[FEAT] Intersection graph.

gargantext/views/api/nodes.py

@@ -15,9 +15,11 @@ import csv
 
 _node_available_fields = ['id', 'parent_id', 'name', 'typename', 'hyperdata', 'ngrams']
 _node_default_fields = ['id', 'parent_id', 'name', 'typename']
-_node_available_formats = ['json', 'csv', 'bibex']
 _node_available_types = NODETYPES
 
+#_hyperdata_available_fields = ['title', 'resourcetype']
+#_node_available_formats = ['json', 'csv', 'bibex']
+
 
 def _query_nodes(request, node_id=None):
     user = cache.User[request.user.id]
@@ -25,6 +27,9 @@ def _query_nodes(request, node_id=None):
     parameters = get_parameters(request)
     parameters = validate(parameters, {'type': dict, 'items': {
         'formated': {'type': str, 'required' : False, 'default': 'json'},
+#        'hyperdata': {'type': list, 'default' : _hyperdata_available_fields, 'items': {
+#            'type': str, 'range' : _node_available_fields,
+#        }},
         'pagination_limit': {'type': int, 'default': 10},
         'pagination_offset': {'type': int, 'default': 0},
         'fields': {'type': list, 'default': _node_default_fields, 'items': {
@@ -75,7 +80,7 @@ class NodeListResource(APIView):
                 'parameters': parameters,
                 'count': count,
                 'records': [
-                    {field: getattr(node, field) for field in parameters['fields']}
+                    { field: getattr(node, field) for field in parameters['fields'] }
                     for node in query
                 ]
             })

graphExplorer/intersection.py

@@ -5,127 +5,71 @@ from gargantext.util.db    import session, aliased, bulk_insert, func
 
 from gargantext.util.lists import WeightedMatrix, UnweightedList, Translations
 from gargantext.util.http  import JsonHttpResponse
-from sqlalchemy            import desc, asc, or_, and_
+from sqlalchemy            import desc, asc, or_, and_, func
 
 import datetime
+import ast
+import networkx as nx
 
 def intersection(request , corpuses_ids, measure='cooc'):
-    FinalDict = False
+    '''
+    intersection :: (str(Int) + "a" str(Int)) -> Dict(Ngram.id :: Int, Score :: Int)
+    intersection = gives the intersection of two graphs
+
+    '''
     if request.method == 'POST' and "nodeids" in request.POST and len(request.POST["nodeids"])>0 :
 
-        import ast
-        import networkx as nx
         node_ids = [int(i) for i in (ast.literal_eval( request.POST["nodeids"] )) ]
         # Here are the visible nodes of the initial semantic map.
 
         corpuses_ids = corpuses_ids.split('a')
 
         corpuses_ids = [int(i) for i in corpuses_ids]
-        print(corpuses_ids)
         # corpus[1] will be the corpus to compare
 
 
             
-        def get_score(corpus_id):
-
-            cooc_ids  = (session.query(Node.id)
-                                .filter(Node.user_id == request.user.id
-                                      , Node.parent_id==corpus_id
-                                      , Node.typename == 'COOCCURRENCES' )
-                                .first()
-                                )
-
-            if len(cooc_ids)==0:
-                return JsonHttpResponse(FinalDict)
-                # If corpus[1] has a coocurrence.id then lets continue
-
-            Coocs  = {}
-
-            G = nx.Graph()
-            # undirected graph only
-            # because direction doesnt matter here
-            # coocs is triangular matrix
-
-            ngrams_data = ( session.query(NodeNgramNgram)
-                                   .filter( NodeNgramNgram.node_id==cooc_ids[0]
-                                          , or_( NodeNgramNgram.ngram1_id.in_( node_ids )
-                                               , NodeNgramNgram.ngram2_id.in_( node_ids )
+        def representativity(corpus_id):
+            ngrams_data = ( session.query(Ngram.id, NodeNgram)
+                                .join(Node, NodeNgram.node_id == Node.id)
+                                .filter( Node.parent_id == corpus_id
+                                       , Node.typename== 'DOCUMENT')
+                                .filter( NodeNgram.weight > 0 )
+                                .filter( Ngram.id.in_(node_ids) )
+                                .group_by(Ngram.id)
+                                .count()
                                 )
+            nodes_count = ( session.query(Node)
+                                   .filter( Node.parent_id == corpus_id
+                                          , Node.typename == 'DOCUMENT'
                                           )
-                                   .group_by(NodeNgramNgram)
-                                   .all()
-                                   )
-
-            for ngram in ngrams_data :
-                # are there visible nodes in the X-axis of corpus to compare ?
-                G.add_edge(  ngram.ngram1_id ,  ngram.ngram2_id , weight=ngram.weight)
-                print(corpus_id, ngram)
-
-            for e in G.edges_iter() :
-                n1 = e[0]
-                n2 = e[1]
-                # print( G[n1][n2]["weight"] , "\t", n1,",",n2 )
-
-                if n1 not in Coocs :
-                    Coocs[n1] = 0
+                                   .count()
+                                   )[0]
         
