Zipnum index: do fail if counting pages with filter params (#631)

- do not apply any filters (param filter, from, to, closest) if counting pages (param showNumPages=true)
webrecorder · Apr 27, 2021 · 73d6735 · 73d6735
1 parent cdb17c4
commit 73d6735
Show file tree

Hide file tree

Showing 2 changed files with 11 additions and 1 deletion.
diff --git a/pywb/warcserver/index/aggregator.py b/pywb/warcserver/index/aggregator.py
@@ -37,7 +37,9 @@ def __call__(self, params):
 
         cdx_iter, errs = self.load_index(query.params)
 
-        cdx_iter = process_cdx(cdx_iter, query)
+        if not query.page_count:
+            cdx_iter = process_cdx(cdx_iter, query)
+
         return cdx_iter, dict(errs)
 
     def load_child_source(self, name, source, params):

diff --git a/pywb/warcserver/index/test/test_zipnum.py b/pywb/warcserver/index/test/test_zipnum.py
@@ -228,6 +228,14 @@ def test_blocks_zero_pages():
     res = zip_ops_test_data(url='http://aaa.zz/', matchType='domain', showNumPages=True)
     assert(res == {"blocks": 0, "pages": 0, "pageSize": 10})
 
+def test_blocks_ignore_filter_params():
+    res = zip_ops_test_data(url='*.iana.org', pageSize='4', showNumPages=True, filter='=status:200')
+    assert(res == {"blocks": 38, "pages": 10, "pageSize": 4})
+
+def test_blocks_ignore_timestamp_params():
+    res = zip_ops_test_data(url='*.iana.org', pageSize='4', showNumPages=True, closest='20140126000000')
+    assert(res == {"blocks": 38, "pages": 10, "pageSize": 4})
+
 
 # Errors