app: rework of leaderboard (now show tabs for each dataset, improve description)

app.py

@@ -2,18 +2,42 @@ import gradio as gr
 import pandas as pd
 
 title = """
-# hmLeaderboard
+# hmLeaderboard: Space for tracking and ranking models on Historic NER Datasets
 
 ![hmLeaderboard](https://huggingface.co/spaces/hmbench/hmLeaderboard/resolve/main/logo.png)
 """
 
 description = """
-## Space for tracking and ranking models on Historic NER Datasets.
+## Models
 
 At the moment the following models are supported:
 
 * hmBERT: [Historical Multilingual Language Models for Named Entity Recognition](https://huggingface.co/hmbert).
 * hmTEAMS: [Historic Multilingual TEAMS Models](https://huggingface.co/hmteams).
+
+## Datasets
+
+We test our pretrained language models on various datasets from HIPE-2020, HIPE-2022 and Europeana. The following table
+shows an overview of used datasets.
+
+
+| Language | Datasets                                                         |
+|----------|------------------------------------------------------------------|
+| English  | [AjMC] - [TopRes19th]                                            |
+| German   | [AjMC] - [NewsEye] - [HIPE-2020]                                 |
+| French   | [AjMC] - [ICDAR-Europeana] - [LeTemps] - [NewsEye] - [HIPE-2020] |
+| Finnish  | [NewsEye]                                                        |
+| Swedish  | [NewsEye]                                                        |
+| Dutch    | [ICDAR-Europeana]                                                |
+
+[AjMC]: https://github.com/hipe-eval/HIPE-2022-data/blob/main/documentation/README-ajmc.md
+[NewsEye]: https://github.com/hipe-eval/HIPE-2022-data/blob/main/documentation/README-newseye.md
+[TopRes19th]: https://github.com/hipe-eval/HIPE-2022-data/blob/main/documentation/README-topres19th.md
+[ICDAR-Europeana]: https://github.com/stefan-it/historic-domain-adaptation-icdar
+[LeTemps]: https://github.com/hipe-eval/HIPE-2022-data/blob/main/documentation/README-letemps.md
+[HIPE-2020]: https://github.com/hipe-eval/HIPE-2022-data/blob/main/documentation/README-hipe2020.md
+
+## Results
 """
 footer = "Made from Bavarian Oberland with ❤️ and 🥨."
 
@@ -63,23 +87,51 @@ def perform_evaluation_for_languages(model_selection, selected_languages):
 
     return result_df
 
+dataset_to_description_mapping = {
+    "AjMC": "#### AjMC\nThe AjMC dataset consists of NE-annotated historical commentaries in the field of Classics, and was created in the context of the [Ajax MultiCommentary](https://mromanello.github.io/ajax-multi-commentary/) project.\n\nThe following NEs were annotated: `pers`, `work`, `loc`, `object`, `date` and `scope`.",
+    "NewsEye": "#### NewsEye\nThe NewsEye dataset is comprised of diachronic historical newspaper material published between 1850 and 1950 in French, German, Finnish, and Swedish. More information can be found [here](https://dl.acm.org/doi/abs/10.1145/3404835.3463255).\n\nThe following NEs were annotated: `PER`, `LOC`, `ORG` and `HumanProd`.",
+    "ICDAR": "#### ICDAR\nThe ICDAR-Europeana NER Dataset is a preprocessed variant of the [Europeana NER Corpora](https://github.com/EuropeanaNewspapers/ner-corpora) for Dutch and French.\n\nThe following NEs were annotated: `PER`, `LOC` and `ORG`.",
+    "LeTemps": "#### LeTemps\nThe LeTemps dataset consists of NE-annotated historical French newspaper articles from mid-19C to mid 20C.\n\nThe following NEs were annotated: `loc`, `org` and `pers`.",
+    "TopRes19th": "#### TopRes19th\nThe TopRes19th dataset consists of NE-annotated historical English newspaper articles from 19C.\n\nThe following NEs were annotated: `BUILDING`, `LOC` and `STREET`.",
+    "HIPE-2020": "#### HIPE-2020\nThe HIPE-2020 dataset is comprised of newspapers from mid 19C to mid 20C. For information can be found [here](https://dl.acm.org/doi/abs/10.1007/978-3-030-58219-7_21).\n\nThe following NEs were annotated: `loc`, `org`, `pers`, `prod`, `time` and `comp`.",
+}
+
+configuration_to_description_mapping = {
+    "Best Configuration": "The best hyper-parameter configuration for each model is used and average F1-score over runs with different seeds is reported here:",
+    "Best Model": "The best hyper-parameter configuration for each model is used, the model with highest F1-score is chosen and its performance is reported here:"
+}
+
 with gr.Blocks() as demo:
     gr.Markdown(title)
     gr.Markdown(description)
 
     with gr.Tab("Overview"):
-        gr.Markdown("### Best Configuration\nThe best hyper-parameter configuration for each model is used and average F1-score over runs with different seeds is reported here:")
+        gr.Markdown("### Best Configuration")
+        gr.Markdown(configuration_to_description_mapping["Best Configuration"])
 
         df_result = perform_evaluation_for_datasets("Best Configuration", dataset_names)
 
         gr.Dataframe(value=df_result)
 
-        gr.Markdown("### Best Model\nThe best hyper-parameter configuration for each model is used, the model with highest F1-score is chosen and its performance is reported here:")
+        gr.Markdown("### Best Model")
+        gr.Markdown(configuration_to_description_mapping["Best Model"])
 
         df_result = perform_evaluation_for_datasets("Best Model", dataset_names)
 
         gr.Dataframe(value=df_result)
 
+    for dataset_name, dataset_description in dataset_to_description_mapping.items():
+        with gr.Tab(dataset_name):
+            selected_datasets = [ds for ds in dataset_names if dataset_name.lower() in ds.lower()]
+
+            gr.Markdown(dataset_description)
+
+            for config in ["Best Configuration", "Best Model"]:
+                gr.Markdown(f"##### Results for {config}")
+                gr.Markdown(configuration_to_description_mapping[config])
+                df_result = perform_evaluation_for_datasets(config, selected_datasets)
+                gr.Dataframe(value=df_result)
+
     with gr.Tab("Filtering"):
 
         gr.Markdown("### Filtering\nSwiss-knife filtering for single datasets and languages is possible.")