First commit

.gitignore

@@ -0,0 +1,16 @@
+venv/
+
+*.pyc
+__pycache__/
+
+instance/
+
+.pytest_cache/
+.coverage
+htmlcov/
+
+dist/
+build/
+*.egg-info/
+
+*~

cce_search/__init__.py

@@ -0,0 +1,44 @@
+import os
+from flask import Flask, send_from_directory
+
+
+def create_app(test_config=None):
+    # create and configure the app
+    app = Flask(__name__, instance_relative_config=True)
+    app.config.from_mapping(
+        SECRET_KEY='dev',
+        #DATABASE=os.path.join(app.instance_path, 'flaskr.sqlite'),
+        API='http://sfr-bardo-copyright-development.us-east-1.elasticbeanstalk.com',
+    )
+
+    if test_config is None:
+        # load the instance config, if it exists, when not testing
+        app.config.from_pyfile('config.py', silent=True)
+    else:
+        # load the test config if passed in
+        app.config.from_mapping(test_config)
+
+    # ensure the instance folder exists
+    try:
+        os.makedirs(app.instance_path)
+    except OSError:
+        pass
+
+    from . import api
+
+    from . import search
+    app.register_blueprint(search.bp)
+    app.add_url_rule('/', endpoint='index')
+    
+    # a simple page that says hello
+    @app.route('/hello')
+    def hello():
+        return api.search()
+
+    @app.route('/fonts/<path:path>')
+    def send_fonts(path):
+        return send_from_directory('fonts', path)
+    
+    return app
+
+

cce_search/api.py

@@ -0,0 +1,43 @@
+from flask import current_app, g
+import requests
+
+def search(term, page=0, per_page=10):
+    r = requests.get(current_app.config['API'] + '/search/fulltext',
+                     params={'query': term,
+                             'source': 'true',
+                             'page': page,
+                             'per_page': per_page})
+    return r.json()
+
+
+def reg_search(term, page=0, per_page=10):
+    api = current_app.config['API']
+    r = requests.get("{}/search/registration/{}".format(api, term),
+                     params={'source': 'true',
+                             'page': page,
+                             'per_page': per_page})
+    return r.json()
+
+
+def ren_search(term, page=0, per_page=10):
+    api = current_app.config['API']
+    r = requests.get("{}/search/renewal/{}".format(api, term),
+                     params={'source': 'true',
+                             'page': page,
+                             'per_page': per_page})
+    return r.json()
+    
+
+
+def registration(cceid):
+    r = requests.get('{}/registration/{}'.format(current_app.config['API'],
+                                                 cceid))
+    r.raise_for_status()
+    return r.json()
+
+
+def renewal(cceid):
+    r = requests.get('{}/renewal/{}'.format(current_app.config['API'],
+                                            cceid))
+    r.raise_for_status()
+    return r.json()

