add dataset

README.md

@@ -27,3 +27,11 @@ configs:
 ---
 
 # XBRL Extraction Dataset
+
+The is the official dataset introduced in the paper [FinLoRA: Benchmarking LoRA Methods for Fine-Tuning LLMs on Financial Datasets](https://arxiv.org/abs/2505.19819)
+
+<p>
+  <a href="https://huggingface.co/spaces/wangd12/xbrl_llm_demo"><img src="https://raw.githubusercontent.com/wangd12rpi/FinLoRA/main/static/demo_btn.svg"></a>
+  <a href="https://huggingface.co/spaces/wangd12/xbrl_llm_demo"><img src="https://raw.githubusercontent.com/wangd12rpi/FinLoRA/main/static/models_btn.svg"></a>
+  <a href="https://finlora-docs.readthedocs.io/en/latest/"><img src="https://raw.githubusercontent.com/wangd12rpi/FinLoRA/main/static/doc_btn.svg"></a>
+</p>

generate_xbrl_extract_hf_split.py

@@ -100,14 +100,14 @@ def gen_xbrl(cat):
 
         print(f"train size: {len(train_data)}, test size: {len(test_data)}\n")
 
-        with open(f"{cat}_test.csv", "w", newline="") as f:
-            w = csv.DictWriter(f, test_data[0].keys(), quoting=csv.QUOTE_ALL)
-            w.writeheader()
-            w.writerows(test_data)
-        with open(f"{cat}_train.csv", "w", newline="") as f:
-            w = csv.DictWriter(f, train_data[0].keys(), quoting=csv.QUOTE_ALL)
-            w.writeheader()
-            w.writerows(train_data)
+        # with open(f"{cat}_test.csv", "w", newline="") as f:
+        #     w = csv.DictWriter(f, test_data[0].keys(), quoting=csv.QUOTE_ALL)
+        #     w.writeheader()
+        #     w.writerows(test_data)
+        # with open(f"{cat}_train.csv", "w", newline="") as f:
+        #     w = csv.DictWriter(f, train_data[0].keys(), quoting=csv.QUOTE_ALL)
+        #     w.writeheader()
+        #     w.writerows(train_data)
 
         return train_data, test_data