nopaque/app/templates/services/tesseract_ocr_pipeline.html.j2

{% extends "base.html.j2" %}
{% import "wtf.html.j2" as wtf %}

{% block main_attribs %}class="service-color lighten" data-service="tesseract-ocr-pipeline"{% endblock main_attribs %}

{% block page_content %}
<div class="container">
  <div class="row">
    <div class="col s12">
      <h1 id="title">{{ title }}</h1>
    </div>

    <div class="col s12 m3 push-m9">
      <div class="center-align">
        <p class="hide-on-small-only">&nbsp;</p>
        <p class="hide-on-small-only">&nbsp;</p>
        <i class="large nopaque-icons service-icons service-color-text text-darken" data-service="tesseract-ocr-pipeline"></i>
      </div>
    </div>

    <div class="col s12 m9 pull-m3">
      <div class="card service-color-border border-darken" data-service="tesseract-ocr-pipeline" style="border-top: 10px solid;">
        <div class="card-content">
          <div class="row">
            <div class="col s12">
              <div class="card-panel z-depth-0">
                <span class="card-title"><i class="left material-icons">layers</i>OCR</span>
                <p>In this process, nopaque converts your image data – like photos or scans – into text data. This step enables you to proceed with the computational analysis of your documents.</p>
              </div>
            </div>
          </div>
        </div>
      </div>
    </div>

    <div class="col s12">
      <h2>Submit a job</h2>
      <div class="card">
        <form class="create-job-form" enctype="multipart/form-data" method="POST">
          <div class="card-content">
            {{ form.hidden_tag() }}
            <div class="row">
              <div class="col s12 l4">
                {{ wtf.render_field(form.title, material_icon='title') }}
              </div>
              <div class="col s12 l8">
                {{ wtf.render_field(form.description, material_icon='description') }}
              </div>
              <div class="col s12 l5">
                {{ wtf.render_field(form.pdf, accept='application/pdf', placeholder='Choose a PDF file') }}
              </div>
              <div class="col s12 l4">
                <div class="input-field">
                  <i class="material-icons prefix">language</i>
                  {{ form.model() }}
                  {{ form.model.label }}
                  <span class="helper-text">
                    <a class="modal-trigger tooltipped" href="#models-modal" data-position="bottom" data-tooltip="See more information about models"><i class="material-icons">help_outline</i></a>
                    <a class="tooltipped" href="{{ url_for('contributions.create_tesseract_ocr_pipeline_model') }}" data-position="bottom" data-tooltip="Add your own Tesseract OCR models"><i class="material-icons">new_label</i></a>
                  </span>
                  {% for error in form.model.errors %}
                  <span class="helper-text error-color-text">{{ error }}</span>
                  {% endfor %}
                </div>
              </div>
              <div class="col s12 l3">
                {{ wtf.render_field(form.version, material_icon='apps') }}
              </div>
              <div class="col s12">
                <span class="card-title">Preprocessing</span>
              </div>
              {% if 'disabled' not in form.binarization.render_kw or not form.binarization.render_kw['disabled'] %}
              <div class="col s9">
                <p>{{ form.binarization.label.text }}</p>
                <p class="light">Based on a brightness threshold pixels are converted into either black or white. It is useful to reduce noise in images. (<b>longer duration</b>)</p>
              </div>
              <div class="col s3 right-align">
                <div class="switch">
                  <label>
                    {{ form.binarization() }}
                    <span class="lever"></span>
                  </label>
                </div>
              </div>
              {% endif %}
              {% if 'disabled' not in form.ocropus_nlbin_threshold.render_kw or not form.ocropus_nlbin_threshold.render_kw['disabled'] %}
              <div class="col s9 hide" id="create-job-form-ocropus_nlbin_threshold-container">
                <br>
                <p>Intensity (between 0 and 1)</p>
                <p class="range-field">{{ form.ocropus_nlbin_threshold() }}</p>
              </div>
              {% endif %}
              <!--
              Seperate each setting with the following
              <div class="col s12"><p>&nbsp;</p></div>
              <div class="col s12 divider"></div>
              <div class="col s12"><p>&nbsp;</p></div>
              -->
            </div>
          </div>
          <div class="card-action right-align">
            {{ wtf.render_field(form.submit, material_icon='send') }}
          </div>
        </form>
      </div>
    </div>
  </div>
</div>
{% endblock page_content %}

{% block modals %}
{{ super() }}
<div id="models-modal" class="modal">
  <div class="modal-content">
    <h4>Tesseract OCR Pipeline models</h4>
    <table>
      <thead>
        <tr>
          <th>Title</th>
          <th>Description</th>
          <th>Biblio</th>
        </tr>
      </thead>
      <tbody>
        {% for m in tesseract_ocr_pipeline_models %}
        <tr id="tesseract-ocr-pipeline-model-{{ m.hashid }}">
          <td>{{ m.title }}</td>
          {% if m.description == '' %}
          <td>Description is not available.</td>
          {% else %}
          <td>{{ m.description }}</td>
          {% endif %}
          <td><a href="{{ m.publisher_url }}">{{ m.publisher }}</a> ({{ m.publishing_year }}), {{ m.title }} {{ m.version}}, <a href="{{ m.publishing_url }}">{{ m.publishing_url }}</a></td>
        </tr>
        {% endfor %}
      </tbody>
    </table>
  </div>
  <div class="modal-footer">
    <a href="#!" class="modal-close waves-effect waves-light btn">Close</a>
  </div>
</div>
{% endblock modals %}

{% block scripts %}
{{ super() }}
<script>

// Disable user model selection if no models are available
{% if user_tesseract_ocr_pipeline_models_count == 0 %}
optionGroupOptions = document.querySelectorAll(".optgroup-option");
optionGroupOptions[0].classList.add("disabled");
{% endif %}

document.querySelector('#create-job-form-binarization').addEventListener('change', (event) => {
  document.querySelector('#create-job-form-ocropus_nlbin_threshold-container').classList.toggle('hide');
});
</script>
{% endblock scripts %}