Add citation and h-index statistics

 - Legacy-Id: 12869

Author: OleLaursen

ietf/person/models.py

@@ -14,7 +14,7 @@
 
 import debug                            # pyflakes:ignore
 
-from ietf.person.name import name_parts, initials
+from ietf.person.name import name_parts, initials, plain_name
 from ietf.utils.mail import send_mail_preformatted
 from ietf.utils.storage import NoLocationMigrationFileSystemStorage
 
@@ -47,8 +47,7 @@ def short(self):
             return (first and first[0]+"." or "")+(middle or "")+" "+last+(suffix and " "+suffix or "")
     def plain_name(self):
         if not hasattr(self, '_cached_plain_name'):
-            prefix, first, middle, last, suffix = name_parts(self.name)
-            self._cached_plain_name = u" ".join([first, last])
+            self._cached_plain_name = plain_name(self.name)
         return self._cached_plain_name
     def ascii_name(self):
         if not hasattr(self, '_cached_ascii_name'):

ietf/person/name.py

@@ -50,6 +50,10 @@ def initials(name):
     initials = u" ".join([ n[0]+'.' for n in given.split() ])
     return initials
 
+def plain_name(name):
+    prefix, first, middle, last, suffix = name_parts(name)
+    return u" ".join([first, last])
+
 if __name__ == "__main__":
     import sys
     name = u" ".join(sys.argv[1:])

ietf/stats/urls.py

@@ -5,7 +5,7 @@
 
 urlpatterns = patterns('',
     url("^$", ietf.stats.views.stats_index),
-    url("^document/(?:(?P<stats_type>authors|pages|words|format|formlang|author/documents|author/affiliation|author/country|author/continent|author/citation)/)?$", ietf.stats.views.document_stats),
+    url("^document/(?:(?P<stats_type>authors|pages|words|format|formlang|author/documents|author/affiliation|author/country|author/continent|author/citations||author/hindex)/)?$", ietf.stats.views.document_stats),
     url("^knowncountries/$", ietf.stats.views.known_countries_list),
     url("^review/(?:(?P<stats_type>completion|results|states|time)/)?(?:%(acronym)s/)?$" % settings.URL_REGEXPS, ietf.stats.views.review_stats),
 )

ietf/stats/views.py

@@ -23,9 +23,10 @@
 from ietf.submit.models import Submission
 from ietf.group.models import Role, Group
 from ietf.person.models import Person
-from ietf.name.models import ReviewRequestStateName, ReviewResultName, CountryName
+from ietf.name.models import ReviewRequestStateName, ReviewResultName, CountryName, DocRelationshipName
+from ietf.person.name import plain_name
 from ietf.doc.models import DocAlias, Document, State
-from ietf.stats.utils import get_aliased_affiliations, get_aliased_countries
+from ietf.stats.utils import get_aliased_affiliations, get_aliased_countries, compute_hirsch_index
 from ietf.ietfauth.utils import has_role
 
 def stats_index(request):
@@ -103,7 +104,8 @@ def build_document_stats_url(stats_type_override=Ellipsis, get_overrides={}):
         ("author/affiliation", "Affiliation"),
         ("author/country", "Country"),
         ("author/continent", "Continent"),
-        ("author/citation", "Citations"),
+        ("author/citations", "Citations"),
+        ("author/hindex", "Impact"),
     ], lambda slug: build_document_stats_url(stats_type_override=slug))
 
 
@@ -346,7 +348,7 @@ def generate_canonical_names(docalias_qs):
 
             person_filters &= Q(documentauthor__document__in=docs_within_time_constraint)
 
-        person_qs = Person.objects.filter(person_filters)
+        person_qs = Person.objects.filter(person_filters, documentauthor__document="draft-arkko-dual-stack-extra-lite")
 
         if document_type == "rfc":
             doc_label = "RFC"
@@ -369,6 +371,8 @@ def count_bins(bins):
 
             bins = defaultdict(list)
 
+            person_qs = Person.objects.filter(person_filters)
+
             for name, document_count in person_qs.values_list("name").annotate(Count("documentauthor")):
                 bins[document_count].append(name)
 
@@ -378,7 +382,7 @@ def count_bins(bins):
             for document_count, names in sorted(bins.iteritems(), key=lambda t: t[0]):
                 percentage = len(names) * 100.0 / (total_persons or 1)
                 series_data.append((document_count, percentage))
-                table_data.append((document_count, percentage, names))
+                table_data.append((document_count, percentage, [plain_name(n) for n in names]))
 
             chart_data.append({
                 "data": series_data,
@@ -390,6 +394,8 @@ def count_bins(bins):
 
             bins = defaultdict(list)
 
