Add similar people to individual persons' pages

pokr/controllers/person.py

@@ -1,11 +1,15 @@
 # -*- coding: utf-8 -*-
+import math
+import numpy
 
 from collections import Counter, defaultdict, OrderedDict
 
 from .base import Controller
+from pokr.cache import cache
 from pokr.database import db_session
 from pokr.models import Bill, Candidacy, cosponsorship, Meeting, Party, Person, Pledge, Statement
-
+from sqlalchemy.orm.exc import NoResultFound
+from sqlalchemy.sql.expression import select
 
 class PersonController(Controller):
     model = 'person'
@@ -100,3 +104,86 @@ def sorted_statements(cls, person):
                   .filter(Statement.person_id==person.id)\
                   .order_by(Meeting.date.desc().nullslast(), Statement.sequence)
 
+    @classmethod
+    def get_similar_assembly_members(cls, person, assembly_id):
+        ideology_tuples = calculate_ideology_tuple(assembly_id)
+        # find the person's ideology level
+        self_ideology_tuple = None
+        for ideology_tuple in ideology_tuples:
+            if ideology_tuple[0] == person.id:
+                self_ideology_tuple = ideology_tuple
+                break
+
+        #remove self from the tuples
+        ideology_tuples.remove(self_ideology_tuple)
+        self_ideology = self_ideology_tuple[1]
+
+        # find the most similar 5 persons
+        diff_ideology_tuples = map(lambda tuple: (tuple[0], math.fabs(tuple[1] - self_ideology)), ideology_tuples)
+        diff_ideology_tuples.sort(cmp=lambda x,y: cmp(x[1], y[1]))
+        similar_people_ids = map(lambda tuple: tuple[0], diff_ideology_tuples[:5])
+
+        return Person.query\
+          .filter(Person.id.in_(similar_people_ids))\
+          .all()
+
+
+
+def rescale(u):
+    u = (u - min(u)) / (max(u) - min(u))
+    return [float(v) for v in u]
+
+# Also draws heavily from govtrack.us
+@cache.memoize(timeout=60*60*24)
+def generate_cosponsorship_matrix(assembly_id):
+    try:
+        bills = Bill.query.filter(Bill.assembly_id==assembly_id).all()
+        rep_to_row = {}
+        cosponsorships = []
+
+        def rownum(id):
+            if not id in rep_to_row:
+                rep_to_row[id] = len(rep_to_row)
+            return rep_to_row[id]
+
+        for bill in bills:
+            reps = bill.representative_people
+            if len(reps) == 0:
+                continue
+
+            for rep in reps:
+                for cosponsor in bill.cosponsors:
+                    rownum(cosponsor.id)
+                    cosponsorships.append((rep.id, cosponsor.id))
+
+        P = numpy.identity(len(rep_to_row), numpy.float)
+        for sponsor, cosponsor in cosponsorships:
+            P[rep_to_row[sponsor], rep_to_row[cosponsor]] += 1.0
+
+    except NoResultFound, e:
+        print e
+
+    return rep_to_row, P
+
+# Got a lot of inspiration from govtrack.us
+@cache.memoize(timeout=60*60*24)
+def calculate_ideology_tuple(assembly_id):
+    try:
+        rep_to_row, P = generate_cosponsorship_matrix(assembly_id)
+        u, s, vh = numpy.linalg.svd(P)
+        spectrum = vh[1,:]
+        spectrum = rescale(spectrum)
+        ids = [None for k in rep_to_row]
+        for k, v in rep_to_row.items():
+            ids[v] = k
+
+        ideology_tuples = []
+        for index, person_id in enumerate(ids):
+            ideology_tuples.append((person_id, spectrum[index]))
+
+        ideology_tuples.sort(cmp=lambda x,y: cmp(x[1], y[1]))
+
+    except NoResultFound, e:
+        print e
+
+    return ideology_tuples

pokr/messages.pot

@@ -8,14 +8,14 @@ msgid ""
 msgstr ""
 "Project-Id-Version: PROJECT VERSION\n"
 "Report-Msgid-Bugs-To: EMAIL@ADDRESS\n"
-"POT-Creation-Date: 2015-04-01 17:11+0900\n"
+"POT-Creation-Date: 2015-04-08 08:14-0700\n"
 "PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
 "Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
 "Language-Team: LANGUAGE <[email protected]>\n"
 "MIME-Version: 1.0\n"
 "Content-Type: text/plain; charset=utf-8\n"
 "Content-Transfer-Encoding: 8bit\n"
-"Generated-By: Babel 0.9.6\n"
+"Generated-By: Babel 1.3\n"
 
 #: pokr/models/meeting.py:49
 #, python-format
@@ -32,7 +32,7 @@ msgstr ""
 msgid "%(sitting_id)sth sitting"
 msgstr ""
 
-#: pokr/templates/bill-layout.html:10 pokr/templates/bill.html:142
+#: pokr/templates/bill-layout.html:10 pokr/templates/bill.html:140
 msgid "The summary of this bill has not been updated."
 msgstr ""
 
@@ -74,50 +74,50 @@ msgstr ""
 msgid "see official page"
 msgstr ""
 
-#: pokr/templates/bill.html:68
+#: pokr/templates/bill.html:67
 msgid "suggest to assembly"
 msgstr ""
 
-#: pokr/templates/bill.html:70
+#: pokr/templates/bill.html:68
 msgid "see original pdf"
 msgstr ""
 
-#: pokr/templates/bill.html:71
+#: pokr/templates/bill.html:69
 msgid "see original text"
 msgstr ""
 
-#: pokr/templates/bill.html:79
+#: pokr/templates/bill.html:77
 msgid "sponsors"
 msgstr ""
 