cce_search/search.py

@@ -0,0 +1,124 @@
+from cce_search.api import search, reg_search, ren_search, registration, renewal
+from flask import (
+    Blueprint, flash, g, redirect, render_template, request, url_for
+)
+import re
+from urllib.parse import urlparse, parse_qs, parse_qsl, urlunparse, urlencode
+from werkzeug.exceptions import abort
+from requests import HTTPError
+
+bp = Blueprint('search', __name__)
+
+@bp.route('/')
+def index():
+    results = None
+    term = None
+    paging=None
+    search_type = "ft"
+    if request.args:
+        term = request.args['term']
+        if request.args['type'] == 'ft':
+            response = search(request.args['term'], request.args.get('page'),
+                              request.args.get('per_page'))
+        elif request.args['type'] == 'reg':
+            search_type = "reg"
+            response = reg_search(request.args['term'],
+                                  request.args.get('page'),
+                                  request.args.get('per_page'))
+        else:
+            search_type = "ren"
+            response = ren_search(request.args['term'],
+                                  request.args.get('page'),
+                                  request.args.get('per_page'))
+            
+        paging = proc_pagination(response['data']['paging'],
+                                 request.args.get('page'))
+        results = proc_results(response)
+    
+    
+    return render_template('search/index.html', results=results, term=term,
+                           paging=paging, search_type=search_type)
+
+
+def proc_results(r):
+    return [enhance_results(res) for res in r['data']['results']]
+
+
+def enhance_results(r):
+    if r.get('type') == 'renewal':
+        return r
+
+    return {**r, **{'original': strip_tags(r.get('xml')),
+                    'source_url': ia_url(r.get('source', {}))}}
+
+
+def strip_tags(xml):
+    if xml:
+        return re.sub(r"</?.+?>", "", xml).replace("\n", "")
+    return ""
+
+
+def ia_url(src):
+    #return src
+    return "{}#page/{:d}/mode/1up".format(src.get('url', ''),
+                                        src.get('page', 0))
+
+def proc_pagination(pg, current):
+    if not pg['next'] and not pg['previous']:
+        return {**pg, **{'has_pages': False}}
+
+    per_page = extract_per_page(pg)
+
+    if current is None:
+        current = 1
+    else:
+        current = int(current) + 1
+        
+    
+    return {**pg, **{'has_pages': True,
+                     'current_page': current,
+                     'last_page': extract_last(pg),
+                     'pages': dict([(p, extract_pg(pg.get(p), per_page))
+                               for p in ['first', 'next', 'last',
+                                         'previous']])}}
+
+def extract_pg(pg, per_page):
+    if pg:
+        oq = dict(parse_qsl(urlparse(pg).query))
+        t = urlparse(request.url)
+        return urlunparse(
+            t._replace(query=urlencode({**dict(parse_qsl(t.query)),
+                                        **{'page': oq['page'],
+                                           'per_page': oq['per_page']}})))
+
+    return None
+
+
+def extract_per_page(pg):
+    return [parse_qs(urlparse(v).query)["per_page"][0]
+            for v in pg.values() if v][0]
+
+
+def extract_last(pg):
+    last = pg.get('last')
+    if last is None:
+        return "1"
+    else:
+        return int(parse_qs(urlparse(last).query)["page"][0]) + 1
+
+
+@bp.route('/cceid/<cceid>')
+def cceid(cceid):
+    try:
+        results = registration(cceid)
+        return render_template('search/cceid.html', results=results)
+    except HTTPError:
+        try:
+            results = renewal(cceid)
+            return render_template('search/cceid.html', results=results)
+        except HTTPError:
+            pass
+
+    return render_template('search/cceid.html', results=None, error=1)
+    
+        

cce_search/static/styles.css

@@ -0,0 +1,31 @@
+.info-card {
+    border: 2px solid black;
+    border-radius: 1em;
+    margin-bottom: 1em;
+    padding-bottom: 1em;
+}
+
+.info-card dt {
+    font-weight: bold;
+}
+
+.info-card dd {
+    padding-bottom: 0.5em;
+}
+
+.info-card ul.child-renewal {
+    padding-left: 0;
+}
+
+fieldset.usa-fieldset legend {
+    margin-bottom: 8px;
+}
+
+.pagination {
+    text-align: center;
+}
+
+nav.pagination li {
+    display: inline-block;
+    padding: 0 1em;
+}

cce_search/templates/base.html

@@ -0,0 +1,46 @@
+<!doctype html>
+<html lang="en">
+  <meta charset="utf-8">
+  <title>{% block title %}{% endblock %} - CCE Search</title>
+  <link rel="stylesheet" href="{{ url_for('static', filename='uswds.css') }}">
+  <link rel="stylesheet" href="{{ url_for('static', filename='styles.css') }}">
+</head>
+<body>
+  <div class="usa-banner">
+    <header class="usa-banner__header">
+      <nav>
+        <h1>CCE Search</h1>
+        <ul>
+          <li><a href="{{ url_for('search.index') }}">Search</a>
+        </ul>
+      </nav>
+    </header>
+  </div>
+
+  <main id="main-content" class="grid-container">
+    {% block header %}{% endblock %}
+    {% for message in get_flashed_messages() %}
+    <div class="flash">{{ message }}</div>
+    {% endfor %}
+    {% block content %}{% endblock %}
+  </main>
+
+  <footer class="usa-footer usa-footer--slim" role="contentinfo">
+    <div class="grid-container usa-footer__return-to-top">
+      <a href="#">Return to top</a>
+    </div>
+    <div class="usa-footer__primary-section">
+      <div class="grid-row grid-row maxw-desktop margin-x-auto desktop:padding-x-4">
+        <div class="mobile-lg:grid-col-8">
+          <nav class="usa-footer__nav">
+            <ul class="add-list-reset grid-row grid-gap">
+              <li class="mobile-lg:grid-col-6 desktop:grid-col-auto usa-footer__primary-content">
+                Footer stuff
+              </li>
+            </ul>
+          </nav>
+        </div>
+      </div>
+    </div>
+  </footer>
+</body>

