Fix: update data file paths to data/raw-*.parquet and data/deduplicated-*.parquet

README.md

@@ -1,33 +1,13 @@
 ---
 configs:
-- config_name: deduplicated
-  data_files:
-  - split: deduplicated
-    path: deduplicated/*.parquet
-- config_name: default
-  data_files:
-  - split: raw
-    path: data/raw-*
 - config_name: raw
   data_files:
   - split: raw
-    path: raw/*.parquet
-dataset_info:
-  features:
-  - name: name
-    dtype: string
-  - name: version
-    dtype: string
-  - name: capabilities
-    dtype: string
-  - name: last_seen
-    dtype: string
-  splits:
-  - name: raw
-    num_bytes: 3979
-    num_examples: 45
-  download_size: 3865
-  dataset_size: 3979
+    path: data/raw-*.parquet
+- config_name: deduplicated
+  data_files:
+  - split: deduplicated
+    path: data/deduplicated-*.parquet
 ---
 
 # MCP Clients Dataset
@@ -84,7 +64,7 @@ The dataset is kept up-to-date via Hugging Face Jobs:
 
 2. **Deduplicate step** (`deduplicated` split):
    ```bash
-   uv run https://huggingface.co/datasets/evalstate/mcp-clients/resolve/main/deduplicate_clients.py --push-to-hub --split deduplicated
+   uv run https://huggingface.co/datasets/evalstate/mcp-clients/resolve/main/deduplicate_clients.py --push-to-hub
    ```
 
 ## License