port tests for takeout from kython

tests/README.org

@@ -0,0 +1,3 @@
+At the moment, some tests here are specific to my data.
+
+If you know any good public data sources I could test against instead, please let me know!

tests/takeout.py

@@ -1,11 +1,14 @@
 #!/usr/bin/env python3
+from datetime import datetime
 from itertools import islice
+import pytz
 
 from my.core.cachew import disable_cachew
 disable_cachew()
 
 import my.location.takeout as LT
-from my.kython.kompress import kopen
+from my.google.takeout.html import read_html
+from my.google.takeout.paths import get_last_takeout
 
 
 def ilen(it):
@@ -32,18 +35,25 @@ import pytest # type: ignore
 )
 def test_parser(path: str):
     path = 'Takeout/' + path
-    from my.google.takeout.html import read_html
-    from my.google.takeout.paths import get_last_takeout
-
     tpath = get_last_takeout(path=path)
-
+    results = list(read_html(tpath, path))
-    results = []
+    # TODO assert len > 100 or something?
-    for res in read_html(tpath, path):
-        results.append(res)
-
     print(len(results))
 
 
+def test_myactivity_search():
+    path = 'Takeout/My Activity/Search/MyActivity.html'
+    tpath = get_last_takeout(path=path)
+    results = list(read_html(tpath, path))
+
+    res = (
+        datetime(year=2018, month=12, day=17, hour=8, minute=16, second=18, tzinfo=pytz.utc),
+        'https://en.wikipedia.org/wiki/Emmy_Noether&usg=AFQjCNGrSW-iDnVA2OTcLsG3I80H_a6y_Q',
+        'Emmy Noether - Wikipedia',
+    )
+    assert res in results
+
+
 def parse_takeout_xmllint(data: str):
     # without xmllint (splitting by '<div class="content-cell' -- 0.68 secs)
     # with xmllint -- 2 seconds

tests/youtube.py

@@ -3,9 +3,18 @@
 # these tests would only make sense with some existing data? although some of them would work for everyone..
 # not sure what's a good way of handling this..
 
-from my.media.youtube import get_watched
+from my.media.youtube import get_watched, Watched
 
 
 def test():
     watched = get_watched()
     assert len(watched) > 1000
+
+    from datetime import datetime
+    import pytz
+    w = Watched(
+        url='https://www.youtube.com/watch?v=hTGJfRPLe08',
+        title='Jamie xx - Gosh',
+        when=datetime(year=2018, month=6, day=21, hour=5, minute=48, second=34, tzinfo=pytz.utc),
+    )
+    assert w in watched