add support for sorting by contact info

4 files changed, 56 insertions, 2 deletions

diff --git a/tor-metrics/generate.py b/tor-metrics/generate.py
index e0cb0c6..ae4a748 100755
--- a/tor-metrics/generate.py
+++ b/tor-metrics/generate.py
@@ -33,6 +33,7 @@ if __name__ == '__main__':
     RELAY_SET.write_pages_by_key('as')
     RELAY_SET.write_pages_by_key('country')
     RELAY_SET.write_pages_by_key('platform')
+    RELAY_SET.write_pages_by_key('contact')
     RELAY_SET.write_pages_by_flag()
     RELAY_SET.write_relay_info()
 
diff --git a/tor-metrics/relays.py b/tor-metrics/relays.py
index a64f097..2b79af5 100644
--- a/tor-metrics/relays.py
+++ b/tor-metrics/relays.py
@@ -5,6 +5,7 @@ Relays class object consisting of relays (list of dict) and onionoo fetch
 timestamp
 '''
 
+import hashlib
 import json
 import os
 import time
@@ -18,6 +19,27 @@ ABS_PATH = os.path.dirname(os.path.abspath(__file__))
 ENV = Environment(loader=FileSystemLoader(os.path.join(ABS_PATH, 'templates')),
                   trim_blocks=True, lstrip_blocks=True)
 
+def hash_filter(value, hash_type='md5'):
+    '''
+    Custom hash filter for jinja; defaults to "md5" if no type specified
+
+    :param value: value to be hashed
+    :param hash_type: valid hash type
+    :return: computed hash as a hexadecimal string
+    '''
+    hash_func = getattr(hashlib, hash_type, None)
+
+    if hash_func:
+        computed_hash = hash_func(value.encode("utf-8")).hexdigest()
+    else:
+        raise AttributeError(
+            "No hashing function named {hname}".format(hname=hash_type)
+        )
+
+    return computed_hash
+
+ENV.filters['hash'] = hash_filter
+
 class Relays:
     '''
     Relay class consisting of relays (list of dict) and onionoo fetch timestamp
@@ -174,6 +196,26 @@ class Relays:
                 else:
                     self.json['sorted']['family'][member]['middle_count'] += 1
 
+            c_str = relay.get('contact', '').encode('utf-8')
+            c_hash = hashlib.md5(c_str).hexdigest()
+            if 'contact' not in self.json['sorted']:
+                self.json['sorted']['contact'] = dict()
+            if not c_hash in self.json['sorted']['contact']:
+                self.json['sorted']['contact'][c_hash] = {
+                    'relays':       list(),
+                    'contact':      c_str,
+                    'bw':           0,
+                    'exit_count':   0,
+                    'middle_count': 0
+                }
+            bw = relay['observed_bandwidth']
+            self.json['sorted']['contact'][c_hash]['relays'].append(idx)
+            self.json['sorted']['contact'][c_hash]['bw'] += bw
+            if 'Exit' in relay['flags']:
+                self.json['sorted']['contact'][c_hash]['exit_count'] += 1
+            else:
+                self.json['sorted']['contact'][c_hash]['middle_count'] += 1
+
     def create_output_dir(self):
         '''
         Ensure config:output_root exists (required for write functions)
diff --git a/tor-metrics/templates/contact.html b/tor-metrics/templates/contact.html
new file mode 100644
index 0000000..10bfd93
--- /dev/null
+++ b/tor-metrics/templates/contact.html
@@ -0,0 +1,11 @@
+{% extends "relay-list.html" %}
+{% if relays.json['relay_subset'][0]['contact'] %}
+    {% set contact = relays.json['relay_subset'][0]['contact']|escape %}
+    {% set contact_hash = relays.json['relay_subset'][0]['contact']|hash %}
+{% else %}
+    {% set contact = 'none' %}
+    {% set contact_hash = ''|hash %}
+{% endif %}
+{% block title %}Tor Relays :: Contact {{ contact_hash }}{% endblock %}
+{% block header %}<a href="../../">Home</a> :: Contact {{ contact_hash }}{% endblock %}
+{% block description %}Relays with contact info <code>{{ contact }}</code> are responsible for ~{{ bandwidth }} MB/s of traffic, with {% if middle_count > 0 %}{{ middle_count }} middle relay{% if middle_count > 1 %}s{% endif %}{% if exit_count > 0 %} and {% endif %}{% endif %}{% if exit_count > 0 %}{{ exit_count }} exit relay{% if exit_count > 1 %}s{% endif %}{% endif %}.{% endblock %}
diff --git a/tor-metrics/templates/relay-info.html b/tor-metrics/templates/relay-info.html
index 2ed50d3..9b8eed1 100644
--- a/tor-metrics/templates/relay-info.html
+++ b/tor-metrics/templates/relay-info.html
@@ -17,9 +17,9 @@
 
             <dt>Contact</dt>
             {% if relay['contact'] %}
-                <dd>{{ relay['contact']|escape }}</dd>
+                <dd><a href="{{ path_prefix }}contact/{{ relay['contact']|hash }}">{{ relay['contact']|escape }}</a></dd>
             {% else %}
-                <dd>none</dd>
+                <dd><a href="{{ path_prefix }}contact/{{ ''|hash }}">none</a></dd>
             {% endif %}
 
             <dt>Dir Address</dt>