---
active_crumb: Docs
layout: documentation
id: overview
---

<!--
 Licensed to the Apache Software Foundation (ASF) under one or more
 contributor license agreements.  See the NOTICE file distributed with
 this work for additional information regarding copyright ownership.
 The ASF licenses this file to You under the Apache License, Version 2.0
 (the "License"); you may not use this file except in compliance with
 the License.  You may obtain a copy of the License at

      http://www.apache.org/licenses/LICENSE-2.0

 Unless required by applicable law or agreed to in writing, software
 distributed under the License is distributed on an "AS IS" BASIS,
 WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License.
-->

<div class="col-md-8 second-column">
    <section id="overview">
        <h2 class="section-title">Library API review <a href="#"><i class="top-link fas fa-fw fa-angle-double-up"></i></a></h2>

        <p>
            NlpCraft library contains two base concepts: <code>Model</code> and <code>Client</code> which have API representations
            <a href="apis/latest/org/apache/nlpcraft/NCModel.html">NCModel</a> and
            <a href="apis/latest/org/apache/nlpcraft/NCModelClient.html">NCModelClient</a>.
            When you work with the system - you should prepare model configuring its parameters and defining its components.
            After you just communicate with this model via client's methods.
        </p>

        <ul>
            <li>
                <code>Model</code> is domain specific object which responsible for user input interpretation.
                <code>Model</code> contains intents, defined via NlpCraft IDL with related code callbacks.
                Intent is user defined callback and rule, according to which this callback should be called.
                Rule is most often some template, based on expected set of entities in user input, but it can be more flexible.
            </li>

            <li>
                <code>Client</code> is object, which allows to communicate with the given model.
            </li>
        </ul>

        <p>Typical part of code:</p>

        <pre class="brush: scala, highlight: []">
              // Prepares domain model.
              val mdl = new CustomNlpModel()

              // Prepares client for given model.
              val client = new NCModelClient(mdl)

              // Sends text request to model by user ID "userId".
              val result = client.ask("Some user command", "userId")

              // Clears dialog session for user with ID "userId".
              client.clearDialog("userId")
        </pre>

        <p>
            <code>Model</code> definition includes two parts:
        </p>
        <ul>
            <li>
                <code>Configuration</code>. Static configuration parameters including name, version, etc.
            </li>
            <li>
                <code>Pipeline</code>. Most important component, which defines user input processing chain.
                <code>Pipeline</code> can be based on standard and custom user defined components.
            </li>
        </ul>
    </section>

    <section id="model">
        <h2 class="section-title">Model overview<a href="#"><i class="top-link fas fa-fw fa-angle-double-up"></i></a></h2>
        <p>
            Let's start with terminology and describe work workflow.
        </p>

        <ul>
            <li>
                <code>Token</code> represented as <a  href="apis/latest/org/apache/nlpcraft/NCToken.html">NCToken</a>.
                It is simple string, part of user input, which split according to some rules,
                for instance by spaces and some additional conditions, which depends on language and some expectations.
                So user input "<b>Where is it?</b>" contains four tokens: "<b>Where</b>", "<b>is</b>", "<b>it</b>", "<b>?</b>".
                Usually <code>tokens</code>are words and punctuation symbols which can also contain some additional
                information like point of speech etc.
                Tokens data are input for searching <code>entities</code>.
            </li>
            <li>
                <code>Entity</code> represented as <a  href="apis/latest/org/apache/nlpcraft/NCEntity.html">NCEntity</a>.
                According to wikipedia, named entity is a real-world object, such as a person, location, organization,
                product, etc., that can be denoted with a proper name. It can be abstract or have a physical existence.
                Each entity can contain one or more tokens.
                Entities data are input for searching <a href="intent-matching.html">Intent matching</a> conditions.
            </li>
            <li>
                <code>Variant</code> represented as <a  href="apis/latest/org/apache/nlpcraft/NCVariant.html">NCVariant</a>.
                List of entities. Potentially, each token can be recognized as different entities,
                so user input can be processed as set of variants.
                For example user input "Mercedes" can be processed as 2 variants,
                both of them contains single element list of entities: car brand or Spanish family name.
                When words are not overlapped with different <code>entities</code> there is only one
                <code>variant</code> detected.
            </li>
        </ul>

        <p>
            <code>Model</code> must be able to do tree following things:
        </p>

        <ul>
            <li>
                Parse user input text as the <code>tokens</code>.
                They are input for searching <code>name entities</code>.
                Tokens parsing components should be included into <a href="#model-pipeline">Model pipeline</a>.
            </li>
            <li>
                Find <code>name entities</code> based on these parsed tokens.
                <code>name entities</code>.
                They are input for searching <code>intents</code>.
                Entity parsing components should be included into <a href="#model-pipeline">Model pipeline</a>.
            </li>
            <li>
                Prepare callback methods which contain business logic
                and rules for matching user requests on them.
                These callbacks with their rules and named as <code>intents</code>.
                These matched callback methods execution with parameters extracted from user text input and
                these execution results returned to user as answer on his request.
                These callbacks methods should be defined in the model or model should have reference in them.
                It will be described below.
            </li>
        </ul>

