From 14a6e2423bee0edbbc2f620ecc7d90727e6186aa Mon Sep 17 00:00:00 2001
From: voje <kristjan.voje@gmail.com>
Date: Sun, 28 Apr 2019 23:24:40 +0200
Subject: [PATCH] logging when indexing

---
 src/backend_flask/app.py | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/src/backend_flask/app.py b/src/backend_flask/app.py
index b4bcae8..3f27281 100644
--- a/src/backend_flask/app.py
+++ b/src/backend_flask/app.py
@@ -453,6 +453,9 @@ def prepare_app_index(appindex_json):
     for corpus in CORPORA:
         res_hws = {}
         res_fns = {}
+
+        nentries = valdb[corpus].count()
+        idx = 0
         for e in valdb[corpus].find({}):
             if "headwords" not in e:
                 continue
@@ -468,6 +471,10 @@ def prepare_app_index(appindex_json):
                     res_fns[fn] += 1
                 else:
                     res_fns[fn] = 1
+            idx += 1
+            if idx % 10000 == 0:
+                log.debug("indexing {}: {}/{}".format(
+                    corpus, idx, nentries))
 
         alphabetical = {}
         for k, e in res_hws.items():