Corey Morris commited on
Commit
e79bcf3
1 Parent(s): d506f10

Fixed type error

Browse files
Files changed (2) hide show
  1. app.py +1 -1
  2. result_data_processor.py +6 -5
app.py CHANGED
@@ -105,7 +105,7 @@ def create_line_chart(df, model_names, metrics):
105
  fig.update_layout(showlegend=True)
106
  return fig
107
 
108
- def find_top_differences_table(df, target_model, closest_models, num_differences=10, exclude_columns=['Parameters']):
109
  # Calculate the absolute differences for each task between the target model and the closest models
110
  new_df = df.drop(columns=exclude_columns)
111
  differences = new_df.loc[closest_models].sub(new_df.loc[target_model]).abs()
 
105
  fig.update_layout(showlegend=True)
106
  return fig
107
 
108
+ def find_top_differences_table(df, target_model, closest_models, num_differences=10, exclude_columns=['Parameters', 'organization']):
109
  # Calculate the absolute differences for each task between the target model and the closest models
110
  new_df = df.drop(columns=exclude_columns)
111
  differences = new_df.loc[closest_models].sub(new_df.loc[target_model]).abs()
result_data_processor.py CHANGED
@@ -89,6 +89,7 @@ class ResultDataProcessor:
89
  def process_data(self):
90
 
91
  dataframes = []
 
92
  for filename in self._find_files(self.directory, self.pattern):
93
  raw_data = self._read_and_transform_data(filename)
94
  split_path = filename.split('/')
@@ -99,13 +100,15 @@ class ResultDataProcessor:
99
  mc2 = self._extract_mc2(raw_data, model_name)
100
  cleaned_data = pd.concat([cleaned_data, mc1])
101
  cleaned_data = pd.concat([cleaned_data, mc2])
102
- # add organization name to the dataframe as a new row
103
- cleaned_data.loc['organization'] = organization_name
104
  dataframes.append(cleaned_data)
105
 
106
 
107
  data = pd.concat(dataframes, axis=1).transpose()
108
-
 
 
 
109
  # Add Model Name and rearrange columns
110
  data['Model Name'] = data.index
111
  cols = data.columns.tolist()
@@ -137,8 +140,6 @@ class ResultDataProcessor:
137
  # remove extreme outliers from column harness|truthfulqa:mc1
138
  data = self._remove_mc1_outliers(data)
139
 
140
- data = data.drop(columns=['organization'])
141
-
142
  return data
143
 
144
  def rank_data(self):
 
89
  def process_data(self):
90
 
91
  dataframes = []
92
+ organization_names = []
93
  for filename in self._find_files(self.directory, self.pattern):
94
  raw_data = self._read_and_transform_data(filename)
95
  split_path = filename.split('/')
 
100
  mc2 = self._extract_mc2(raw_data, model_name)
101
  cleaned_data = pd.concat([cleaned_data, mc1])
102
  cleaned_data = pd.concat([cleaned_data, mc2])
103
+ organization_names.append(organization_name)
 
104
  dataframes.append(cleaned_data)
105
 
106
 
107
  data = pd.concat(dataframes, axis=1).transpose()
108
+
109
+ # Add organization column
110
+ data['organization'] = organization_names
111
+
112
  # Add Model Name and rearrange columns
113
  data['Model Name'] = data.index
114
  cols = data.columns.tolist()
 
140
  # remove extreme outliers from column harness|truthfulqa:mc1
141
  data = self._remove_mc1_outliers(data)
142
 
 
 
143
  return data
144
 
145
  def rank_data(self):