Merge pull request #182 from dalf/enh-proxy

[enh] image-proxy : handle ETag and date related headers, add hash to URL

searx/templates/courgette/result_templates/code.html

@@ -1,7 +1,7 @@
 <div class="result {{ result.class }}">
     <h3 class="result_title">{% if result['favicon'] %}<img width="14" height="14" class="favicon" src="static/{{theme}}/img/icon_{{result['favicon']}}.ico" alt="{{result['favicon']}}" />{% endif %}<a href="{{ result.url }}">{{ result.title|safe }}</a></h3>
     {% if result.publishedDate %}<span class="published_date">{{ result.publishedDate }}</span>{% endif %}
-    <p class="content">{% if result.img_src %}<img src="{{ result.img_src }}" class="image" />{% endif %}{% if result.content %}{{ result.content|safe }}<br class="last"/>{% endif %}</p>
+    <p class="content">{% if result.img_src %}<img src="{{ image_proxify(result.img_src) }}" class="image" />{% endif %}{% if result.content %}{{ result.content|safe }}<br class="last"/>{% endif %}</p>
     {% if result.repository %}<p class="content"><a href="{{ result.repository|safe }}">{{ result.repository }}</a></p>{% endif %}
     {{ result.codelines|code_highlighter(result.code_language)|safe }}
 

searx/templates/courgette/result_templates/videos.html

@@ -5,6 +5,6 @@
 
     <h3 class="result_title"><a href="{{ result.url }}">{{ result.title|safe }}</a></h3>
     {% if result.publishedDate %}<span class="published_date">{{ result.publishedDate }}</span><br />{% endif %}
-    <a href="{{ result.url }}"><img width="400" src="{{ result.thumbnail }}" title="{{ result.title|striptags }}" alt="{{ result.title|striptags }}"/></a>
+    <a href="{{ result.url }}"><img width="400" src="{{ image_proxify(result.thumbnail) }}" title="{{ result.title|striptags }}" alt="{{ result.title|striptags }}"/></a>
     <p class="url">{{ result.pretty_url }}</p>
 </div>

searx/templates/default/infobox.html

@@ -1,6 +1,6 @@
 <div class="infobox">
     <h2>{{ infobox.infobox }}</h2>
-    {% if infobox.img_src %}<img src="{{ infobox.img_src }}" title="{{ infobox.infobox|striptags }}" alt="{{ infobox.infobox|striptags }}" />{% endif %}
+    {% if infobox.img_src %}<img src="{{ image_proxify(infobox.img_src) }}" title="{{ infobox.infobox|striptags }}" alt="{{ infobox.infobox|striptags }}" />{% endif %}
     <p>{{ infobox.entity }}</p>
     <p>{{ infobox.content | safe }}</p>
     {% if infobox.attributes %}

searx/templates/default/result_templates/code.html

@@ -2,7 +2,7 @@
     <h3 class="result_title"> {% if result['favicon'] %}<img width="14" height="14" class="favicon" src="static/{{theme}}/img/icon_{{result['favicon']}}.ico" alt="{{result['favicon']}}" />{% endif %}<a href="{{ result.url }}">{{ result.title|safe }}</a></h3>
     <p class="url">{{ result.pretty_url }} <a class="cache_link" href="https://web.archive.org/web/{{ result.url }}">cached</a></p>
     {% if result.publishedDate %}<p class="published_date">{{ result.publishedDate }}</p>{% endif %}
-    <p class="content">{% if result.img_src %}<img src="{{ result.img_src }}" class="image" />{% endif %}{% if result.content %}{{ result.content|safe }}<br class="last"/>{% endif %}</p>
+    <p class="content">{% if result.img_src %}<img src="{{ image_proxify(result.img_src) }}" class="image" />{% endif %}{% if result.content %}{{ result.content|safe }}<br class="last"/>{% endif %}</p>
     {% if result.repository %}<p class="result-content"><a href="{{ result.repository|safe }}">{{ result.repository }}</a></p>{% endif %}
 
     {{ result.codelines|code_highlighter(result.code_language)|safe }}

searx/templates/default/result_templates/default.html

@@ -2,5 +2,5 @@
     <h3 class="result_title">{% if "icon_"~result.engine~".ico" in favicons %}<img width="14" height="14" class="favicon" src="{{ url_for('static', filename='img/icons/icon_'+result.engine+'.ico') }}" alt="{{result.engine}}" />{% endif %}<a href="{{ result.url }}">{{ result.title|safe }}</a></h3>
     <p class="url">{{ result.pretty_url }} <a class="cache_link" href="https://web.archive.org/web/{{ result.url }}">cached</a>
     {% if result.publishedDate %}<span class="published_date">{{ result.publishedDate }}</span>{% endif %}</p>
-    <p class="content">{% if result.img_src %}<img src="{{ result.img_src }}" class="image" />{% endif %}{% if result.content %}{{ result.content|safe }}<br class="last"/>{% endif %}</p>
+    <p class="content">{% if result.img_src %}<img src="{{ image_proxify(result.img_src) }}" class="image" />{% endif %}{% if result.content %}{{ result.content|safe }}<br class="last"/>{% endif %}</p>
 </div>

searx/templates/default/result_templates/map.html

@@ -8,6 +8,6 @@
         <h3 class="result_title"><a href="{{ result.url }}">{{ result.title|safe }}</a></h3>
         <p class="url">{{ result.pretty_url }} <a class="cache_link" href="https://web.archive.org/web/{{ result.url }}">cached</a>
         {% if result.publishedDate %}<span class="published_date">{{ result.publishedDate }}</span>{% endif %}</p>
-        <p class="content">{% if result.img_src %}<img src="{{ result.img_src }}" class="image" />{% endif %}{% if result.content %}{{ result.content|safe }}<br class="last"/>{% endif %}</p>
+        <p class="content">{% if result.img_src %}<img src="{{ image_proxify(result.img_src) }}" class="image" />{% endif %}{% if result.content %}{{ result.content|safe }}<br class="last"/>{% endif %}</p>
     </div>
 </div>

