GSK-2737-GSK-2735-support-model-url-remove-misleading-info (#90)

- add support to input model url & remove suggestions when no related ds (77da4cdd32b75cee859556a213421840cedc380e)


Co-authored-by: zcy <ZeroCommand@users.noreply.huggingface.co>

text_classification.py

@@ -376,4 +376,9 @@ def text_classification_fix_column_mapping(column_mapping, ppl, d_id, config, sp
         prediction_result,
         id2label_df,
         feature_map_df,
-    )
+    )
+
+def strip_model_id_from_url(model_id):
+    if model_id.startswith("https://huggingface.co/"):
+        return "/".join(model_id.split("/")[-2])
+    return model_id

text_classification_ui_helpers.py

@@ -11,6 +11,7 @@ import leaderboard
 from io_utils import read_column_mapping, write_column_mapping
 from run_jobs import save_job_to_pipe
 from text_classification import (
+    strip_model_id_from_url,
     check_model_task,
     preload_hf_inference_api,
     get_example_prediction,
@@ -21,6 +22,7 @@ from wordings import (
     CHECK_CONFIG_OR_SPLIT_RAW,
     CONFIRM_MAPPING_DETAILS_FAIL_RAW,
     MAPPING_STYLED_ERROR_WARNING,
+    NOT_TEXT_CLASSIFICATION_MODEL_RAW,
     get_styled_input,
 )
 
@@ -32,12 +34,12 @@ ds_config = None
 
 def get_related_datasets_from_leaderboard(model_id):
     records = leaderboard.records
+    model_id = strip_model_id_from_url(model_id)
     model_records = records[records["model_id"] == model_id]
     datasets_unique = list(model_records["dataset_id"].unique())
 
     if len(datasets_unique) == 0:
-        all_unique_datasets = list(records["dataset_id"].unique())
-        return gr.update(choices=all_unique_datasets, value="")
+        return gr.update(choices=[], value="")
     
     return gr.update(choices=datasets_unique, value=datasets_unique[0])
 
@@ -161,10 +163,11 @@ def list_labels_and_features_from_dataset(ds_labels, ds_features, model_labels,
 def precheck_model_ds_enable_example_btn(
     model_id, dataset_id, dataset_config, dataset_split
 ):
+    model_id = strip_model_id_from_url(model_id)
     model_task = check_model_task(model_id)
     preload_hf_inference_api(model_id)
     if model_task is None or model_task != "text-classification":
-        gr.Warning("Please check your model.")
+        gr.Warning(NOT_TEXT_CLASSIFICATION_MODEL_RAW)
         return (gr.update(), gr.update(),"")
 
     if dataset_config is None or dataset_split is None or len(dataset_config) == 0:
@@ -195,9 +198,10 @@ def align_columns_and_show_prediction(
     run_inference,
     inference_token,
 ):
+    model_id = strip_model_id_from_url(model_id)
     model_task = check_model_task(model_id)
     if model_task is None or model_task != "text-classification":
-        gr.Warning("Please check your model.")
+        gr.Warning(NOT_TEXT_CLASSIFICATION_MODEL_RAW)
         return (
             gr.update(visible=False),
             gr.update(visible=False),
@@ -338,7 +342,7 @@ def try_submit(m_id, d_id, config, split, inference, inference_token, uid):
         eval_str,
         threading.Lock(),
     )
-    gr.Info("Your evaluation is submitted")
+    gr.Info("Your evaluation has been submitted")
 
     return (
         gr.update(interactive=False),  # Submit button

wordings.py

@@ -38,6 +38,10 @@ MAPPING_STYLED_ERROR_WARNING = """
                         </h3>
                         """
 
+NOT_TEXT_CLASSIFICATION_MODEL_RAW = """
+                        Your model does not fall under the category of text classification. This page is specifically designated for the evaluation of text classification models.  
+                      """
+
 USE_INFERENCE_API_TIP = """
                 We recommend to use
                 <a href="https://huggingface.co/docs/api-inference/detailed_parameters#text-classification-task">