+            person_qs = Person.objects.filter(person_filters)
+
             # Since people don't write the affiliation names in the
             # same way, and we don't want to go back and edit them
             # either, we transform them here.
@@ -410,7 +416,7 @@ def count_bins(bins):
                 percentage = len(names) * 100.0 / (total_persons or 1)
                 if affiliation:
                     series_data.append((affiliation, len(names)))
-                table_data.append((affiliation, percentage, names))
+                table_data.append((affiliation, percentage, [plain_name(n) for n in names]))
 
             series_data.sort(key=lambda t: t[1], reverse=True)
             series_data = series_data[:30]
@@ -428,6 +434,8 @@ def count_bins(bins):
 
             bins = defaultdict(list)
 
+            person_qs = Person.objects.filter(person_filters)
+
             # Since people don't write the country names in the
             # same way, and we don't want to go back and edit them
             # either, we transform them here.
@@ -457,7 +465,7 @@ def count_bins(bins):
                 percentage = len(names) * 100.0 / (total_persons or 1)
                 if country:
                     series_data.append((country, len(names)))
-                table_data.append((country, percentage, names))
+                table_data.append((country, percentage, [plain_name(n) for n in names]))
 
             series_data.sort(key=lambda t: t[1], reverse=True)
             series_data = series_data[:30]
@@ -477,6 +485,8 @@ def count_bins(bins):
 
             bins = defaultdict(list)
 
+            person_qs = Person.objects.filter(person_filters)
+
             name_country_set = set((name, country)
                                    for name, country in person_qs.values_list("name", "documentauthor__country"))
 
@@ -497,7 +507,7 @@ def count_bins(bins):
                 percentage = len(names) * 100.0 / (total_persons or 1)
                 if continent:
                     series_data.append((continent, len(names)))
-                table_data.append((continent, percentage, names))
+                table_data.append((continent, percentage, [plain_name(n) for n in names]))
 
             series_data.sort(key=lambda t: t[1], reverse=True)
 
@@ -506,6 +516,59 @@ def count_bins(bins):
                 "animation": False,
             })
 
+        elif stats_type == "author/citations":
+            stats_title = "Number of citations of {}s written by author".format(doc_label)
+
+            bins = defaultdict(list)
+
+            cite_relationships = list(DocRelationshipName.objects.filter(slug__in=['refnorm', 'refinfo', 'refunk', 'refold']))
+            person_filters &= Q(documentauthor__document__docalias__relateddocument__relationship__in=cite_relationships)
+
+            person_qs = Person.objects.filter(person_filters)
+
+            for name, citations in person_qs.values_list("name").annotate(Count("documentauthor__document__docalias__relateddocument")):
+                bins[citations].append(name)
+
+            total_persons = count_bins(bins)
+
+            series_data = []
+            for citations, names in sorted(bins.iteritems(), key=lambda t: t[0], reverse=True):
+                percentage = len(names) * 100.0 / (total_persons or 1)
+                series_data.append((citations, percentage))
+                table_data.append((citations, percentage, [plain_name(n) for n in names]))
+
+            chart_data.append({
+                "data": sorted(series_data, key=lambda t: t[0]),
+                "animation": False,
+            })
+
+        elif stats_type == "author/hindex":
+            stats_title = "h-index for {}s written by author".format(doc_label)
+
+            bins = defaultdict(list)
+
+            cite_relationships = list(DocRelationshipName.objects.filter(slug__in=['refnorm', 'refinfo', 'refunk', 'refold']))
+            person_filters &= Q(documentauthor__document__docalias__relateddocument__relationship__in=cite_relationships)
+
+            person_qs = Person.objects.filter(person_filters)
+
+            values = person_qs.values_list("name", "documentauthor__document").annotate(Count("documentauthor__document__docalias__relateddocument"))
+            for name, ts in itertools.groupby(values.order_by("name"), key=lambda t: t[0]):
+                h_index = compute_hirsch_index([citations for _, document, citations in ts])
+                bins[h_index].append(name)
+
+            total_persons = count_bins(bins)
+
+            series_data = []
+            for citations, names in sorted(bins.iteritems(), key=lambda t: t[0], reverse=True):
+                percentage = len(names) * 100.0 / (total_persons or 1)
+                series_data.append((citations, percentage))
+                table_data.append((citations, percentage, [plain_name(n) for n in names]))
+
+            chart_data.append({
+                "data": sorted(series_data, key=lambda t: t[0]),
+                "animation": False,
+            })
 
     return render(request, "stats/document_stats.html", {
         "chart_data": mark_safe(json.dumps(chart_data)),