cce_search/templates/search/#render_item.html#

@@ -0,0 +1,58 @@
+{% if result['registrations'] %}
+<dl>
+  <dt>Title</dt>
+  <dd>{{ result['title'] }}</dd>
+
+  <dt>Registrations</dt>
+  <dd>
+    <ul>
+      {% for reg in result['registrations'] %}
+      <li>
+        <dl>
+          <dt>ID</dt>
+          <dd>{{ reg['number'] }}</dd>
+
+          <dt>Date</dt>
+          <dd>{{ reg['date'] }}</dd>
+        </dl>
+      </li>
+      {% endfor %}
+    </ul>
+  </dd>
+is  
+  {% if result['renewals'] %}
+  <dt>Renewals</dt>
+  <dd>
+    <ul>
+      {% for ren in result['renewals'] %}
+      <li>
+        <dl>
+          <dt>ID</dt>
+          <dd>{{ ren['renewal_num'] }}</dd>
+
+          <dt>Date</dt>
+          <dd>{{ ren['renewal_date'] }}</dd>
+
+          <dt>JSON</dt>
+          <dd>{{ ren }}</dd>
+        </dl>
+      </li>
+      {% endfor %}
+    </ul>
+  </dd>
+  {% endif %}
+
+  <dt>Source</dt>
+  <dd><a href="{{ result['source']['url'] }}#page/{{ result['source']['page'] }}">src</a></dd>
+
+
+  <dt>JSON</dt>
+  <dd>{{ result }}</dd>
+</dl>
+{% else %}
+RENEWAL
+<dl>
+  <dt>JSON</dt>
+  <dd>{{ result }}</dd>
+</dl>
+{% endif %}

cce_search/templates/search/.#render_item.html

@@ -0,0 +1 @@
+sean@lugdunum.local.13184

cce_search/templates/search/cceid.html

@@ -0,0 +1,10 @@
+{% extends 'base.html' %}
+
+{% block header %}
+<h2>{% block title %}Find Copyright Entries{% endblock %}</h2>
+{% endblock %}
+
+{% block content %}
+{{ results }}
+{% endblock %}
+  

cce_search/templates/search/index.html

@@ -0,0 +1,74 @@
+{% extends 'base.html' %}
+
+{% block header %}
+  <h2>{% block title %}Find Copyright Entries{% endblock %}</h2>
+{% endblock %}
+
+  {% block content %}
+  <div class="grid-row">
+    <div class="grid-col-8">
+      <h3>Search</h3>
+      <form method="get" class="usa-form">
+        <fieldset class="usa-fieldset">
+          <legend>Search Type</legend>
+          <div class="usa-radio">
+            <input class="usa-radio__input" id="type_ft" type="radio"
+                   {% if search_type == "ft" %}checked{% endif %} name="type"
+                   value="ft">
+            <label class="usa-radio__label" for="type_ft">Full Text</label>
+          </div>
+          <div class="usa-radio">
+            <input class="usa-radio__input" id="type_registration" type="radio"
+                   {% if search_type == "reg" %}checked{% endif %} name="type"
+                   value="reg">
+            <label class="usa-radio__label" for="type_registration">Registration Number</label>
+          </div>
+          <div class="usa-radio">
+            <input class="usa-radio__input" id="type_renewal" type="radio"
+                   {% if search_type == "ren" %}checked{% endif %} name="type"
+                   value="ren">
+            <label class="usa-radio__label" for="type_renewal">Renewal Number</label>
+          </div>
+        </fieldset>
+
+
+        <label for="term" class="usa-label">Search For</label>
+        <input name="term" id="term" required class="usa-input"
+               {% if term %} value="{{ term }}" {% endif %} >
+        
+        <input type="submit" value="Search" class="usa-button">
+      </form>
+    </div>
+    
+    <div class="grid-col-4">
+      <h3>Search Hints</h3>
+
+      <ul>
+        <li>Use “Full Text” for authors, titles, and publishers</li>
+        <li>Use quotation marks to search phrases (for example &quot;Grapes of Wrath&quot;)</li>
+        <li>Searches are case insensitive</li>
+        <li>Renewal numbers look like <tt>R12345</tt> or <tt>RE12345</tt></li>
+        <li>Registration numbers look like <tt>A12345</tt>, <tt>BB12345</tt>, <tt>D12345</tt>. That is, they start with many letters or combination of letters, but not “R” or “RE”</li>
+      </ul>
+    </div>
+  </div>
+
+  {% if results %}
+  <h2>Results</h2>
+
+  {% include "search/paging.html" %}
+
+  <ul>
+    {% for result in results %}
+      {% if result['type'] %}
+        {% include "search/render_renewal.html" %}
+      {% else %}
+        {% include "search/render_registration.html" %}
+      {% endif %}
+    {% endfor %}
+  </ul>
+
+  {% endif %}
+
+  {% include "search/paging.html" %}
+{% endblock %}

