Make TWBlue to respect persist_size when loading the cache database. Added tests for both reading from disk and loading in memory caches

This commit is contained in:
Manuel Cortez 2021-06-27 02:32:28 -05:00
parent 35d6010298
commit 0612c653b8
4 changed files with 225 additions and 7 deletions

18
src/run_tests.py Normal file
View File

@ -0,0 +1,18 @@
# -*- coding: utf-8 -*-
import unittest
testmodules = ["test.test_cache"]
suite = unittest.TestSuite()
for t in testmodules:
try:
# If the module defines a suite() function, call it to get the suite.
mod = __import__(t, globals(), locals(), ['suite'])
suitefn = getattr(mod, 'suite')
suite.addTest(suitefn())
except (ImportError, AttributeError):
# else, just load all the test cases from the module.
suite.addTest(unittest.defaultTestLoader.loadTestsFromName(t))
unittest.TextTestRunner(verbosity=2).run(suite)

View File

@ -72,10 +72,10 @@ class baseSession(object):
""" Returns a list with the amount of items specified by size.""" """ Returns a list with the amount of items specified by size."""
if isinstance(buffer, list) and size != -1 and len(buffer) > size: if isinstance(buffer, list) and size != -1 and len(buffer) > size:
log.debug("Requesting {} items from a list of {} items. Reversed mode: {}".format(size, len(buffer), reversed)) log.debug("Requesting {} items from a list of {} items. Reversed mode: {}".format(size, len(buffer), reversed))
if reversed == False: if reversed == True:
return buffer[size:]
else:
return buffer[:size] return buffer[:size]
else:
return buffer[len(buffer)-size:]
else: else:
return buffer return buffer
@ -91,17 +91,20 @@ class baseSession(object):
# Let's check if we need to create a new SqliteDict object (when loading db in memory) or we just need to call to commit in self (if reading from disk).db. # Let's check if we need to create a new SqliteDict object (when loading db in memory) or we just need to call to commit in self (if reading from disk).db.
# If we read from disk, we cannot modify the buffer size here as we could damage the app's integrity. # If we read from disk, we cannot modify the buffer size here as we could damage the app's integrity.
# We will modify buffer's size (managed by persist_size) upon loading the db into memory in app startup. # We will modify buffer's size (managed by persist_size) upon loading the db into memory in app startup.
if self.settings["general"]["load_cache_in_memory"]: if self.settings["general"]["load_cache_in_memory"] and isinstance(self.db, dict):
log.debug("Opening database to dump memory contents...") log.debug("Opening database to dump memory contents...")
db=sqlitedict.SqliteDict(dbname, 'c') db=sqlitedict.SqliteDict(dbname, 'c')
for k in self.db.keys(): for k in self.db.keys():
db[k] = self.get_sized_buffer(self.db[k], self.settings["general"]["persist_size"], self.settings["general"]["reverse_timelines"]) sized_buff = self.get_sized_buffer(self.db[k], self.settings["general"]["persist_size"], self.settings["general"]["reverse_timelines"])
db[k] = sized_buff
db.commit(blocking=True)
db.close() db.close()
log.debug("Data has been saved in the database.") log.debug("Data has been saved in the database.")
else: else:
try: try:
log.debug("Syncing new data to disk...") log.debug("Syncing new data to disk...")
self.db.commit() if hasattr(self.db, "commit"):
self.db.commit()
except: except:
output.speak(_("An exception occurred while saving the {app} database. It will be deleted and rebuilt automatically. If this error persists, send the error log to the {app} developers.").format(app=application.name),True) output.speak(_("An exception occurred while saving the {app} database. It will be deleted and rebuilt automatically. If this error persists, send the error log to the {app} developers.").format(app=application.name),True)
log.exception("Exception while saving {}".format(dbname)) log.exception("Exception while saving {}".format(dbname))
@ -128,6 +131,7 @@ class baseSession(object):
log.debug("Loading database contents into memory...") log.debug("Loading database contents into memory...")
for k in db.keys(): for k in db.keys():
self.db[k] = db[k] self.db[k] = db[k]
db.commit(blocking=True)
db.close() db.close()
log.debug("Contents were loaded successfully.") log.debug("Contents were loaded successfully.")
else: else:
@ -136,7 +140,8 @@ class baseSession(object):
# We must make sure we won't load more than the amount of buffer specified. # We must make sure we won't load more than the amount of buffer specified.
log.debug("Checking if we will load all content...") log.debug("Checking if we will load all content...")
for k in self.db.keys(): for k in self.db.keys():
self.db[k] = self.get_sized_buffer(self.db[k], self.settings["general"]["persist_size"], self.settings["general"]["reverse_timelines"]) sized_buffer = self.get_sized_buffer(self.db[k], self.settings["general"]["persist_size"], self.settings["general"]["reverse_timelines"])
self.db[k] = sized_buffer
if self.db.get("cursors") == None: if self.db.get("cursors") == None:
cursors = dict(direct_messages=-1) cursors = dict(direct_messages=-1)
self.db["cursors"] = cursors self.db["cursors"] = cursors

1
src/test/__init__.py Normal file
View File

@ -0,0 +1 @@
# -*- coding: utf-8 -*-

194
src/test/test_cache.py Normal file
View File

@ -0,0 +1,194 @@
# -*- coding: utf-8 -*-
""" Test case to check some of the scenarios we might face when storing tweets in cache, both loading into memory or rreading from disk. """
import unittest
import os
import paths
import sqlitedict
import shutil
# The base session module requires sound as a dependency, and this needs libVLC to be locatable.
os.environ['PYTHON_VLC_MODULE_PATH']=os.path.abspath(os.path.join(paths.app_path(), "..", "windows-dependencies", "x86"))
os.environ['PYTHON_VLC_LIB_PATH']=os.path.abspath(os.path.join(paths.app_path(), "..", "windows-dependencies", "x86", "libvlc.dll"))
from sessions import base
class cacheTestCase(unittest.TestCase):
def setUp(self):
""" Configures a fake session to check caching objects here. """
self.session = base.baseSession("testing")
if os.path.exists(os.path.join(paths.config_path(), "testing")) == False:
os.mkdir(os.path.join(paths.config_path(), "testing"))
self.session.get_configuration()
def tearDown(self):
""" Removes the previously configured session. """
session_folder = os.path.join(paths.config_path(), "testing")
if os.path.exists(session_folder):
shutil.rmtree(session_folder)
def generate_dataset(self):
""" Generates a sample dataset"""
dataset = dict(home_timeline=["message" for i in range(10000)], mentions_timeline=["mention" for i in range(20000)])
return dataset
### Testing database being read from disk.
def test_cache_in_disk_unlimited_size(self):
""" Tests cache database being read from disk, storing the whole datasets. """
dataset = self.generate_dataset()
self.session.settings["general"]["load_cache_in_memory"] = False
self.session.settings["general"]["persist_size"] = -1
self.session.load_persistent_data()
self.session.db["home_timeline"] = dataset["home_timeline"]
self.session.db["mentions_timeline"] = dataset["mentions_timeline"]
self.session.save_persistent_data()
self.assertIsInstance(self.session.db, sqlitedict.SqliteDict)
self.assertTrue(self.session.db.get("home_timeline") != None)
self.assertTrue(self.session.db.get("mentions_timeline") != None)
self.assertEquals(len(self.session.db.get("home_timeline")), 10000)
self.assertEquals(len(self.session.db.get("mentions_timeline")), 20000)
self.session.db.close()
def test_cache_in_disk_limited_dataset(self):
""" Tests wether the cache stores only the amount of items we ask it to store. """
dataset = self.generate_dataset()
self.session.settings["general"]["load_cache_in_memory"] = False
self.session.settings["general"]["persist_size"] = 100
self.session.load_persistent_data()
self.session.db["home_timeline"] = dataset["home_timeline"]
self.session.db["mentions_timeline"] = dataset["mentions_timeline"]
# We need to save and load the db again because we cannot modify buffers' size while the database is opened.
# As TWBlue reads directly from db when reading from disk, an attempt to modify buffers size while Blue is reading the db
# Might cause an out of sync error between the GUI lists and the database.
# So we perform the changes to buffer size when loading data during app startup if the DB is read from disk.
self.session.save_persistent_data()
self.session.load_persistent_data()
self.assertIsInstance(self.session.db, sqlitedict.SqliteDict)
self.assertTrue(self.session.db.get("home_timeline") != None)
self.assertTrue(self.session.db.get("mentions_timeline") != None)
self.assertEquals(len(self.session.db.get("home_timeline")), 100)
self.assertEquals(len(self.session.db.get("mentions_timeline")), 100)
self.session.db.close()
def test_cache_in_disk_limited_dataset_unreversed(self):
"""Test if the cache is saved properly in unreversed buffers, when newest items are at the end of the list. """
dataset = dict(home_timeline=[i for i in range(20)], mentions_timeline=[i for i in range(20)])
self.session.settings["general"]["load_cache_in_memory"] = False
self.session.settings["general"]["persist_size"] = 10
self.session.load_persistent_data()
self.session.db["home_timeline"] = dataset["home_timeline"]
self.session.db["mentions_timeline"] = dataset["mentions_timeline"]
# We need to save and load the db again because we cannot modify buffers' size while the database is opened.
# As TWBlue reads directly from db when reading from disk, an attempt to modify buffers size while Blue is reading the db
# Might cause an out of sync error between the GUI lists and the database.
# So we perform the changes to buffer size when loading data during app startup if the DB is read from disk.
self.session.save_persistent_data()
self.session.load_persistent_data()
self.assertIsInstance(self.session.db, sqlitedict.SqliteDict)
self.assertTrue(self.session.db.get("home_timeline") != None)
self.assertTrue(self.session.db.get("mentions_timeline") != None)
self.assertEquals(self.session.db.get("home_timeline")[0], 10)
self.assertEquals(self.session.db.get("mentions_timeline")[0], 10)
self.assertEquals(self.session.db.get("home_timeline")[-1], 19)
self.assertEquals(self.session.db.get("mentions_timeline")[-1], 19)
self.session.db.close()
def test_cache_in_disk_limited_dataset_reversed(self):
"""Test if the cache is saved properly in reversed buffers, when newest items are at the start of the list. """
dataset = dict(home_timeline=[i for i in range(19, -1, -1)], mentions_timeline=[i for i in range(19, -1, -1)])
self.session.settings["general"]["load_cache_in_memory"] = False
self.session.settings["general"]["persist_size"] = 10
self.session.settings["general"]["reverse_timelines"] = True
self.session.load_persistent_data()
self.session.db["home_timeline"] = dataset["home_timeline"]
self.session.db["mentions_timeline"] = dataset["mentions_timeline"]
# We need to save and load the db again because we cannot modify buffers' size while the database is opened.
# As TWBlue reads directly from db when reading from disk, an attempt to modify buffers size while Blue is reading the db
# Might cause an out of sync error between the GUI lists and the database.
# So we perform the changes to buffer size when loading data during app startup if the DB is read from disk.
self.session.save_persistent_data()
self.session.load_persistent_data()
self.assertIsInstance(self.session.db, sqlitedict.SqliteDict)
self.assertTrue(self.session.db.get("home_timeline") != None)
self.assertTrue(self.session.db.get("mentions_timeline") != None)
self.assertEquals(self.session.db.get("home_timeline")[0], 19)
self.assertEquals(self.session.db.get("mentions_timeline")[0], 19)
self.assertEquals(self.session.db.get("home_timeline")[-1], 10)
self.assertEquals(self.session.db.get("mentions_timeline")[-1], 10)
self.session.db.close()
### Testing database being loaded into memory. Those tests should give the same results than before
### but as we have different code depending whether we load db into memory or read it from disk,
### We need to test this anyways.
def test_cache_in_memory_unlimited_size(self):
""" Tests cache database being loaded in memory, storing the whole datasets. """
dataset = self.generate_dataset()
self.session.settings["general"]["load_cache_in_memory"] = True
self.session.settings["general"]["persist_size"] = -1
self.session.load_persistent_data()
self.session.db["home_timeline"] = dataset["home_timeline"]
self.session.db["mentions_timeline"] = dataset["mentions_timeline"]
self.session.save_persistent_data()
self.session.load_persistent_data()
self.assertIsInstance(self.session.db, dict)
self.assertTrue(self.session.db.get("home_timeline") != None)
self.assertTrue(self.session.db.get("mentions_timeline") != None)
self.assertEquals(len(self.session.db.get("home_timeline")), 10000)
self.assertEquals(len(self.session.db.get("mentions_timeline")), 20000)
def test_cache_in_memory_limited_dataset(self):
""" Tests wether the cache stores only the amount of items we ask it to store, when loaded in memory. """
dataset = self.generate_dataset()
self.session.settings["general"]["load_cache_in_memory"] = True
self.session.settings["general"]["persist_size"] = 100
self.session.load_persistent_data()
self.session.db["home_timeline"] = dataset["home_timeline"]
self.session.db["mentions_timeline"] = dataset["mentions_timeline"]
self.session.save_persistent_data()
self.session.load_persistent_data()
self.assertIsInstance(self.session.db, dict)
self.assertTrue(self.session.db.get("home_timeline") != None)
self.assertTrue(self.session.db.get("mentions_timeline") != None)
self.assertEquals(len(self.session.db.get("home_timeline")), 100)
self.assertEquals(len(self.session.db.get("mentions_timeline")), 100)
def test_cache_in_memory_limited_dataset_unreversed(self):
"""Test if the cache is saved properly when loaded in memory in unreversed buffers, when newest items are at the end of the list. """
dataset = dict(home_timeline=[i for i in range(20)], mentions_timeline=[i for i in range(20)])
self.session.settings["general"]["load_cache_in_memory"] = True
self.session.settings["general"]["persist_size"] = 10
self.session.load_persistent_data()
self.assertTrue(len(self.session.db)==1)
self.session.db["home_timeline"] = dataset["home_timeline"]
self.session.db["mentions_timeline"] = dataset["mentions_timeline"]
self.session.save_persistent_data()
self.session.load_persistent_data()
self.assertIsInstance(self.session.db, dict)
self.assertTrue(self.session.db.get("home_timeline") != None)
self.assertTrue(self.session.db.get("mentions_timeline") != None)
self.assertEquals(self.session.db.get("home_timeline")[0], 10)
self.assertEquals(self.session.db.get("mentions_timeline")[0], 10)
self.assertEquals(self.session.db.get("home_timeline")[-1], 19)
self.assertEquals(self.session.db.get("mentions_timeline")[-1], 19)
def test_cache_in_memory_limited_dataset_reversed(self):
"""Test if the cache is saved properly in reversed buffers, when newest items are at the start of the list. This test if for db read into memory. """
dataset = dict(home_timeline=[i for i in range(19, -1, -1)], mentions_timeline=[i for i in range(19, -1, -1)])
self.session.settings["general"]["load_cache_in_memory"] = True
self.session.settings["general"]["persist_size"] = 10
self.session.settings["general"]["reverse_timelines"] = True
self.session.load_persistent_data()
self.session.db["home_timeline"] = dataset["home_timeline"]
self.session.db["mentions_timeline"] = dataset["mentions_timeline"]
self.session.save_persistent_data()
self.session.load_persistent_data()
self.assertIsInstance(self.session.db, dict)
self.assertTrue(self.session.db.get("home_timeline") != None)
self.assertTrue(self.session.db.get("mentions_timeline") != None)
self.assertEquals(self.session.db.get("home_timeline")[0], 19)
self.assertEquals(self.session.db.get("mentions_timeline")[0], 19)
self.assertEquals(self.session.db.get("home_timeline")[-1], 10)
self.assertEquals(self.session.db.get("mentions_timeline")[-1], 10)
if __name__ == "__main__":
unittest.main()