[UPDATE] corpora single-comparison

gargantext_web/urls.py

@@ -79,8 +79,8 @@ urlpatterns = patterns('',
     url(r'^tests/istextquery$', pubmedscrapper.getGlobalStatsISTEXT), # api/query?type=istext ?
     url(r'^tests/pubmedquery$', pubmedscrapper.getGlobalStats),
     url(r'^tests/project/(\d+)/pubmedquery/go$', pubmedscrapper.doTheQuery),
-    url(r'^tests/project/(\d+)/ISTEXquery/go$', pubmedscrapper.testISTEX)
-
+    url(r'^tests/project/(\d+)/ISTEXquery/go$', pubmedscrapper.testISTEX),
+    url(r'^tests/paginator/project/(\d+)/corpus/(\d+)/$', views.newpaginatorJSON)
 )
 
 

gargantext_web/views.py

@@ -27,6 +27,7 @@ from collections import defaultdict
 
 from parsing.FileParsers import *
 import os
+import json
 
 # SOME FUNCTIONS
 
@@ -305,6 +306,56 @@ def corpus(request, project_id, corpus_id):
     return HttpResponse(html)
 
 
+def newpaginatorJSON(request , project_id , corpus_id):
+
+    results = ["hola" , "mundo"]
+
+    # t = get_template('tests/newpag/thetable.html')
+    
+    project = session.query(Node).filter(Node.id==project_id).first()
+    corpus  = session.query(Node).filter(Node.id==corpus_id).first()
+    type_document_id = cache.NodeType['Document'].id
+    documents  = session.query(Node).filter(Node.parent_id==corpus_id , Node.type_id == type_document_id ).all()
+
+    filtered_docs = []
+    for doc in documents:
+        if "publication_date" in doc.metadata:
+            try:
+                realdate = doc.metadata["publication_date"].split(" ")[0] # in database is = (year-month-day = 2015-01-06 00:00:00 = 06 jan 2015 00 hrs)
+                realdate = datetime.datetime.strptime(str(realdate), '%Y-%m-%d').date() # finalform = (yearmonthday = 20150106 = 06 jan 2015)
+                # doc.date = realdate
+                resdict = {}
+                resdict["id"] = doc.id 
+                resdict["date"] = realdate
+                resdict["name"] =  doc.name
+                filtered_docs.append( resdict )
+            except Exception as e:
+                print ("pag2 error01 detail:",e)
+                print("pag2 error01 doc:",doc)
+    results = sorted(filtered_docs, key=lambda x: x["date"])
+    for i in results:
+        i["date"] = i["date"].strftime("%Y-%m-%d")
+
+    finaldict = {
+        "records":results,
+        "queryRecordCount":10,
+        "totalRecordCount":len(results)
+    }
+    return JsonHttpResponse(finaldict)
+    # html = t.render(Context({\
+    #         # 'user': user,\
+    #         # 'date': date,\
+    #         'project': project_id,\
+    #         'corpus' : corpus_id,\
+    #         # 'documents': results,\
+    #         # 'number' : len(filtered_docs),\
+    #         # 'dates' : chart,\
+    #         }))
+    
+    # return HttpResponse(html)
+
+
+# Im using this actually
 from django.core.paginator import Paginator, EmptyPage, PageNotAnInteger
 def subcorpus(request, project_id, corpus_id, start , end ):
     if not request.user.is_authenticated():
@@ -380,7 +431,6 @@ def subcorpus(request, project_id, corpus_id, start , end ):
     return HttpResponse(html)
 
 
-import json
 def subcorpusJSON(request, project_id, corpus_id, start , end ):
     if not request.user.is_authenticated():
         return redirect('/login/?next=%s' % request.path)
@@ -559,11 +609,34 @@ def graph(request, project_id, corpus_id):
     project = session.query(Node).filter(Node.id==project_id).first()
     corpus  = session.query(Node).filter(Node.id==corpus_id).first()
 
+
+    user_id         = cache.User[request.user.username].id
+    project_type_id = cache.NodeType['Project'].id
+    corpus_type_id = cache.NodeType['Corpus'].id
+
+    results = {}
+    projs = session.query(Node).filter(Node.type_id==project_type_id).all()
+    for i in projs:
+        # print(i.id , i.name)
+        if i.id not in results: results[i.id] = {}
+        results[i.id]["proj_name"] = i.name
+        results[i.id]["corpuses"] = []
+        corpuses = session.query(Node).filter(Node.parent_id==i.id , Node.type_id==corpus_type_id).all()
+        for j in corpuses:
+            if int(j.id)!=int(corpus_id):
+                info = { "id":j.id , "name":j.name }
+                results[i.id]["corpuses"].append(info)
+                # print("\t",j.id , j.name)
+
+    import pprint
+    pprint.pprint(results)
+
     html = t.render(Context({\
             'user'      : user,\
             'date'      : date,\
             'corpus'    : corpus,\
             'project'   : project,\
+            'corpusinfo'   : results,\
             'graphfile' : "hola_mundo",\
             }))
     
@@ -667,7 +740,6 @@ def send_csv(request, corpus_id):
 # To get the data
 from gargantext_web.api import JsonHttpResponse
 from analysis.functions import get_cooc
-import json
 def node_link(request, corpus_id):
     '''
     Create the HttpResponse object with the node_link dataset.