cce_search/templates/search/paging.html

@@ -0,0 +1,27 @@
+{% if paging["has_pages"] %}
+<nav aria-label="pagination" class="pagination">
+  <ul class="usa-list usa-list--unstyled">
+    {% if paging['pages']['first'] %}
+    <li><a href="{{ paging['pages']['first'] }}">First<a></li>
+    {% endif %}
+
+    {% if paging['pages']['previous'] %}
+    <li><a href="{{ paging['pages']['previous'] }}">Previous<a></li>
+    {% endif %}
+
+    <li>Page {{ paging["current_page"] }} of {{ paging["last_page"] }}</li>
+
+    {% if paging['pages']['next'] %}
+    <li><a href="{{ paging['pages']['next'] }}">Next<a></li>
+    {% endif %}
+
+    {% if paging['pages']['last'] %}
+    <li><a href="{{ paging['pages']['last'] }}">Last<a></li>
+    {% endif %}
+
+  </ul>
+</nav>
+{% else %}
+<div class="pagination">Page 1 of 1</div>
+{% endif %}
+

cce_search/templates/search/render_child_renewal.html

@@ -0,0 +1,17 @@
+<div class="grid-col-4">
+  <a href="/cceid/{{ renewal['uuid'] }}">
+    {{ renewal['renewal_num'] }} {{ renewal['renewal_date'] }}
+  </a>
+</div>
+
+<div class="grid-col-6">
+  <div>
+    {{ renewal['author'] }} {{ renewal['title'] }}
+  </div>
+
+  <ul class="usa-list usa-list--unstyled">
+    {% for cl in renewal['claimants'] %}
+    <li>{{ cl['name'] }} ({{ cl['type'] }})</li>
+    {% endfor %}
+  </ul>
+</div>

cce_search/templates/search/render_item.html

@@ -0,0 +1,58 @@
+{% if result['registrations'] %}
+<dl>
+  <dt>Title</dt>
+  <dd>{{ result['title'] }}</dd>
+
+  <dt>Registrations</dt>
+  <dd>
+    <ul>
+      {% for reg in result['registrations'] %}
+      <li>
+        <dl>
+          <dt>ID</dt>
+          <dd>{{ reg['number'] }}</dd>
+
+          <dt>Date</dt>
+          <dd>{{ reg['date'] }}</dd>
+        </dl>
+      </li>
+      {% endfor %}
+    </ul>
+  </dd>
+  
+  {% if result['renewals'] %}
+  <dt>Renewals</dt>
+  <dd>
+    <ul>
+      {% for ren in result['renewals'] %}
+      <li>
+        <dl>
+          <dt>ID</dt>
+          <dd>{{ ren['renewal_num'] }}</dd>
+
+          <dt>Date</dt>
+          <dd>{{ ren['renewal_date'] }}</dd>
+
+          <dt>JSON</dt>
+          <dd>{{ ren }}</dd>
+        </dl>
+      </li>
+      {% endfor %}
+    </ul>
+  </dd>
+  {% endif %}
+
+  <dt>Source</dt>
+  <dd><a href="{{ result['source']['url'] }}#page/{{ result['source']['page'] }}">src</a></dd>
+
+
+  <dt>JSON</dt>
+  <dd>{{ result }}</dd>
+</dl>
+{% else %}
+RENEWAL
+<dl>
+  <dt>JSON</dt>
+  <dd>{{ result }}</dd>
+</dl>
+{% endif %}