        <p>
            Let's prepare the system which can call persons from your contact list.
            Typical commands are: "Please call to John Smith" or "Connect me with Barbara Dillan".
            This model should be able to recognize in user text following entities:
            <code>command</code> and <code>person</code> to apply this command.
        </p>

        <ul>
            <li>
                Parsing split your input on tokens ["<code>please</code>", "<code>call</code>", "<code>to</code>", "<code>john</code>", "<code>smith</code>"]
            </li>
            <li>
                By these tokens model should be able to found two named entities:
                <ul>
                    <li>
                        <code>command</code> by token <code>call</code>.
                    </li>
                    <li>
                        <code>person</code> by tokens <code>john</code> and <code>smith</code>.
                    </li>
                </ul>
            </li>
            <li>
                Also intents should be prepared:
                <pre class="brush: scala, highlight: [1, 2, 5, 6]">
                    @NCIntent("intent=call term(command)={# == command'} term(person)={# == 'person'}")
                    def onMatch(
                        ctx: NCContext,
                        im: NCIntentMatch,
                        @NCIntentTerm("command") command: NCEntity,
                        @NCIntentTerm("person") person: NCEntity
                    ): NCResult = ...
                </pre>

                <ul>
                    <li>
                        <code>Line 1</code> defines intent <code>ls</code> with two conditions.
                    </li>
                    <li>
                        <code>Line 2</code> defines related callback method.
                    </li>
                    <li>
                        <code>Lines 4 and 5</code> define two callback method's arguments which are corresponded to
                        <code>call</code> intent conditions. You can extract normalized value
                        <code>john smith</code> from the <code>person</code> parameter and use in the method body.
                    </li>
                </ul>

                <p>
                    Note that there are a lot of options of defining intents and their callback methods.
                    They can be defined in configuration files, nested classes, etc.
                    Look at the chapter <a href="intent-matching.html">Intent Matching</a> content for get more details.
                </p>
            </li>
        </ul>

    </section>

    <section id="client">
        <h2 class="section-title">Client overview<a href="#"><i class="top-link fas fa-fw fa-angle-double-up"></i></a></h2>

        <p>
            Base client methods:
        </p>
        <ul>
            <li>
                <code>ask()</code> passes user text input to the model and receives back execution
                <a href="apis/latest/org/apache/nlpcraft/NCResult.html">NCResult</a> or
                rejection exception if there isn't any triggered intents.
                <a href="apis/latest/org/apache/nlpcraft/NCResult.html">NCResult</a> is wrapper on
                callback method execution result with additional information.
            </li>
            <li>
                <code>debugAsk()</code> passes user text input to the model and receives back callback and its parameters or
                rejection exception if there isn't any triggered intents.
                Main difference from <code>ask</code> that triggered intent callback method is not called.
                This method and this parameter can be useful for tests scenarios.
            </li>
            <li>
                <code>clearStm()</code> clears STM state. Memory is cleared wholly or with some predicate.
                Loot at <a href="short-term-memory.html">Conversation</a> chapter to get more details.
            </li>
            <li>
                <code>clearDialog()</code> clears dialog state. Dialog is cleared wholly or with some predicate.
                Loot at <a href="short-term-memory.html">Conversation</a> chapter to get more details.
            </li>
            <li>
                <code>close()</code> - Closes client. You can't call another client's methods after this method was closed.
            </li>
        </ul>
    </section>

    <section id="model-configuration">
        <h2 class="section-title">Model configuration <a href="#"><i class="top-link fas fa-fw fa-angle-double-up"></i></a></h2>

        <p>
            <code>Model configuration</code> which is represented as is set of model parameter values.
            Its API representation is <a href="apis/latest/org/apache/nlpcraft/NCModelConfig.html">NCModelConfig</a>.
        </p>
        <ul>
            <li>
                <code>id</code>, <code>name</code> and <code>version</code> are mandatory model descriptors.
            </li>
            <li>
                <code>description</code> and <code>origin</code> are optional model descriptors.
            </li>
            <li>
                <code>conversationTimeout</code> - timeout of the user's conversation.
                If user doesn't communicate with the model this time period STM is going to be cleared.
                Loot at <a href="short-term-memory.html">Conversation</a> chapter to get more details.
                Mandatory parameter with default value.
            </li>
            <li>
                <code>conversationDepth</code> - maximum supported depth the user's conversation.
                Loot at <a href="short-term-memory.html">Conversation</a> chapter to get more details.
                Mandatory parameter with default value.
            </li>
        </ul>
    </section>

    <section id="model-pipeline">
        <h2 class="section-title">Model pipeline <a href="#"><i class="top-link fas fa-fw fa-angle-double-up"></i></a></h2>

