Clean up files and Syntax

.gitattributes

@@ -1 +0,0 @@
-data/wandb/tzebw6rb.parquet filter=lfs diff=lfs merge=lfs -text

test.py

@@ -1,17 +0,0 @@
-import pytest
-
-
-def test_query_network():
-    pass
-
-
-def test_filter_completions():
-    pass
-
-
-def test_guess_task_name():
-    pass
-
-
-def test_ensemble_completions():
-    pass

utils.py

@@ -52,12 +52,12 @@ EXTRACTORS = {
     'created_at': lambda x: pd.Timestamp(x.created_at),
     'last_event_at': lambda x: pd.Timestamp(x.summary.get('_timestamp'), unit='s'),
 
-    'netuid': lambda x: x.config.get('netuid'),
-    'mock': lambda x: x.config.get('neuron').get('mock'),
-    'sample_size': lambda x: x.config.get('neuron').get('sample_size'),
-    'timeout': lambda x: x.config.get('neuron').get('timeout'),
-    'epoch_length': lambda x: x.config.get('neuron').get('epoch_length'),
-    'disable_set_weights': lambda x: x.config.get('neuron').get('disable_set_weights'),
+    # 'netuid': lambda x: x.config.get('netuid'),
+    # 'mock': lambda x: x.config.get('neuron').get('mock'),
+    # 'sample_size': lambda x: x.config.get('neuron').get('sample_size'),
+    # 'timeout': lambda x: x.config.get('neuron').get('timeout'),
+    # 'epoch_length': lambda x: x.config.get('neuron').get('epoch_length'),
+    # 'disable_set_weights': lambda x: x.config.get('neuron').get('disable_set_weights'),
 
     # This stuff is from the last logged event
     'num_steps': lambda x: x.summary.get('_step'),
@@ -176,8 +176,9 @@ def build_data(timestamp=None, path=BASE_PATH, min_steps=MIN_STEPS, use_cache=Tr
         n_events += num_steps
         prog_msg = f'Loading data {i/len(runs)*100:.0f}%, (total {n_events:,.0f} events)'
         progress.progress(i/len(runs),text=f'{prog_msg}... **downloading** `{os.path.join(*run.path)}`')
+        if 'netuid_1' in run.tags or 'netuid_61' in run.tags or 'netuid_102' in run.tags:
+            run_data.append(run)
 
-        run_data.append(run)
 
     progress.empty()
 
@@ -249,7 +250,7 @@ def get_productivity(df_runs):
 
     total_duration = df_runs.last_event_at.max() - df_runs.created_at.min()
     total_steps = df_runs.num_steps.sum()
-    total_completions = (df_runs.num_steps*df_runs.sample_size).sum()
+    total_completions = (df_runs.num_steps*100).sum() #TODO: Parse from df 
     total_completion_words = (df_runs.num_steps*df_runs.completion_words).sum()
     total_completion_tokens = round(total_completion_words/0.75)
     total_validator_words = (df_runs.num_steps*df_runs.apply(lambda x: len(str(x.query).split()) + len(str(x.challenge).split()) + len(str(x.reference).split()), axis=1 )).sum()
@@ -266,7 +267,7 @@ def get_productivity(df_runs):
     }
 
 @st.cache_data(show_spinner=False)
-def get_reward_stats(df, exclude_multiturn=True, freq='1D', remove_zero_rewards=True, agg='mean', date_min='2024-01-22', date_max='2024-06-25'):
+def get_reward_stats(df, exclude_multiturn=True, freq='D', remove_zero_rewards=True, agg='mean', date_min='2024-01-22', date_max='2024-08-12'): #TODO: Set the date_max to the current date
 
     df = df.loc[df._timestamp.between(pd.Timestamp(date_min), pd.Timestamp(date_max))]
     if exclude_multiturn:
@@ -378,7 +379,7 @@ def load_state_vars(username=USERNAME, percentile=0.95):
 
     df_runs = build_data(time.time()//UPDATE_INTERVAL, use_cache=False)
     
-    df_runs = df_runs.loc[df_runs.netuid.isin([1,61,102])]
+    # df_runs = df_runs.loc[df_runs.netuid.isin([1,61,102])] # Now we filter for the netuid tag in build_data
     st.toast(f'Loaded {len(df_runs)} runs')
 
     df_vali = df_runs.loc[df_runs.username == username]