Improve dataset card: Add metadata, links to paper, project, code, and usage guidance

This PR enhances the `ClueWeb-Reco` dataset card by:
- Adding `task_categories: ['text-retrieval', 'recommendation']` and `tags: ['recommender-systems', 'webpage-recommendation']` to the metadata for improved discoverability.
- Including direct links to the associated paper (https://huggingface.co/papers/2510.26095), project page (https://www.open-reco-bench.ai), and GitHub repository (https://github.com/cxcscmu/RecSys-Benchmark) in the content.
- Restructuring the introductory text for better clarity and context.
- Converting the "Utility files" section into a "Sample Usage" guide, explaining how to use the provided `ClueWeb22Api.py` and `example_dataset.py` for data loading without fabricating code snippets.
- Removing a commented-out YAML block from the Markdown content.

These changes provide more comprehensive information about the dataset and its usage, aligning with Hugging Face Hub best practices.

README.md

@@ -1,48 +1,36 @@
 ---
+license: mit
+task_categories:
+- text-retrieval
+- recommendation
+tags:
+- recommender-systems
+- webpage-recommendation
 configs:
 - config_name: input
   data_files:
   - split: valid
     path:
-    - "interaction_splits/valid_inter_input.tsv"
+    - interaction_splits/valid_inter_input.tsv
   - split: test
-    path: 
-    - "interaction_splits/test_inter_input.tsv"
+    path:
+    - interaction_splits/test_inter_input.tsv
   default: true
-- config_name: target 
+- config_name: target
   data_files:
   - split: valid
     path:
-    - "interaction_splits/valid_inter_target.tsv"
-- config_name: mapping 
-  data_files: "cwid_to_id.tsv"
-license: mit
+    - interaction_splits/valid_inter_target.tsv
+- config_name: mapping
+  data_files: cwid_to_id.tsv
 ---
 
+# ClueWeb-Reco Dataset
 
-<!-- configs:
-- config_name: ordered_id_splits
-  data_files:
-  - split: train
-    path:
-    - "ordered_id_splits/valid_input.tsv"
-    - "ordered_id_splits/valid_target.tsv"
-  - split: test
-    path: "ordered_id_splits/test_input.tsv"
-  default: true
-- config_name: interaction_splits
-  data_files:
-  - split: valid
-    path:
-    - "interaction_splits/valid_inter_input.tsv"
-    - "interaction_splits/valid_inter_target.csv"
-  - split: test
-    path: "interaction_splits/test_input.tsv"
-- config_name: clueweb_id_mapping
-  data_files: "cwid_to_id.tsv" -->
-
+This dataset is part of the [ORBIT - Open Recommendation Benchmark for Reproducible Research with Hidden Tests](https://huggingface.co/papers/2510.26095) paper.
 
-# ClueWeb-Reco: 
+Project page: https://www.open-reco-bench.ai
+Code: https://github.com/cxcscmu/RecSys-Benchmark
 
 ## ClueWeb-Reco as Webpage Recommendation Hidden Test 
 ClueWeb-Reco is a dataset constructed by mapping real-life U.S. browsing history to publicly available websites in the English subset of [ClueWeb22-B](https://lemurproject.org/clueweb22/) dataset.
@@ -75,10 +63,12 @@ We provide input and ground truth for the validation set. The target of the test
     - valid_target.tsv: validation dataset ground truth
     - test_input.tsv: input for testing dataset (ground truth hidden) 
 
-## Utility files for ClueWebApi usage and example processing on the ordered ClueWeb id list format  
-- cw_data_processing: 
-    - ClueWeb22Api.py: API to retrieve ClueWeb document information from official ClueWeb22 docids
-    - example_dataset.py: example to load input data sequences with ClueWeb22Api
+## Sample Usage
+The repository provides utility files in the `cw_data_processing` folder for interacting with the ClueWeb-Reco dataset:
+- `ClueWeb22Api.py`: An API to retrieve ClueWeb document information from official ClueWeb22 docids.
+- `example_dataset.py`: An example script to load input data sequences using the `ClueWeb22Api`.
+
+You can refer to these files within the GitHub repository (linked above) for detailed examples on how to load and process the dataset.
 
 ## Note
-The ClueWeb-Reco dataset was collected, stored, released, and is maintained by our team at Carnegie Mellon University.
+The ClueWeb-Reco dataset was collected, stored, released, and is maintained by our team at Carnegie Mellon University.