        <p>
            Model <code>Pipeline</code> is represented as <a href="apis/latest/org/apache/nlpcraft/NCPipeline.html">NCPipeline</a> and
            contains following components:
        </p>
        <ul>
            <li>
                <code>Token parser</code> represented as <a href="apis/latest/org/apache/nlpcraft/NCTokenParser.html">NCTokenParser</a>.
                Mandatory pipeline component, it is required for parsing plain text, user input, and split this text
                into tokens  list. NlpCraft provides default EN implementation of token parser.
                Also, project contain various examples for <a href="examples/light_switch_fr.html">French</a> and
                <a href="examples/light_switch_ru.html">Russia</a> languages.
            </li>
            <li>
                <code>Tokens enrichers</code> represented as <a href="apis/latest/org/apache/nlpcraft/NCTokenEnricher.html">NCTokenEnricher</a> optional list.
                Tokens enricher is component which allows to add additional properties to prepared tokens,
                like part of speech, quote, stop-words flags or any other.
                NlpCraft provides default set of EN tokens enrichers implementations.
            </li>
            <li>
                <code>Tokens validators</code> represented as <a href="apis/latest/org/apache/nlpcraft/NCTokenValidator.html">NCTokenValidator</a> optional list.
                Tokens validator is user defined component, where tokens are inspected and exception can be thrown
                from user code to break user input processing.
            </li>
            <li>
                <code>Entity parsers</code> represented as <a href="apis/latest/org/apache/nlpcraft/NCEntityParser.html">NCEntityParser</a> mandatory list.
                At least one entity parser must be defined. Having prepared tokens as input,
                each entity parser tries to find user defined named entities.
                NlpCraft provides wrappers for named-entity recognition components of OpenNLP and Stanford libraries.
            </li>
            <li>
                <code>Entity enrichers</code> represented as <a href="apis/latest/org/apache/nlpcraft/NCEntityEnricher.html">NCEntityEnricher</a> optional list.
                Entity enricher is component which allows to add additional properties to prepared entities.
                Can be useful for extending existing entity enrichers functionality.
            </li>
            <li>
                <code>Entity mappers</code> represented as <a href="apis/latest/org/apache/nlpcraft/NCEntityMapper.html">NCEntityMapper</a> optional list.
                Entity mapper is component which allows to map one set of entities into another after the entities
                were parsed and enriched. Can be useful for building complex parsers based on existing.
            </li>
            <li>
                <code>Entity validators</code> represented as <a href="apis/latest/org/apache/nlpcraft/NCEntityValidator.html">NCEntityValidator</a> optional list.
                Entities validator is user defined component, where prepared entities are inspected and
                exceptions can be thrown from user code to break user input processing.
            </li>
            <li>
                <code>Variant filter</code> represented as <a href="apis/latest/org/apache/nlpcraft/NCVariantFilter.html">NCVariantFilter</a>.
                Optional component which allows filtering detected variants, rejecting undesirable.
            </li>
        </ul>

        <p>
            Below <code>Model</code> creation example.
            <code>Pipeline</code> is prepared using <code>NCPipelineBuilder</code> class helper.
        </p>

        <pre class="brush: scala, highlight: []">
            val pipeline =
                new NCPipelineBuilder().
                    withTokenParser(new NCFrTokenParser()).
                    withTokenEnricher(new NCFrLemmaPosTokenEnricher()).
                    withTokenEnricher(new NCFrStopWordsTokenEnricher()).
                    withEntityParser(new NCFrSemanticEntityParser("lightswitch_model_fr.yaml")).
                    build
            val cfg = NCModelConfig("nlpcraft.lightswitch.fr.ex", "LightSwitch Example Model FR", "1.0")

            val mdl = new NCModelAdapter(cfg, pipeline)
        </pre>

        <p>
            This flexible system allows to create any pipelines on any language.
            You can collect NlpCraft predefined components, write your own and easy reuse custom components.
        </p>
    </section>

    <section id="model-behavior">
        <h2 class="section-title">Model behavior overriding<a href="#"><i class="top-link fas fa-fw fa-angle-double-up"></i></a></h2>

        <p>
            There are also several <a href="/apis/latest/org/apache/nlpcraft/NCModel.html">NCModel</a>
            callbacks that you can override to affect model behavior during
            <a href="/intent-matching.html#model_callbacks">intent matching</a>
            to perform logging, debugging, statistic or usage collection, explicit update or initialization of
            conversation context, security audit or validation:
        </p>
        <ul>
            <li>
                Overriding <code>onMatchedIntent()</code> allows to reject matched intent and continue matching process.
            </li>
            <li>
                Overriding <code>onResult()</code> allows to replace callback method execution result.
            </li>
            <li>
                Overriding <code>onRejection()</code> allows to change operation result when rejection occurs.
            </li>
            <li>
                Overriding  <code>onError()</code> allows to change operation result when any error occurs.
            </li>
        </ul>
    </section>
</div>
<div class="col-md-2 third-column">
    <ul class="side-nav">
        <li class="side-nav-title">On This Page</li>
        <li><a href="#overview">Overview</a></li>
        <li><a href="#model">Model overview</a></li>
        <li><a href="#client">Client overview</a></li>
        <li><a href="#model-configuration">Model configuration</a></li>
        <li><a href="#model-pipeline">Model pipeline</a></li>
        <li><a href="#model-behavior">Model behavior overriding</a></li>
        {% include quick-links.html %}
    </ul>
</div>