]
        , listFields = []
        }

The SimpleConfig parameter to new is

• ref
  • The unique document reference will be extracted from each document using .cid.
• fields
  • Define which fields contain a strings to be indexed.
  • The following fields will be indexed from each document
    • .title
    • .body
  • When searching the index any word matches found in the .title field (boost value 5.0) raise the document match score more than if found in the .body field (boost value 1.0).
    • The document match score determines the order of the list of matching documents returned.
• listFields
  • Define which fields contain list of strings to be indexed.

Create new index with additional configuration.

Example.

import ElmTextSearch
import Index.Defaults
import StopWordFilter


type alias ExampleDocType =
    { cid : String
    , title : String
    , author : String
    , body : String
    }


createMyStopWordFilter =
    StopWordFilter.createFilterFuncWith
        [ "explanations" ]


createNewWithIndexExample : ElmTextSearch.Index ExampleDocType
createNewWithIndexExample =
    ElmTextSearch.newWith
        { indexType = "ElmTextSearch - Customized Stop Words v1"
        , ref = .cid
        , fields =
            [ ( .title, 5.0 )
            , ( .body, 1.0 )
            ]
        , listFields = []
        , initialTransformFactories = Index.Defaults.defaultInitialTransformFactories
        , transformFactories = Index.Defaults.defaultTransformFactories
        , filterFactories = [ createMyStopWordFilter ]
        }

Add a document to an index.

Starting with the ElmTextSearch.new example above this adds a document.

addDocToIndexExample : Result String (ElmTextSearch.Index ExampleDocType)
addDocToIndexExample =
    ElmTextSearch.add
        { cid = "id1"
        , title = "First Title"
        , author = "Some Author"
        , body = "Words in this example document with explanations."
        }
        createNewWithIndexExample

Conditions that cause a result Err with message.

• Error document ref is empty.
• Error after tokenisation there are no terms to index.
• Error adding document that allready exists.

Add multiple documents. Tries to add all docs and collects errors.. It does not stop adding at first error encountered.

The result part List (Int, String) is the list of document index and the error string message result of adding. Returns the index unchanged if all documents error when added. Returns the updated index after adding the documents.

Remove a document from an index.

Starting with the ElmTextSearch.new example above this removes a document.

removeDocFromIndexExample =
    ElmTextSearch.remove
        { cid = "123"
        , title = "Examples of a Banana"
        , author = "Sally Apples"
        , body = "Sally writes words about a banana."
        }
        createNewIndexExample

Conditions that cause a result Err with message.

• Error document has an empty unique id (ref).
• Error document is not in index.

Update a document in an index.

Starting with the ElmTextSearch.new example above this updates a document.

  updatedIndex =
    ElmTextSearch.update
      { cid = "123"
      , title = "Examples of a Bananas in every day life."
      , author = "Sally Apples"
      , body = "Sally writes more words about a banana."
      }
      createNewIndexExample

Conditions that cause an error result are those for ElmTextSearch.remove and ElmTextSearch.add.

Add or Update a document in an index. This removes the document first if it is allready in index then adds it.

Search an index with query.

Tokens are extracted from the query string and passed through the same processing used when indexing documents.

Each token is expanded, so that the term "he" might be expanded to "hello" and "help" if those terms were already included in the document index.

Multiple tokens are allowed and will lead to an AND based query.

The following example runs a search for documents containing both "apple" and "banana".

searchResult =
    Index.search "Apple banana" createNewIndexExample

Results are a list of matching document reference identifiers with there similarity to query score, ordered by score descending, so the best matches are earliest in the list.

An index is returned from search as well. This is because the data model may be updated to improve performance for later searches.

Adding or removing a new document will cause some of the internal caching to be reset.

Conditions that cause a result Err with message.

• Error there are no documents in index to search.
• Error query is empty.
• Error after tokenisation there are no terms to search for.

An Index holds the data to be able search for added documents.

A Config is required to create an Index.

A SimpleConfig is the least amount of configuration data required to create an Index.

See ElmTextSearch.new for fields.

Store an index to a Value. You can also use ElmTextSearch.Json.Encoder.

Store an index to a String. You can also use ElmTextSearch.Json.Encoder.

Create an Index from a String which has a stored Index in it and the supplied basic configurations. See ElmTextSearch.fromStringWith for possible Err results.

Create an Index from a Value which has a stored Index in it. See ElmTextSearch.fromStringWith for possible Err results.

Create an Index from a String which has a stored Index in it.

If none of the indexVersion in the list of Config match the index type being loaded it will return an Err.

The list of configurations wil be searched for a matching indexType so you should provide configs for all types you may be trying to load. No more than the config that matches is required though.

If the none of the supplied Config match the loaded Index then it will try if the index being loaded matches the default version if so it will still load the index.

The following Err results may be returned.

• "Error cannot load Index. Tried to load index of type "__IndexTest Type -". It is not in supported index configurations."
  • It contains the loaded version index type which comes from input.
• "Error cannot load Index. Version supported is 1.0.0. Version tried to load is 1.0.1."
  • It includes both expected and loaded versions which may vary.

Create an Index from a String which has a stored Index in it. If none of the indexVersion in the list of SimpleConfig match the index being decoded it will return an Err.

See ElmTextSearch.fromStringWith for possible Err results.