cce_search/templates/search/render_registration.html

@@ -0,0 +1,94 @@
+<div class="info-card registration grid-container">
+  <h3>Registration {{ result['registrations'][0]['number'] }}</h3>
+
+  <div>
+    <dl class="grid-row">
+      <dt class="grid-col-2">Title</dt>
+      <dd class="grid-col-9">{{ result['title'] }}</dd>
+      
+      <dt class="grid-col-2">Authors</dt>
+      <dd class="grid-col-9">
+        <ul class="usa-list usa-list--unstyled">
+          {% for a in result['authors'] %}
+          <li>{{ a }}</li>
+          {% endfor %}
+        </ul>
+      </dd>
+
+      <dt class="grid-col-2">Registrations</dt>
+      <dd class="grid-col-9">
+        <ul class="usa-list usa-list--unstyled">
+          {% for r in result['registrations'] %}
+          <li>{{ r['number'] }} {{ r['date'] }}</li>
+          {% endfor %}
+        </ul>
+      </dd>
+
+      {% if result['copies'] %}
+      <dt class="grid-col-2">Copies</dt>
+      <dd class="grid-col-9">{{ result['copies'] }}</dd>
+      {% endif %}
+      
+      {% if result['copy_date'] %}
+      <dt class="grid-col-2">Copy Date</dt>
+      <dd class="grid-col-9">{{ result['copy_date'] }}</dd>
+      {% endif %}
+      
+      {% if result['description'] %}
+      <dt class="grid-col-2">Description</dt>
+      <dd class="grid-col-9">{{ result['description'] }}</dd>      
+      {% endif %}
+      
+      {% if result['pub_date'] %}
+      <dt class="grid-col-2">Publication Date</dt>
+      <dd class="grid-col-9">{{ result['pub_date'] }}</dd>
+      {% endif %}
+      
+      {% if result['publishers'] %}
+      <dt class="grid-col-2">Publishers</dt>
+      <dd class="grid-col-9">
+        <ul class="usa-list usa-list--unstyled">
+          {% for p in result['publishers'] %}
+          <li>{{ p }}</li>
+          {% endfor %}
+        </ul>
+      </dd>
+      {% endif %}
+      
+      <dt class="grid-col-2">Renewals</dt>
+      <dd class="grid-col-9">
+        {% if result['renewals'] %}
+        <ul class="child-renewal grid-container usa-list usa-list--unstyled">
+          {% for renewal in result['renewals'] %}
+          <li class="grid-row">
+            {% include "search/render_child_renewal.html" %}
+          </li>
+          {% endfor %}
+        </ul>
+        {% else %}
+        <em>No renewals found</em>
+        {% endif %}
+        
+      </dd>
+      
+      <dt class="grid-col-2">Source</dt>
+      <dd class="grid-col-9">
+        <a href="{{ result['source_url'] }}">
+          {{ result['source']['year'] }} p. {{ result['source']['page'] }}
+        </a>
+      </dd>
+      
+      <dt class="grid-col-2">Original</dt>
+      <dd class="grid-col-9">{{ result['original'] }}</dd>
+      
+      <dt class="grid-col-2">CCEID</dt>
+      <dd class="grid-col-9">
+        <a href="/cceid/{{ result['uuid'] }}">{{ result['uuid'] }}</a>
+      </dd>
+
+      
+    </dl>
+
+  </div>
+
+</div>

cce_search/templates/search/render_renewal.html

@@ -0,0 +1,48 @@
+<div class="info-card renewal grid-container">
+  <h3>Renewal {{ result['renewal_num'] }}</h3>
+
+  <div>
+    <dl class="grid-row">
+      <dt class="grid-col-2">Title</dt>
+      <dd class="grid-col-9">{{ result['title'] }}</dd>
+      
+      <dt class="grid-col-2">Authors</dt>
+      <dd class="grid-col-9">{{ result['author'] }}</dd>
+
+      <dt class="grid-col-2">Renewal</dt>
+      <dd class="grid-col-9">
+        {{ result['renewal_num'] }} {{ result['renewal_date'] }}
+      </dd>
+
+      <dt class="grid-col-2">Claimants</dt>
+      <dd class="grid-col-9">
+        <ul class="usa-list usa-list--unstyled">
+          {% for c in result['claimants'] %}
+          <li>{{ c['name'] }} ({{ c['type'] }})</li>
+          {% endfor %}
+        </ul>
+      </dd>
+
+      {% if result['new_matter'] %}
+      <dt class="grid-col-2">New Matter</dt>
+      <dd class="grid-col-9">{{ result['new_matter'] }}</dd>
+      {% endif %}
+
+      {% if result['notes'] %}
+      <dt class="grid-col-2">Notes</dt>
+      <dd class="grid-col-9">{{ result['notes'] }}</dd>
+      {% endif %}
+
+      {% if result['source'] %}
+      <dt class="grid-col-2">Original</dt>
+      <dd class="grid-col-9">{{ result['source'] }}</dd>
+      {% endif %}
+      
+      <dt class="grid-col-2">CCEID</dt>
+      <dd class="grid-col-9">
+        <a href="/cceid/{{ result['uuid'] }}">{{ result['uuid'] }}</a>
+      </dd>
+    </dl>
+  </div>
+</div>
+

requirements.txt

@@ -0,0 +1,6 @@
+Click==7.0
+Flask==1.1.1
+itsdangerous==1.1.0
+Jinja2==2.10.1
+MarkupSafe==1.1.1
+Werkzeug==0.15.4