ietf/templates/stats/document_stats_author_citations.html

@@ -0,0 +1,66 @@
+<h3>{{ stats_title }}</h3>
+
+<div id="chart"></div>
+
+<script>
+ var chartConf = {
+     chart: {
+         type: 'area'
+     },
+     title: {
+         text: '{{ stats_title|escapejs }}'
+     },
+     xAxis: {
+         title: {
+             text: 'Number of citations of {{ doc_label }}s by author'
+         },
+         max: 500
+     },
+     yAxis: {
+         title: {
+             text: 'Percentage of authors'
+         },
+         labels: {
+             formatter: function () {
+                 return this.value + '%';
+             }
+         }
+     },
+     tooltip: {
+         formatter: function () {
+             var s = '<b>' + this.x + ' ' + (this.x == 1 ? "citation" : 'citations') + '</b>';
+
+             $.each(this.points, function () {
+                 s += '<br/>' + chartConf.yAxis.title.text + ': ' + this.y.toFixed(1) + '%';
+             });
+
+             return s;
+         },
+         shared: true
+     },
+     series: {{ chart_data }}
+ };
+</script>
+
+<h3>Data</h3>
+
+<table class="table table-condensed stats-data">
+  <thead>
+    <tr>
+      <th>Citations</th>
+      <th>Percentage of authors</th>
+      <th>Authors</th>
+    </tr>
+  </thead>
+  <tbody>
+    {% for citations, percentage, names in table_data %}
+      <tr>
+        <td>{{ citations }}</td>
+        <td>{{ percentage|floatformat:2 }}%</td>
+        <td>{% include "stats/includes/number_with_details_cell.html" with content_limit=10 %}</td>
+      </tr>
+    {% endfor %}
+  </tbody>
+</table>
+
+<p>Note that the citation counts do not exclude self-references.</p>

ietf/templates/stats/document_stats_author_documents.html

@@ -58,7 +58,7 @@ <h3>Data</h3>
       <tr>
         <td>{{ document_count }}</td>
         <td>{{ percentage|floatformat:2 }}%</td>
-        <td>{% include "stats/includes/number_with_details_cell.html" %}</td>
+        <td>{% include "stats/includes/number_with_details_cell.html" with content_limit=10 %}</td>
       </tr>
     {% endfor %}
   </tbody>

ietf/templates/stats/document_stats_author_hindex.html

@@ -0,0 +1,74 @@
+<h3>{{ stats_title }}</h3>
+
+<div id="chart"></div>
+
+<script>
+ var chartConf = {
+     chart: {
+         type: 'column'
+     },
+     title: {
+         text: '{{ stats_title|escapejs }}'
+     },
+     xAxis: {
+         tickInterval: 1,
+         title: {
+             text: 'h-index of {{ doc_label }}s by author'
+         }
+     },
+     yAxis: {
+         title: {
+             text: 'Percentage of authors'
+         },
+         labels: {
+             formatter: function () {
+                 return this.value + '%';
+             }
+         }
+     },
+     tooltip: {
+         formatter: function () {
+             var s = '<b>' + ' h-index ' + this.x + '</b>';
+
+             $.each(this.points, function () {
+                 s += '<br/>' + chartConf.yAxis.title.text + ': ' + this.y.toFixed(1) + '%';
+             });
+
+             return s;
+         },
+         shared: true
+     },
+     series: {{ chart_data }}
+ };
+</script>
+
+<h3>Data</h3>
+
+<table class="table table-condensed stats-data">
+  <thead>
+    <tr>
+      <th>h-index</th>
+      <th>Percentage of authors</th>
+      <th>Authors</th>
+    </tr>
+  </thead>
+  <tbody>
+    {% for h_index, percentage, names in table_data %}
+      <tr>
+        <td>{{ h_index }}</td>
+        <td>{{ percentage|floatformat:2 }}%</td>
+        <td>{% include "stats/includes/number_with_details_cell.html" with content_limit=25 %}</td>
+      </tr>
+    {% endfor %}
+  </tbody>
+</table>
+
+<p>Hirsch index or h-index is a
+  <a href="https://www.wikipedia.org/wiki/H-index">measure of the
+    productivity and impact of the publications of an author</a>. An
+  author with an h-index of 5 has had 5 publications each cited at
+  least 5 times - to increase the index to 6, the 5 publications plus
+  1 more would have to have been cited at least 6 times, each. Thus a
+  high h-index requires many highly-cited publications.</p>
+
+<p>Note that the h-index calculations do not exclude self-references.</p>

ietf/templates/stats/includes/number_with_details_cell.html

@@ -1 +1,7 @@
-<a class="popover-details" href="" data-elements="{% for n in names|slice:":20" %}{{ n }}{% if not forloop.last %}|{% endif %}{% endfor %}" data-sliced="{% if names|length > 20 %}1{% endif %}">{{ names|length }}</a>
+{% if content_limit and names|length <= content_limit %}
+  {% for n in names %}
+    {{ n }}<br>
+  {% endfor %}
+{% else %}
+  <a class="popover-details" href="" data-elements="{% for n in names|slice:":20" %}{{ n }}{% if not forloop.last %}|{% endif %}{% endfor %}" data-sliced="{% if names|length > 20 %}1{% endif %}">{{ names|length }}</a>
+{% endif %}