GSK-2844-to-2841-fix-for-model-url-add-doc-and-UI

#120
app_leaderboard.py CHANGED
@@ -90,7 +90,7 @@ def get_demo(leaderboard_tab):
90
  column_names = records.columns.tolist()
91
  issue_columns = column_names[:11]
92
  info_columns = column_names[15:]
93
- default_columns = ["dataset_id", "total_issues", "report_link"]
94
  default_df = records[default_columns] # extract columns selected
95
  types = get_types(default_df)
96
  display_df = get_display_df(default_df) # the styled dataframe to display
 
90
  column_names = records.columns.tolist()
91
  issue_columns = column_names[:11]
92
  info_columns = column_names[15:]
93
+ default_columns = ["model_id", "dataset_id", "total_issues", "report_link"]
94
  default_df = records[default_columns] # extract columns selected
95
  types = get_types(default_df)
96
  display_df = get_display_df(default_df) # the styled dataframe to display
app_text_classification.py CHANGED
@@ -59,7 +59,7 @@ def get_demo():
59
  with gr.Row():
60
  first_line_ds = gr.DataFrame(label="Dataset Preview", visible=False)
61
  with gr.Row():
62
- loading_status = gr.HTML(visible=True)
63
  with gr.Row():
64
  example_btn = gr.Button(
65
  "Validate Model & Dataset",
@@ -67,9 +67,10 @@ def get_demo():
67
  variant="primary",
68
  interactive=False,
69
  )
70
-
71
  with gr.Row():
72
- example_input = gr.HTML(visible=False)
 
 
73
  with gr.Row():
74
  example_prediction = gr.Label(label="Model Sample Prediction", visible=False)
75
 
@@ -153,7 +154,7 @@ def get_demo():
153
  triggers=[dataset_id_input.input, dataset_id_input.select],
154
  fn=check_dataset,
155
  inputs=[dataset_id_input],
156
- outputs=[dataset_config_input, dataset_split_input, loading_status]
157
  )
158
 
159
  dataset_config_input.change(fn=get_dataset_splits, inputs=[dataset_id_input, dataset_config_input], outputs=[dataset_split_input])
@@ -197,7 +198,12 @@ def get_demo():
197
  dataset_config_input,
198
  dataset_split_input,
199
  ],
200
- outputs=[example_btn, first_line_ds, loading_status],
 
 
 
 
 
201
  )
202
 
203
  gr.on(
@@ -215,11 +221,11 @@ def get_demo():
215
  inference_token,
216
  ],
217
  outputs=[
218
- example_input,
219
  example_prediction,
220
  column_mapping_accordion,
221
  run_btn,
222
- loading_status,
223
  *column_mappings,
224
  ],
225
  )
 
59
  with gr.Row():
60
  first_line_ds = gr.DataFrame(label="Dataset Preview", visible=False)
61
  with gr.Row():
62
+ loading_dataset_info = gr.HTML(visible=True)
63
  with gr.Row():
64
  example_btn = gr.Button(
65
  "Validate Model & Dataset",
 
67
  variant="primary",
68
  interactive=False,
69
  )
 
70
  with gr.Row():
71
+ loading_validation = gr.HTML(visible=True)
72
+ with gr.Row():
73
+ validation_result = gr.HTML(visible=False)
74
  with gr.Row():
75
  example_prediction = gr.Label(label="Model Sample Prediction", visible=False)
76
 
 
154
  triggers=[dataset_id_input.input, dataset_id_input.select],
155
  fn=check_dataset,
156
  inputs=[dataset_id_input],
157
+ outputs=[dataset_config_input, dataset_split_input, loading_dataset_info]
158
  )
159
 
160
  dataset_config_input.change(fn=get_dataset_splits, inputs=[dataset_id_input, dataset_config_input], outputs=[dataset_split_input])
 
198
  dataset_config_input,
199
  dataset_split_input,
200
  ],
201
+ outputs=[
202
+ example_btn,
203
+ first_line_ds,
204
+ validation_result,
205
+ example_prediction,
206
+ column_mapping_accordion,],
207
  )
208
 
209
  gr.on(
 
221
  inference_token,
222
  ],
223
  outputs=[
224
+ validation_result,
225
  example_prediction,
226
  column_mapping_accordion,
227
  run_btn,
228
+ loading_validation,
229
  *column_mappings,
230
  ],
231
  )
text_classification.py CHANGED
@@ -380,7 +380,7 @@ def text_classification_fix_column_mapping(column_mapping, ppl, d_id, config, sp
380
 
381
  def strip_model_id_from_url(model_id):
382
  if model_id.startswith("https://huggingface.co/"):
383
- return "/".join(model_id.split("/")[-2])
384
  return model_id
385
 
386
  def check_hf_token_validity(hf_token):
@@ -393,21 +393,4 @@ def check_hf_token_validity(hf_token):
393
  response = requests.get(AUTH_CHECK_URL, headers=headers)
394
  if response.status_code != 200:
395
  return False
396
- return True
397
-
398
- def get_dataset_info_from_server(dataset_id):
399
- url = "https://datasets-server.huggingface.co/splits?dataset=" + dataset_id
400
- response = requests.get(url)
401
- if response.status_code != 200:
402
- return None
403
- return response.json()
404
-
405
- def get_dataset_splits(dataset_id, dataset_config):
406
- dataset_info = get_dataset_info_from_server(dataset_id)
407
- if dataset_info is None:
408
- return None
409
- try:
410
- splits = dataset_info["splits"]
411
- return [split["split"] for split in splits if split["config"] == dataset_config]
412
- except Exception:
413
- return None
 
