fix details of hashtags filter to sql constraints to map procedure

services/dbdatapi.py

@@ -12,7 +12,7 @@ from networkx  import Graph, DiGraph
 from random    import randint
 from math      import floor, log, log1p
 from cgi       import escape
-from re        import sub
+from re        import sub, match
 from traceback import format_tb
 
 if __package__ == 'services':
@@ -392,6 +392,25 @@ class BipartiteExtractor:
                             # "LIKE_relation" or "EQ_relation"
                             rel_type = FIELDS_FRONTEND_TO_SQL[key]['type']
 
+
+                        # pre-treatment: rewrite tables' names if they're inside the sub-query
+
+                        # exemple:
+                        # scholars.country   ~~~~~> scholars_n_hashtags.country
+                        # hashtags.htstr     ~~~~~> scholars_n_hashtags.htstr
+                        # (see cascaded join below for explanation)
+
+                        if match("scholars", sql_column):
+                            (sql_table, sql_field) = sql_column.split('.')
+                            sql_column = 'scholars_n_hashtags.'+sql_field
+
+                            mlog('DBG', "rewrote sql col", sql_column)
+                        elif match("hashtags.htstr", sql_column):
+                            sql_column = 'scholars_n_hashtags.hashtags_list'
+
+                            mlog('DBG', "rewrote sql col", sql_column)
+
+                        # now create the constraints
                         val = filter_dict[known_filter]
 
                         if len(val):
@@ -442,19 +461,35 @@ class BipartiteExtractor:
                     mlog("INFO", "SELECTing active users with sql_constraints", sql_constraints)
 
                     # use constraints as WHERE-clause
+
+                    # NB we must cascade join because
+                    #    both hashtags and keywords are one-to-many
+                    #   => it renames scholars and hashtag tables
+                    #      into 'scholars_n_hashtags'
                     sql_query = """
                         SELECT
-                            scholars.luid,
+                            scholars_n_hashtags.luid,
+                            scholars_n_hashtags.affiliation_id,
 
                             -- kws info
-                            COUNT(keywords.kwid) AS keywords_nb,
-                            GROUP_CONCAT(keywords.kwstr) AS keywords_list,
-                            GROUP_CONCAT(keywords.kwid) AS keywords_ids
-
-                        FROM scholars
+                            GROUP_CONCAT(keywords.kwstr) AS keywords_list
+
+                        FROM (
+                            SELECT
+                                scholars.*,
+                                -- hts info
+                                GROUP_CONCAT(hashtags.htstr) AS hashtags_list
+
+                            FROM scholars
+                            LEFT JOIN sch_ht
+                                ON uid = luid
+                            JOIN hashtags
+                                ON sch_ht.htid = hashtags.htid
+                            GROUP BY luid
+                        ) AS scholars_n_hashtags
 
                         -- two step JOIN for keywords
-                        JOIN sch_kw
+                        LEFT JOIN sch_kw
                             ON uid = luid
                         JOIN keywords
                             ON sch_kw.kwid = keywords.kwid

services/main.py

@@ -42,7 +42,6 @@ if __package__ == 'services':
     from services          import tools, dbcrud, dbdatapi
     from services.user     import User, login_manager, \
                                   doors_login, doors_register
-    from services.dbdatapi import BipartiteExtractor
     from services.text.utils import sanitize
 else:
     # when this script is run directly
@@ -51,7 +50,6 @@ else:
     import tools, dbcrud, dbdatapi
     from user           import User, login_manager, \
                                doors_login, doors_register
-    from dbdatapi       import BipartiteExtractor
     from text.utils      import sanitize
 
 # ============= app creation ============
@@ -220,7 +218,13 @@ def graph_api():
     (original author S. Castillo)
     """
     if 'qtype' in request.args:
-        graphdb = BipartiteExtractor(config['SQL_HOST'])
+        graphdb = dbdatapi.BipartiteExtractor(config['SQL_HOST'])
+
+        # request.query_string
+        #                => b'qtype=filters&tags[]=%23iscpif'
+        # tools.restparse(request.query_string.decode())
+        #                => {'qtype': 'filters', 'tags': ['#iscpif']}
+
         scholars = graphdb.getScholarsList(
                     request.args['qtype'],
                     tools.restparse(

static/tinawebJS/main.js

@@ -303,7 +303,7 @@ function bringTheNoise(sourceinfo,type){
                 var nameSubElts = []
                 for (var value of filteringKeyArrayPairs[fieldName]) {
                     // exemple: "countries[]=France"
-                    restParams.push(fieldName+"[]="+value)
+                    restParams.push(fieldName+'[]='+encodeURIComponent(value))
                     nameSubElts.push ('"'+value+'"')
                 }
                 nameElts.push("("+nameSubElts.join(" or ")+")")