app.py

app.py

@@ -7,8 +7,10 @@ import pandas as pd
 from datasets import Dataset
 from pypdf import PdfReader
 from huggingface_hub import HfApi
+
 # import template
 from string import Template
+
 to_be_removed = ["ͳ", "•", "→", "□", "▪", "►", "�", "", "", "", ""]
 to_be_replaced = {
     "½": "1/2",
@@ -37,8 +39,6 @@ to_be_replaced = {
 }
 
 
-
-
 def clean(text):
     # Remove all the unwanted characters
     for char in to_be_removed:
@@ -113,32 +113,32 @@ caution_text = """⚠️ Caution:
 - Anyone (including you) will be able to delete the dataset once it is uploaded.
 """
 
-instructions_template = Template("""
-Your dataset is now available on Hugging Face Datasets at [pdf2dataset/$dataset_name](https://huggingface.co/datasets/pdf2dataset/$dataset_name).
-
-You can load the dataset using the following code:
+instructions_template = Template(
+    """
+🔗: https://huggingface.co/datasets/pdf2dataset/$dataset_name.
 
 ```python
 from datasets import load_dataset
 
 dataset = load_dataset("pdf2dataset/$dataset_name")
 ```
-    """)
+    """
+)
 
 with gr.Blocks() as demo:
-    # Convert a PDF to a dataset
-    gr.Markdown("## Convert a PDF to a dataset")
+    gr.Markdown("# PDF to 🤗 Dataset")
+    gr.Markdown("## 1️⃣ Upload a PDF")
     file = gr.File(file_types=["pdf"], height=50)
     gr.Markdown(caution_text)
-    convert_button = gr.Button("🔄 Convert and upload") 
-    instructions = gr.Markdown(instructions_template.substitute(dataset_name="generated_dataset_name"))
+    gr.Markdown("## 2️⃣ Convert the PDF and upload")
+    convert_button = gr.Button("🔄 Convert and upload")
     preview = gr.Dataframe(label="Preview (first 10 rows)", headers=["text"], datatype=["str"], row_count=10, wrap=True, height=200)
-    
-    # Delete a dataset
-    gr.Markdown("### Delete a dataset")
-    dataset_name_to_delete = gr.Textbox("", placeholder="Enter dataset name to delete")
+    gr.Markdown("## 3️⃣ Use the dataset in your code")
+    instructions = gr.Markdown(instructions_template.substitute(dataset_name="generated_dataset_name"))
+    gr.Markdown("## 4️⃣ Delete the (optional)")
+    dataset_name_to_delete = gr.Textbox("", placeholder="Enter dataset name to delete", label="Dataset to delete")
     delete_button = gr.Button("🗑️ Delete dataset")
-    
+
     # Define the actions
     convert_button.click(pdf2dataset, inputs=[file], outputs=[instructions, preview, dataset_name_to_delete])
     delete_button.click(delete_dataset, inputs=[dataset_name_to_delete], outputs=[delete_button])