fixed unit test typo

README.md

@@ -5,12 +5,12 @@ This is a work-in-progress package that retrieves training data from the [liches
 Currently the app is not functional, and has not been deployed. If cloning this repo for personal use, the structure of the python scripts assumes that there is a folder called `lichess-games-database` to which .pgn and .pgn.zst files are downloaded and unzipped (this may be automated in the future using a bash script), and that there is a folder called `lichess_player_data` to which .csv files are saved (this folder is created by `parse_pgn.py` if it doesn't exist).
 
 ### Model Description
-This is a simple statistical model that flags players who have performed a certain threshold above their expected performance under the Glicko-2 rating system. The expected performance takes into account all player's complete game history and opponents in the span of the training data. The thresholds are initialized to default values, but are then adjusted separately for each 100 point rating bin in the training data.
+This is a simple statistical model that flags players who have performed a certain threshold above their expected performance under the Glicko-2 rating system. The expected performance takes into account each player's complete game history and opponents in the span of the training data. The thresholds are initialized to default values, and then adjusted separately for each 100 point rating bin in the training data.
 
 ### Model Training
-We define `N` as the number of players who have performed above some threshold, and the estimated number of cheaters as `X = 0.00 * N_open + 0.75 * N_closed + 1.00 * N_violation` where `N_open` is the number of players with open accounts, `N_closed` is the number of players with closed accounts, and `N_violation` is the number of players with a terms of service violation (where `N = N_open + N_closed + N_violation`), the metric used to evaluate the performance of the threshold is the `log(N+1) * X / N`. This is a simple metric intended to reward the model for `high accuracy = X / N` in detecting suspicious players without flagging too many players (observationally, if the threshold is too low, the accuracy will decrease faster than log(N)). Note that for a threshold that is too high and flags 0 players, the metric will be 0. This metric may be fine-tuned in the future, but is sufficient for a POC.
+We define `N` as the number of players who have performed above some threshold, and the estimated number of cheaters as `X = 0.00 * N_open + 0.75 * N_closed + 1.00 * N_violation` where `N_open` is the number of players with open accounts, `N_closed` is the number of players with closed accounts, and `N_violation` is the number of players with a terms of service violation (where `N = N_open + N_closed + N_violation`), the metric used to evaluate the performance of the threshold is the `log(N+1) * X / N`. This is a simple metric intended to reward the model for `high accuracy = X / N` in detecting suspicious players without flagging too many players (observationally, if the threshold is too low, the accuracy will decrease faster than `log(N)`). Note that for a threshold that is too high and flags 0 players, the metric will be 0. This metric may be fine-tuned in the future, but is sufficient for a POC.
 
-Below is an example of the threshold vs accuracy plot below for players in the 1200-1300 range based on training data from the month of Jan 2015.
+Below is an example of the threshold vs accuracy plot below for players in the 1400-1500 range for classical chess based on training data from the month of Jan 2015.
 
 ![sample threshold vs accuracy plot](images/sample_model_threshold.png)
 

requirements.txt

@@ -8,6 +8,5 @@ python-lichess==0.10 # Client for lichess.org API
 mock==5.1.0
 pylint==3.0.3
 pytest==7.0.1
-pytest-dependency==0.6.0
 
 debugpy # Required for debugging.

tests/test_model.py

@@ -72,7 +72,7 @@ def get_sample_test_data():
 
 
 @pytest.mark.usefixtures("get_sample_train_data", "build_training_data")
-@pytest.mark.usefixtures("get_sample_test_data", "build_training_data")
+@pytest.mark.usefixtures("get_sample_test_data", "build_test_data")
 class TestPlayerAnomalyDetectionModel(unittest.TestCase):
     @mock.patch(
         "player_account_handler.PlayerAccountHandler.update_player_account_status"