-#: pokr/templates/bill.html:82 pokr/templates/bill.html:105
+#: pokr/templates/bill.html:80 pokr/templates/bill.html:103
 #: pokr/templates/macros.html:105 pokr/templates/person-legislations.html:64
 msgid "representative sponsor"
 msgstr ""
 
-#: pokr/templates/bill.html:93 pokr/templates/bill.html:119
+#: pokr/templates/bill.html:91 pokr/templates/bill.html:117
 msgid "cosponsors"
 msgstr ""
 
-#: pokr/templates/bill.html:134
+#: pokr/templates/bill.html:132
 msgid "contents"
 msgstr ""
 
-#: pokr/templates/bill.html:137 pokr/templates/party.html:43
+#: pokr/templates/bill.html:135 pokr/templates/party.html:43
 msgid "summary"
 msgstr ""
 
 #: pokr/templates/bills.html:5 pokr/templates/bills.html:9
 msgid "bills"
 msgstr ""
 
-#: pokr/templates/bills.html:29 pokr/templates/parties.html:30
+#: pokr/templates/bills.html:30 pokr/templates/parties.html:30
 #: pokr/templates/people.html:29 pokr/templates/region.html:54
 #: pokr/templates/region.html:99
 msgid "th"
 msgstr ""
 
-#: pokr/templates/bills.html:57 pokr/templates/macro-statement.html:6
+#: pokr/templates/bills.html:62 pokr/templates/macro-statement.html:6
 #: pokr/templates/macros.html:95 pokr/templates/mypage.html:88
 #: pokr/templates/mypage.html:113 pokr/templates/person-legislations.html:29
 #: pokr/templates/person-legislations.html:73
@@ -126,7 +126,7 @@ msgstr ""
 msgid "more"
 msgstr ""
 
-#: pokr/templates/bills.html:58 pokr/templates/district-feeds.html:19
+#: pokr/templates/bills.html:63 pokr/templates/district-feeds.html:19
 msgid "No summary"
 msgstr ""
 
@@ -554,6 +554,7 @@ msgstr ""
 
 #: pokr/templates/person-legislations.html:13
 #: pokr/templates/person-legislations.html:53
+#: pokr/templates/person-similar-people.html:18
 msgid "th assembly"
 msgstr ""
 
@@ -566,6 +567,14 @@ msgstr ""
 msgid "role"
 msgstr ""
 
+#: pokr/templates/person-similar-people.html:4
+msgid "similar people"
+msgstr ""
+
+#: pokr/templates/person-similar-people.html:8
+msgid "based on cosponsorships"
+msgstr ""
+
 #: pokr/templates/person-statements.html:6
 msgid "Statements"
 msgstr ""
@@ -620,11 +629,11 @@ msgstr ""
 msgid "Wikipedia"
 msgstr ""
 
-#: pokr/templates/person.html:161
+#: pokr/templates/person.html:162
 msgid "trends"
 msgstr ""
 
-#: pokr/templates/person.html:163
+#: pokr/templates/person.html:164
 #, python-format
 msgid "Results from Google Trends searched by query %(name)s"
 msgstr ""
@@ -736,7 +745,7 @@ msgid "sources"
 msgstr ""
 
 #: pokr/templates/includes/header.html:15
-#: pokr/templates/includes/header.html:17 pokr/views/bill.py:19
+#: pokr/templates/includes/header.html:17 pokr/views/bill.py:26
 msgid "bill"
 msgstr ""
 

pokr/static/less/person.less

@@ -51,6 +51,24 @@
             }
         }
 
+        #section-similar-people {
+            div.content {
+                div.group {
+                    div {
+                        display: inline-block;
+
+                        &.similarity {
+                            width: 13%;
+                        }
+
+                        &.people {
+
+                        }
+                    }
+                }
+            }
+        }
+
         #section-elections {
             tr:first-child th,
             tr:first-child td {

pokr/templates/person-similar-people.html

@@ -0,0 +1,32 @@
+{% import 'macros.html' as macro with context %}
+
+<div id="section-similar-people" class="section">
+    <h3 id="similar-people">{{ gettext('similar people') }}</h3>
+    <table id="person-legislations" class="table table-condensed">
+        <thead>
+            <tr>
+                <th>{{ gettext('based on cosponsorships') }}</th>
+                <th></th>
+            </tr>
+        </thead>
+        <tbody>
+            {% for candidacy in person.candidacies %}
+                {% if candidacy.is_elected %}
+                    {% set similar_assembly_members = PersonController.get_similar_assembly_members(person, candidacy.assembly_id) %}
+                    <tr>
+                        <td>
+                            <a href="{{ url_for('person_main', assembly_id=candidacy.assembly_id) }}">{{ candidacy.assembly_id }}{{ gettext('th assembly') }}</a>
+                        </td>
+                        <td>
+                            {% for similar_assembly_member in similar_assembly_members %}
+                            <span class="people">
+                                <a href="{{ url_for('person', id=similar_assembly_member.id) }}">{{ similar_assembly_member.name }}</a>
+                            </span>
+                            {% endfor %}
+                        </td>
+                    </tr>
+                {% endif %}
+            {% endfor %}
+        </tbody>
+    </table>
+</div>

pokr/templates/person.html

@@ -152,6 +152,7 @@ <h3 id="profile">
             </div>
             {% endif %}
         </div>
+        {% if has_bills %}{% include 'person-similar-people.html' with context %}{% endif %}
         {% if has_bills %}{% include 'person-legislations.html' with context %}{% endif %}
         {% include 'person-statements.html' with context %}
         {% include 'person-elections.html' with context %}