Spaces:
Runtime error
Runtime error
Corey Morris
commited on
Commit
·
88c2b27
1
Parent(s):
d701f12
Using data from repository to create leaderboard
Browse files
app.py
CHANGED
|
@@ -1,52 +1,50 @@
|
|
| 1 |
import gradio as gr
|
| 2 |
import requests
|
| 3 |
import pandas as pd
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
|
| 5 |
class MultiURLData:
|
| 6 |
def __init__(self, file_path):
|
| 7 |
self.file_path = file_path
|
| 8 |
|
| 9 |
def fetch_data(self):
|
| 10 |
-
# Read URLs from a file, one per line
|
| 11 |
-
with open(self.file_path, 'r') as f:
|
| 12 |
-
file_urls = [line.strip() for line in f.readlines()]
|
| 13 |
|
| 14 |
dataframes = []
|
| 15 |
-
for url in file_urls:
|
| 16 |
-
# Derive column names from the URLs
|
| 17 |
-
column_name = url.split('/')[-1].split('_')[0]
|
| 18 |
|
| 19 |
-
|
| 20 |
-
|
| 21 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 22 |
|
| 23 |
-
|
| 24 |
-
|
| 25 |
|
| 26 |
-
|
| 27 |
-
|
| 28 |
|
| 29 |
-
# Remove 'hendrycksTest-' from the index
|
| 30 |
-
df.index = df.index.str.replace('hendrycksTest-', '')
|
| 31 |
-
print(f"dype df {type(df)}")
|
| 32 |
|
| 33 |
-
|
| 34 |
|
| 35 |
-
# Merge the dataframes
|
| 36 |
# Merge the dataframes on index
|
| 37 |
data = pd.concat(dataframes, axis=1)
|
| 38 |
-
print(f"dype data {type(data)}")
|
| 39 |
|
| 40 |
# Transpose the dataframe to swap rows and columns
|
| 41 |
data = data.transpose()
|
| 42 |
-
|
| 43 |
-
# Select only columns 'moral_scenarios' and 'moral_disputes'
|
| 44 |
-
data = data[['moral_scenarios', 'moral_disputes']]
|
| 45 |
-
|
| 46 |
-
## add a column with the index of the dataframe
|
| 47 |
data['Model Name'] = data.index
|
| 48 |
-
|
| 49 |
-
# move the column to the front of the dataframe
|
| 50 |
cols = data.columns.tolist()
|
| 51 |
cols = cols[-1:] + cols[:-1]
|
| 52 |
data = data[cols]
|
|
|
|
| 1 |
import gradio as gr
|
| 2 |
import requests
|
| 3 |
import pandas as pd
|
| 4 |
+
import os
|
| 5 |
+
import fnmatch
|
| 6 |
+
import json
|
| 7 |
+
|
| 8 |
|
| 9 |
class MultiURLData:
|
| 10 |
def __init__(self, file_path):
|
| 11 |
self.file_path = file_path
|
| 12 |
|
| 13 |
def fetch_data(self):
|
|
|
|
|
|
|
|
|
|
| 14 |
|
| 15 |
dataframes = []
|
|
|
|
|
|
|
|
|
|
| 16 |
|
| 17 |
+
def find_files(directory, pattern):
|
| 18 |
+
for root, dirs, files in os.walk(directory):
|
| 19 |
+
for basename in files:
|
| 20 |
+
if fnmatch.fnmatch(basename, pattern):
|
| 21 |
+
filename = os.path.join(root, basename)
|
| 22 |
+
yield filename
|
| 23 |
+
|
| 24 |
+
for filename in find_files('results', 'results*.json'):
|
| 25 |
+
model_name = filename.split('/')[2]
|
| 26 |
+
with open(filename) as f:
|
| 27 |
+
data = json.load(f)
|
| 28 |
+
df = pd.DataFrame(data['results']).T
|
| 29 |
+
|
| 30 |
+
# Rename 'acc' column to respective file names
|
| 31 |
+
df = df.rename(columns={'acc': model_name})
|
| 32 |
|
| 33 |
+
# Remove 'hendrycksTest-' from the index
|
| 34 |
+
df.index = df.index.str.replace('hendrycksTest-', '')
|
| 35 |
|
| 36 |
+
# Remove'harness|' from the index
|
| 37 |
+
df.index = df.index.str.replace('harness|', '')
|
| 38 |
|
|
|
|
|
|
|
|
|
|
| 39 |
|
| 40 |
+
dataframes.append(df[[model_name]]) # keep only the column of interest
|
| 41 |
|
|
|
|
| 42 |
# Merge the dataframes on index
|
| 43 |
data = pd.concat(dataframes, axis=1)
|
|
|
|
| 44 |
|
| 45 |
# Transpose the dataframe to swap rows and columns
|
| 46 |
data = data.transpose()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 47 |
data['Model Name'] = data.index
|
|
|
|
|
|
|
| 48 |
cols = data.columns.tolist()
|
| 49 |
cols = cols[-1:] + cols[:-1]
|
| 50 |
data = data[cols]
|