-                if n2 not in Coocs :
-                    Coocs[n2] = 0
+            result = dict()
+            for ngram_id, somme in ngrams_data:
+                result[ngram_id] = somme / nodes_count
 
-                Coocs[n1] += G[n1][n2]["weight"]
-                Coocs[n2] += G[n1][n2]["weight"]
+            return result
 
-            return(Coocs,G)
+        data_0 = representativity(corpuses_ids[0])
+        data_1 = representativity(corpuses_ids[1])
         
-        Coocs_0,G_0 = get_score( corpuses_ids[0] )
-        Coocs_1,G_1 = get_score( corpuses_ids[1] )
+        queue     = list()
 
-        FinalDict = {}
-
-        if measure == 'jacquard':
-            for node in node_ids :
-                if node in G_1.nodes() and node in G_0.nodes():
-                    neighbors_0 = set(G_0.neighbors(node))
-                    neighbors_1 = set(G_1.neighbors(node))
-                    jacquard = len(neighbors_0.intersection(neighbors_1)) / len(neighbors_0.union(neighbors_1))
-                    FinalDict[node] = jacquard * 3
-                elif node in G_0.nodes() and node not in G_1.nodes() :
-                    FinalDict[node] = 2
-                elif node not in G_0.nodes() and node in G_1.nodes() :
-                    FinalDict[node] = 1
-                else:
-                    FinalDict[node] = 0
-
-        elif measure == 'degree':
-            for node in node_ids :
-                if node in G_1.nodes() and node in G_0.nodes():
-                    score_0 = Coocs_0[node] / G_0.degree(node)
-                    score_1 = Coocs_1[node] / G_1.degree(node)
-                    FinalDict[node] = 5 * score_0 / score_1
-                elif node in G_0.nodes() and node not in G_1.nodes() :
-                    FinalDict[node] = 0.5
-                elif node not in G_0.nodes() and node in G_1.nodes() :
-                    FinalDict[node] = 0.2
-                else:
-                    FinalDict[node] = 0
-
-        elif measure == 'cooc':
-            for node in node_ids :
-                if node in G_1.nodes() and node in G_0.nodes():
-                    #FinalDict[node] = Coocs_1[node] / Coocs_0[node]
-                    FinalDict[node] = Coocs_0[node] / Coocs_1[node]
-                elif node in G_0.nodes() and node not in G_1.nodes() :
-                    FinalDict[node] = 0.0
-                elif node not in G_0.nodes() and node in G_1.nodes() :
-                    FinalDict[node] = 0.0
+        for k in data_0.keys():
+            if data_1[k] == 0:
+                queue.append(k)
             else:
-                    FinalDict[node] = 0
+                FinalDict[k] = data_0[k] / data_1[k]
 
+        maximum = max([ FinalDict[k] for k in FinalDict.keys()])
 
+        for k in queue:
+            FinalDict[k] = maximum +1
 
+        print("-" * 100)
         print(FinalDict)
-                #print(node,score)
-                # Getting AVG-COOC of each ngram that exists in the cooc-matrix of the compared-corpus.
 
         return JsonHttpResponse(FinalDict)
 

static/lib/graphExplorer/extras_explorerjs.js

@@ -815,6 +815,7 @@ function GetUserPortfolio() {
         return true;
 
     var query_url = window.location.origin+'/api/nodes?types[]=PROJECT&types[]=CORPUS&pagination_limit=100'
+    // var query_url = window.location.origin+'/api/nodes?types[]=PROJECT&types[]=CORPUS&pagination_limit=100&fields[]=hyperdata'
     $.ajax({
         type: 'GET',
         dataType : 'JSON',