memcached for localDB CSV searchbase

with this optional cache, indexation happens only once every 1800s, and subsequent queries retrieve the search base (aka postings) from a serialized version in cache

LOCALDB/README.md

+
+## Related Documents PHP backend
+
+##### Provides an API for "topPapers" search queries in tinawebJS
+
+Main use case is support for one-doc-by-row CSV files.
+
+#### Prerequisites
+
+You need any kind of php server support with php > 5.0.
+
+For instance on an ubuntu 16 with an nginx server, I'd install php 7 fpm:
+```
+sudo apt install php7.0-fpm
+```
+
+And then add this kind of configuration entry in `nginx.conf`:
+```
+location ~ \.php$ {
+  include snippets/fastcgi-php.conf;
+  fastcgi_pass unix:/run/php/php7.0-fpm.sock;
+
+ fastcgi_param SCRIPT_FILENAME /your/path/to/LOCALDB/$fastcgi_script_name;
+ #                             ---------------------
+}
+```
+
+It's enough to run any kind of gargantext-style gexf + CSV sets.
+
+Optionally, especially in production, you should add memcached support for faster CSV search (it allows caching the CSV postings base).
+
+```
+sudo apt install php-memcached
+sudo service php7.0-fpm restart
+```
+
+For legacy cortext-style databases, you'll also need sqlite:
+```
+sudo apt install sqlite3
+```
+
+
+#### Usage
+For a given graph source and database, you will need to **fill in the right settings in `db.json`**  in order to associate the graph source file (eg gexf) with this related docs API.
+
+£TODO explain after new specifications implemented

LOCALDB/info_div.php

@@ -2,7 +2,7 @@
 // manage the dynamical additional information in the left panel.
 ini_set('display_errors',1);
 ini_set('display_startup_errors',1);
-error_reporting(-1);
+// error_reporting(-1);
 
 // relative path to dirname "/line"
 $project_root = "../";
@@ -27,24 +27,49 @@ else {
 
     // echodump("columns to index",$idxcolsbytype);
 
-    include('csv_indexation.php');
 
-    // DO THE INDEXATION
+    // DO THE INDEXATION (or RETRIEVE CACHED ONE)
     // we use cache if memcached is present (and if we indexed the csv already)
-    // $can_use_cache = False
+    include('csv_indexation.php');
+
+    $csv_search_base = NULL;
+    if(class_exists('Memcached')){
+      $mcd = new Memcached;
+      $mcd->addServer($memserver, $memport);
 
-    // £TODO use memcached or something to store a serialized version of csv_search_base
-    // + add all (sem+soc) columns for the index to work !!
-    $csv_search_base = parse_and_index_csv($project_root.$graphdb, $idxcolsbytype, $csvsep, $csvquote);
+      // test if we indexed it already
+      $csv_search_base_seri = $mcd->get(mem_entry_name($graphdb));
+
+      if ($csv_search_base_seri !== False) {
+        // echo("Using cached base<br>");
+        $csv_search_base = json_decode($csv_search_base_seri, $assoc=true);
+      }
+    }
+
+    if (! $csv_search_base) {
+      // echo("Creating new base<br>");
+
+      // indexing ----------------------------------------------------------------------
+      // must know about all (sem+soc) cols typed, even if each search doesn't need them
+      $csv_search_base = parse_and_index_csv($project_root.$graphdb,
+                                             $idxcolsbytype,
+                                            $csvsep, $csvquote);
+      // -------------------------------------------------------------------------------
+
+      if(class_exists('Memcached')){
+        // **store** in cache for 1/2h
+        $mcd->set(mem_entry_name($graphdb), json_encode($csv_search_base), 1800);
+      }
+    }
 
     $base = $csv_search_base[0];
     $postings = $csv_search_base[1];
-
     // echodump("postings", $postings);
     // echodump("base", $base);
 
-    // DO THE SEARCH
 
+    // DO THE SEARCH
+    // -------------
     $searchcols = json_decode($_GET['searchin']);
 
     // a - split the query
@@ -105,6 +130,12 @@ else {
 }
 
 
+// just to make sure we use the same conventional name at cache read/write
+function mem_entry_name($graphdbname) {
+  return "twbackendmem/".$graphdbname."/csv_search_base";
+}
+
+
 
 
 ?>

settings_explorerjs.js

@@ -37,7 +37,7 @@ TW.conf = (function(TW){
   TWConf.getRelatedDocs = true
   TWConf.relatedDocsMax = 10
 
-  TWConf.relatedDocsType = "twitter"   // accepted: "twitter" | "wosLocalDB"
+  TWConf.relatedDocsType = "wosLocalDB"   // accepted: "twitter" | "wosLocalDB"
                                           // POSSible: "elastic"
 
   TWConf.relatedDocsAPIS = {