380
 
381
  def strip_model_id_from_url(model_id):
382
  if model_id.startswith("https://huggingface.co/"):
383
+ return "/".join(model_id.split("/")[-2:])
384
  return model_id
385
 
386
  def check_hf_token_validity(hf_token):
 
393
  response = requests.get(AUTH_CHECK_URL, headers=headers)
394
  if response.status_code != 200:
395
  return False
396
+ return True
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
text_classification_ui_helpers.py CHANGED
@@ -179,29 +179,57 @@ def precheck_model_ds_enable_example_btn(
179
  model_task = check_model_task(model_id)
180
  preload_hf_inference_api(model_id)
181
 
182
- if model_task is None or model_task != "text-classification":
183
- gr.Warning(NOT_TEXT_CLASSIFICATION_MODEL_RAW)
184
- return (gr.update(interactive=False), gr.update(visible=False),"")
185
-
186
  if dataset_config is None or dataset_split is None or len(dataset_config) == 0:
187
- return (gr.update(interactive=False), gr.update(visible=False), "")
188
-
 
 
 
 
 
 
189
  try:
190
  ds = datasets.load_dataset(dataset_id, dataset_config, trust_remote_code=True)
191
  df: pd.DataFrame = ds[dataset_split].to_pandas().head(5)
192
  ds_labels, ds_features = get_labels_and_features_from_dataset(ds[dataset_split])
193
- if model_id == "" or model_id is None:
194
- return (gr.update(interactive=False), gr.update(value=df, visible=True), "")
 
 
 
 
 
 
 
 
195
 
196
  if not isinstance(ds_labels, list) or not isinstance(ds_features, list):
197
  gr.Warning(CHECK_CONFIG_OR_SPLIT_RAW)
198
- return (gr.update(interactive=False), gr.update(value=df, visible=True), "")
 
 
 
 
 
 
199
 
200
- return (gr.update(interactive=True), gr.update(value=df, visible=True), "")
 
 
 
 
 
 
201
  except Exception as e:
202
  # Config or split wrong
203
  logger.warn(f"Check your dataset {dataset_id} and config {dataset_config} on split {dataset_split}: {e}")
204
- return (gr.update(interactive=False), gr.update(visible=False), "")
 
 
 
 
 
 
205
 
206
 
207
  def align_columns_and_show_prediction(
 
179
  model_task = check_model_task(model_id)
180
  preload_hf_inference_api(model_id)
181
 
 
 
 
 
182
  if dataset_config is None or dataset_split is None or len(dataset_config) == 0:
183
+ return (
184
+ gr.update(interactive=False),
185
+ gr.update(visible=False),
186
+ gr.update(visible=False),
187
+ gr.update(visible=False),
188
+ gr.update(visible=False),
189
+ )
190
+
191
  try:
192
  ds = datasets.load_dataset(dataset_id, dataset_config, trust_remote_code=True)
193
  df: pd.DataFrame = ds[dataset_split].to_pandas().head(5)
194
  ds_labels, ds_features = get_labels_and_features_from_dataset(ds[dataset_split])
195
+
196
+ if model_task is None or model_task != "text-classification":
197
+ gr.Warning(NOT_TEXT_CLASSIFICATION_MODEL_RAW)
198
+ return (
199
+ gr.update(interactive=False),
200
+ gr.update(value=df, visible=True),
201
+ gr.update(visible=False),
202
+ gr.update(visible=False),
203
+ gr.update(visible=False),
204
+ )
205
 
206
  if not isinstance(ds_labels, list) or not isinstance(ds_features, list):
207
  gr.Warning(CHECK_CONFIG_OR_SPLIT_RAW)
208
+ return (
209
+ gr.update(interactive=False),
210
+ gr.update(value=df, visible=True),
211
+ gr.update(visible=False),
212
+ gr.update(visible=False),
213
+ gr.update(visible=False),
214
+ )
215
 
216
+ return (
217
+ gr.update(interactive=True),
218
+ gr.update(value=df, visible=True),
219
+ gr.update(visible=False),
220
+ gr.update(visible=False),
221
+ gr.update(visible=False),
222
+ )
223
  except Exception as e:
224
  # Config or split wrong
225
  logger.warn(f"Check your dataset {dataset_id} and config {dataset_config} on split {dataset_split}: {e}")
226
+ return (
227
+ gr.update(interactive=False),
228
+ gr.update(visible=False),
229
+ gr.update(visible=False),
230
+ gr.update(visible=False),
231
+ gr.update(visible=False),
232
+ )
233
 
234
 
235
  def align_columns_and_show_prediction(
wordings.py CHANGED
@@ -2,7 +2,8 @@ INTRODUCTION_MD = """
2
  <h1 style="text-align: center;">
3
  🐢Giskard Evaluator - Text Classification
4
  </h1>
5
- Welcome to the Giskard Evaluator Space! Get a model vulnerability report immediately by simply sharing your model and dataset id below.
 
6
  """
7
  CONFIRM_MAPPING_DETAILS_MD = """
8
  <h1 style="text-align: center;">
 
2
  <h1 style="text-align: center;">
3
  🐢Giskard Evaluator - Text Classification
4
  </h1>
5
+ Welcome to the Giskard Evaluator Space! Get a model vulnerability report immediately by simply sharing your model and dataset id below.
6
+ You can also checkout our library documentation <a href="https://docs.giskard.ai/en/latest/getting_started/quickstart/index.html">here</a>.
7
  """
8
  CONFIRM_MAPPING_DETAILS_MD = """
9
  <h1 style="text-align: center;">