docker: update to pywb rc7

content type redirects: support default block list, with specific allow rules, per #54

Dockerfile

@@ -1,6 +1,6 @@
 
 # Based on standard pywb fork
-FROM webrecorder/pywb:2.4.0-rc6-test
+FROM webrecorder/pywb:2.4.0-rc7
 
 USER root
 WORKDIR /ukwa_pywb

config.yaml

@@ -16,10 +16,20 @@ collections:
           Expires: 'Thu, 01 Jan 1970 00:00:00 GMT'
 
         content_type_redirects:
+          # allows
+          'text/': 'allow'
+          'image/': 'allow'
+          'video/': 'allow'
+          'audio/': 'allow'
+          'application/javascript': 'allow'
+
           'text/rtf': 'https://example.com/viewer?{query}'
           'application/pdf': 'https://example.com/viewer?{query}'
           'application/': 'https://example.com/blocked?{query}'
+
+          # default redirects
           '<any-download>': 'https://example.com/blocked?{query}'
+          '*': 'https://example.com/blocked?{query}'
 
 
     # open access (with access controls)

ukwa_pywb/test/config_test.yaml

@@ -13,10 +13,18 @@ collections:
           Expires: 'Thu, 01 Jan 1970 00:00:00 GMT'
 
         content_type_redirects:
+          'text/': 'allow'
+          'image/': 'allow'
+          'video/': 'allow'
+          'audio/': 'allow'
+
           'text/rtf': 'https://example.com/viewer?{query}'
           'application/pdf': 'https://example.com/viewer?{query}'
           'application/': 'https://example.com/blocked?{query}'
+
+          # default redirects
           '<any-download>': 'https://example.com/blocked?{query}'
+          '*': 'https://example.com/blocked?{query}'
 
     pywb-no-locks:
         index_paths: ./integration-test/test-data/

ukwa_pywb/ukwa_app.py

@@ -196,25 +196,33 @@ def render_content(self, wb_url_str, coll_config, environ):
         if default_response.status_headers.get('preference-applied') == 'raw':
             return default_response
 
-        content_type = default_response.status_headers.get("content-type")
 
         redirect_url = None
 
-        if content_type:
-            content_type = content_type.split(";", 1)[0]
-            redirect_url = ct_redirects.get(content_type)
-            if redirect_url is None:
-                redirect_url = ct_redirects.get(content_type.split("/")[0] + "/")
-
-        # if no content-type match, check content-disposition
-        if not redirect_url:
-            content_disp = default_response.status_headers.get("content-disposition")
-            if content_disp and 'attachment' in content_disp:
-                redirect_url = ct_redirects.get('<any-download>')
-
-        if not redirect_url:
+        # if we have a content-disposition, takes precedence using the <any-download> option
+        content_disp = default_response.status_headers.get("content-disposition")
+        if content_disp and 'attachment' in content_disp:
+            redirect_url = ct_redirects.get('<any-download>')
+
+        # attempt to find rule by content-type
+        if redirect_url is None:
+            content_type = default_response.status_headers.get("content-type")
+            if content_type:
+                content_type = content_type.split(";", 1)[0]
+                redirect_url = ct_redirects.get(content_type)
+                # find by content-type prefix, eg: text/
+                if redirect_url is None:
+                    redirect_url = ct_redirects.get(content_type.split("/")[0] + "/")
+
+        # default rule if no other matches
+        if redirect_url is None:
+            redirect_url = ct_redirects.get('*')
+
+        # if no redirect or rule is 'allow', then continue
+        if not redirect_url or redirect_url == 'allow':
             return default_response
 
+        # otherwise, redirect to specified url
         wb_url = WbUrl(wb_url_str)
         wb_url.mod = 'id_'
         loc = self.get_full_prefix(environ) + str(wb_url)