gargantext_web/views_optimized.py

@@ -199,8 +199,12 @@ def tfidf(request, corpus_id, ngram_ids):
         .order_by(func.sum(NodeNodeNgram.score).desc())
         .limit(limit)
     )
+    # print("\n")
+    # print("in TFIDF:")
+    # print("\tcorpus_id:",corpus_id)
     # convert query result to a list of dicts
     for node, score in nodes_query:
+        print("\t corpus:",corpus_id,"\t",node.name)
         node_dict = {
             'id': node.id,
             'score': score,
@@ -210,5 +214,6 @@ def tfidf(request, corpus_id, ngram_ids):
                 node_dict[key] = node.metadata[key]
         nodes_list.append(node_dict)
 
+    # print("= = = = = = = = \n")
     data = json.dumps(nodes_list) 
     return JsonHttpResponse(data)

templates/explorer.html

@@ -4,7 +4,7 @@
 <link rel="stylesheet" href="{% static "css/bootstrap.css" %}">
 <link rel="stylesheet" href="{% static "css/bootstrap-theme.min.css" %}">
 <link rel="stylesheet" href="{% static "js/libs/jquery/jquery-ui.css" %}" media="screen">
-<link rel="stylesheet" href="{% static "js/libs/bootstrap/css/bootstrap.css" %}" media="screen">
+<!-- <link rel="stylesheet" href="{% static "js/libs/bootstrap/css/bootstrap.css" %}" media="screen"> -->
 <link rel="stylesheet" href="{% static "js/libs/css2/freshslider.css" %}" media="screen">
 <link rel="stylesheet" href="{% static "js/libs/css2/custom.css" %}" media="screen">
 <link rel="stylesheet" href="{% static "js/libs/css2/sidebar.css" %}" media="screen">
@@ -260,13 +260,13 @@
 
         </div>
         
-        <div id="topPapers"></div>
-        <!--
+        <!-- <div id="topPapers"></div> -->
+        
         <div id="tab-container-top" class='tab-container'>
 
            <ul class='etabs'>
              <li id="tabmed" class='tab active'><a href="#tabs3">Medline Pubs</a></li>
-             <li id="tabgps" class='tab'><a href="#tabs4">+</a></li>
+             <li id="tabgps" class='tab'><a onclick="$('#corpuses').modal('show');">+</a></li>
            </ul>
 
            <div class='panel-container'>
@@ -278,7 +278,7 @@
               </div>
            </div>
         </div>
-        -->
+        
 
 
 
@@ -350,6 +350,70 @@
 </div> 
 
 
+
+<div id="corpuses" class="modal fade">
+  <div class="modal-dialog">
+    <div class="modal-content">
+
+
+      <div class="modal-header">
+        <button type="button" class="close" data-dismiss="modal" aria-hidden="true">×</button>
+        <h4 class="modal-title">Corpus Comparison</h4>
+      </div>
+
+
+
+      <div class="modal-body form-horizontal">
+        Chose another corpus to compare with:
+        <div class="form-group">
+          <label class="col-lg-2 control-label"> </label>
+          <div class="col-lg-10">
+
+
+              <form id="corpuses_form" role="form">
+
+                <ul>
+
+                  {% if corpusinfo  %}
+                    {% for k1, v1 in corpusinfo.items %}
+                        {% if v1.corpuses|length > 0  %}
+                            <br><li><a href="/project/{{k1}}/">{{v1.proj_name}}</a><br>
+                              <ul style="list-style-type: none;">
+                              {% for c in v1.corpuses %}
+                                <li>
+                                  <div class="radio">
+                                    <label><input type="radio" id="{{c.id}}" name="optradio">
+                                      <a href="/project/{{k1}}/corpus/{{c.id}}/">{{c.name}}</a>
+                                    </label>
+                                  </div>
+                                </li>
+                              {% endfor %}
+                              </ul>
+                            </li>
+                        {% endif %}
+                    {% endfor %}
+                  {% endif %}
+
+                </ul>
+
+              </form>
+
+
+          </div>
+        </div>
+      </div>
+
+
+      <div class="modal-footer">
+        <button id="closecorpuses" type="button" class="btn btn-default" data-dismiss="modal">Close</button>
+        <button type="button" class="btn btn-primary" onclick='printCorpuses();'>Add Tab</button>
+      </div>
+
+    </div>
+  </div>
+</div> 
+
+
 <div id="modalloader" class="modal fade">
     <div id="loader" class="loader">
         <img src="{% static "js/libs/img2/loader.gif" %}" ></img>

templates/project.html

@@ -282,7 +282,7 @@
 				$("#simpleloader").html('<img width="30px" src="{% static "js/libs/img2/loading-bar.gif" %}"></img>')
 				$("#submit_thing").prop('onclick',null);
 				var theType = $("#id_type option:selected").html();
-				if(theType=="pubmed") doTheQuery();
+				if(theType=="Pubmed (xml format)") doTheQuery();
 				if(theType=="istex") {
 					testISTEX(origQuery.replace(" ","+"),1000)
 				}