searx/templates/default/result_templates/videos.html

@@ -1,6 +1,6 @@
 <div class="result">
     <h3 class="result_title">{% if "icon_"~result.engine~".ico" in favicons %}<img width="14" height="14" class="favicon" src="{{ url_for('static', filename='img/icons/icon_'+result.engine+'.ico') }}" alt="{{result.engine}}" />{% endif %}<a href="{{ result.url }}">{{ result.title|safe }}</a></h3>
     {% if result.publishedDate %}<span class="published_date">{{ result.publishedDate }}</span><br />{% endif %}
-    <a href="{{ result.url }}"><img class="thumbnail" src="{{ result.thumbnail }}" title="{{ result.title|striptags }}" alt="{{ result.title|striptags }}"/></a>
+    <a href="{{ result.url }}"><img class="thumbnail" src="{{ image_proxify(result.thumbnail) }}" title="{{ result.title|striptags }}" alt="{{ result.title|striptags }}"/></a>
     <p class="url">{{ result.url }}</p>
 </div>

searx/templates/oscar/infobox.html

@@ -3,7 +3,7 @@
         <h4 class="panel-title">{{ infobox.infobox }}</h4>
     </div>
     <div class="panel-body">
-        {% if infobox.img_src %}<img class="img-responsive center-block infobox_part" src="{{ infobox.img_src }}" alt="{{ infobox.infobox }}" />{% endif %}
+        {% if infobox.img_src %}<img class="img-responsive center-block infobox_part" src="{{ image_proxify(infobox.img_src) }}" alt="{{ infobox.infobox }}" />{% endif %}
         {% if infobox.content %}<p class="infobox_part">{{ infobox.content }}</p>{% endif %}
 
         {% if infobox.attributes %}

searx/templates/oscar/result_templates/videos.html

@@ -15,7 +15,7 @@
 
 <div class="container-fluid">
     <div class="row">
-        <a href="{{ result.url }}"><img class="thumbnail col-xs-6 col-sm-4 col-md-4 result-content" src="{{ result.thumbnail|safe }}" alt="{{ result.title|striptags }} {{ result.engine }}" /></a>
+        <a href="{{ result.url }}"><img class="thumbnail col-xs-6 col-sm-4 col-md-4 result-content" src="{{ image_proxify(result.thumbnail) }}" alt="{{ result.title|striptags }} {{ result.engine }}" /></a>
         {% if result.content %}<p class="col-xs-12 col-sm-8 col-md-8 result-content">{{ result.content|safe }}</p>{% endif %}
     </div>
 </div>

searx/utils.py

@@ -206,3 +206,11 @@ def format_date_by_locale(date_string, locale_string):
     except:
         logger.warning('cannot set original locale: {0}'.format(orig_locale))
     return formatted_date
+
+
+def dict_subset(d, properties):
+    result = {}
+    for k in properties:
+        if k in d:
+            result[k] = d[k]
+    return result

searx/webapp.py

@@ -25,6 +25,7 @@ if __name__ == '__main__':
 import json
 import cStringIO
 import os
+import hashlib
 
 from datetime import datetime, timedelta
 from requests import get as http_get
@@ -41,7 +42,7 @@ from searx.engines import (
 )
 from searx.utils import (
     UnicodeWriter, highlight_content, html_to_text, get_themes,
-    get_static_files, get_result_templates, gen_useragent
+    get_static_files, get_result_templates, gen_useragent, dict_subset
 )
 from searx.version import VERSION_STRING
 from searx.languages import language_codes
@@ -216,8 +217,10 @@ def image_proxify(url):
     if not settings['server'].get('image_proxy') and not request.cookies.get('image_proxy'):
         return url
 
+    h = hashlib.sha256(url + settings['server']['secret_key']).hexdigest()
+
     return '{0}?{1}'.format(url_for('image_proxy'),
-                            urlencode(dict(url=url)))
+                            urlencode(dict(url=url, h=h)))
 
 
 def render(template_name, override_theme=None, **kwargs):
@@ -562,10 +565,21 @@ def image_proxy():
     if not url:
         return '', 400
 
+    h = hashlib.sha256(url + settings['server']['secret_key']).hexdigest()
+
+    if h != request.args.get('h'):
+        return '', 400
+
+    headers = dict_subset(request.headers, {'If-Modified-Since', 'If-None-Match'})
+    headers['User-Agent'] = gen_useragent()
+
     resp = http_get(url,
                     stream=True,
                     timeout=settings['server'].get('request_timeout', 2),
-                    headers={'User-Agent': gen_useragent()})
+                    headers=headers)
+
+    if resp.status_code == 304:
+        return '', resp.status_code
 
     if resp.status_code != 200:
         logger.debug('image-proxy: wrong response code: {0}'.format(resp.status_code))
@@ -586,7 +600,9 @@ def image_proxy():
             return '', 502  # Bad gateway - file is too big (>5M)
         img += chunk
 
-    return Response(img, mimetype=resp.headers['content-type'])
+    headers = dict_subset(resp.headers, {'Content-Length', 'Length', 'Date', 'Last-Modified', 'Expires', 'Etag'})
+
+    return Response(img, mimetype=resp.headers['content-type'], headers=headers)
 
 
 @app.route('/stats', methods=['GET'])