update

README.md

@@ -39,7 +39,7 @@ task_categories:
 
 The UTS_Text_v1 dataset is a collection of 10,000 sentences sourced from various news articles.
 
-Out of the 10,000 sentences in the dataset, 5,000 sentences have a length greater than 50, while the other 5,000 sentences have a length ranging from 20 to 50. This distribution of sentence lengths provides a diverse range of text samples that can be used to train and test natural language processing models.
+Out of the 10,000 sentences in the dataset, 5,000 sentences have a length ranging from 50 to 150, while the other 5,000 sentences have a length ranging from 20 to 50. This distribution of sentence lengths provides a diverse range of text samples that can be used to train and test natural language processing models.
 
 ### Dataset Summary
 

eval.py

@@ -0,0 +1,18 @@
+from datasets import load_dataset
+import matplotlib.pyplot as plt
+
+dataset = load_dataset("undertheseanlp/UTS_Text_v1")
+sentences = dataset["train"]["text"]
+
+# compute histogram of sentence lengths with bin size = 10
+lengths = [len(s) for s in sentences]
+plt.hist(lengths, bins=range(0, max(lengths), 10))
+plt.show()
+
+# n_sample = 0
+# for s in sentences:
+#     if len(s) > 150:
+#         print(s)
+#         n_sample += 1
+#     if n_sample == 10:
+#         break

generate_dataset.py

@@ -3,17 +3,23 @@ from underthesea.file_utils import DATASETS_FOLDER
 import random
 
 random.seed(10)
-# sampling data
 text_file = join(DATASETS_FOLDER, "VNESES", "VNESEScorpus.txt")
 with open(text_file) as f:
     lines = f.read().splitlines()
 NUM_LONG_TOKENS = 50
 NUM_SHORT_TOKENS = 20
-long_lines = [
-    line
-    for line in lines
-    if len(line) >= NUM_LONG_TOKENS and line[0].isupper() and line[-1] == "."
-]
+NUM_MAX_TOKENS = 150
+
+
+def longline_conditions(line):
+    if len(line) < NUM_LONG_TOKENS or len(line) > NUM_MAX_TOKENS:
+        return False
+    if not (line[0].isupper() and line[-1] == "."):
+        return False
+    return True
+
+
+long_lines = [line for line in lines if longline_conditions(line)]
 # get random 1000 lines
 random_long_lines = random.sample(long_lines, 5000)
 for line in random_long_lines[:20]: