eng-elan.cmdi

<?xml version="1.0" encoding="UTF-8"?>
<!-- Anything outside of the RECOGNIZER element can be left untouched -->
<CMD CMDVersion="1.1" xmlns="http://www.clarin.eu/cmd/"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.clarin.eu/cmd/ http://catalog.clarin.eu/ds/ComponentRegistry/rest/registry/profiles/clarin.eu:cr1:p_1430905751617/xsd">
    <!-- See http://catalog.clarin.eu/ds/ComponentRegistry?registrySpace=published&itemId=clarin.eu:cr1:p_1430905751617 -->
    <Header></Header>
    <Resources>
        <ResourceProxyList></ResourceProxyList>
        <JournalFileProxyList></JournalFileProxyList>
        <ResourceRelationList></ResourceRelationList>
    </Resources>
    <Components>
        <RECOGNIZER>
            <!--
            For Recognizer API extensions the regonizerType is "direct", for stand-alone components running on the same
            computer the type should be "local". That are currently the types supported by ELAN.
            For "direct" the "run" attributes (runMac etc.) should have the value of the fully qualified name of the 
            class implementing "Recognizer.java". For "local" the executable command per platform should be the value.
            The "info" element contains the name as it will be shown in the user interface. Try to define a name
            that makes it likely to be unique, e.g by including an institution or university acronym.
            -->
            <recognizer recognizerType="local" runLinux="./eng-elan.sh" runMac="./eng-elan.sh" runWin="C:\Program Files\ELAN_6.4\app\extensions\ASR_English\eng-elan.bat"
                info="Eng automatic speech transcription">Eng_Automatic_Speech_Transcription</recognizer>

            <!-- For a more detailed documentation, a HTML file linked to the recognizer. This HTML ships with the 
                 recognizer and gives extra explanations for using the recognizer and information about licenses, 
                 authors and contacting them, version numbers etc. The link can be either a relative link to a file
                 in the same directory tree as the CMDI file or (not recommended) a direct http link to a recognizer 
                 homepage elsewhere. 
            -->
            <documentation>eng-elan.html</documentation>

            <!--
            Any number of input and output parameters can be specified.
            At least one input element (of type audio, video or timeseries) should be specified.
            Other input and output types (tier, timeseries etc.) and their r4equired formats are described 
            in the interfacing specification at https://tla.mpi.nl/projects_info/auvis/#more-7066 and 
            https://tla.mpi.nl/projects_info/avatech/
            (https://tla.mpi.nl/wp-content/uploads/2012/08/Avatech-interface-spec-2014-03-06.pdf)
            -->
            <input level="basic" type="audio" optional="false"
                mimetypes="audio/x-wav"
                info="Input audio file (WAV)">source</input>
            
            <!--
                Specify the url to your Google Cloud Project
                -->
            
            <textparam level="basic" optional="false" mimetypes="text/plain"
                info="Url to Google Cloud project">project</textparam>
                
            <!--
                Select your transcription service
                -->

            <textparam convoc="AWS GCloud Whisper" type="auxiliary" optional="false"
                info="Transcription service">transcription_service</textparam>
            
            <!--
                Select the annotation level
                -->

            <textparam convoc="Utterance Word Both" type="auxiliary" optional="false"
                info="Annotation level">level</textparam>

            <!--
                Select tier for annotation at utterance level
                -->

            <input level="basic" type="tier" optional="true" mimetypes="text/plain" 
            info="Select a tier for utterance-level annotation">tier</input>
            
            <!--
                Select the annotation level
                -->

            <numparam level="basic" type="int" min="1" max="500" default="100"
                info="Utterance threshold (smaller value = less silence between utterances)">threshold</numparam>

            <!--
                Name of the bucket where your audio should be stored
            -->
            <textparam level="basic" type="auxiliary" mimetypes="text/plain" optional="false" 
                info="Bucket name">bucket</textparam>

            <!--
                Name of your transcription file
                The name should be unique
            -->
            <textparam level="basic" type="auxiliary" mimetypes="text/plain" optional="false"
                info="Filename">filename</textparam>

             <!--
                Path where your transcription should be stored
            -->
            <input level="basic" type="auxiliary" optional="true" default="/transcriptions/transcription.json"
                info="Output path">output_path</input>

            <output level="basic" type="tier" optional="true"
                mimetypes="text/xml" default="/var/tmp/eng.xml" 
                info="Output recognized segments">output_segments</output>
        </RECOGNIZER>
    </Components>
</CMD>