markdown edits

app.py

@@ -180,16 +180,18 @@ glycosyltransferase_db = {
 
 def get_family_info(family_name):
     family_info = glycosyltransferase_db.get(family_name, {})
-    # convert information to markdown formatted string
-    markdown_text = ""
+    
+    output = ""
     for key, value in family_info.items():
         if key == "more_info":
-            markdown_text += "**{}:**".format(key.title().replace("_", " ")) + "\n"
+            output += "**{}:**".format(key.title().replace("_", " ")) + "\n"
             for link in value:
-                markdown_text += "[{}]({})\n".format(link, link)
+                output += "[{}]({})  ".format(link, link)
         else:
-            markdown_text += "**{}:** {}\n".format(key.title().replace("_", " "), value)
-    return markdown_text
+            output += "**{}:** {}  ".format(key.title().replace("_", " "), value)
+    
+    return output
+
 
 def fig_to_img(fig):
     """Converts a matplotlib figure to a PIL Image and returns it"""
@@ -363,7 +365,7 @@ def generate_heatmap(protein_fasta):
 
     headers = [line for line in lines if line.startswith('>')]
     if len(headers) > 1:
-        return None, "Multiple fasta sequences detected. Please upload a fasta file with multiple sequences, otherwise only include one fasta sequence.", None
+        return None, "Multiple fasta sequences detected. Please only include a single fasta sequence.", None
 
     protein_sequence = ''.join(line for line in lines if not line.startswith('>'))
 
@@ -378,12 +380,14 @@ def generate_heatmap(protein_fasta):
     with torch.no_grad():
         original_output = model(encoded_input["input_ids"], attention_mask=encoded_input["attention_mask"])
     original_probabilities = F.softmax(original_output.logits, dim=1).cpu().numpy()[0]
-
+    
     # Define the size of each group
-    group_size = 10  # You can change this
+    group_size = 10  # allow user to change this
 
-    # Initialize an array to hold the importance scores
+    # Calculate the number of groups
     num_groups = len(protein_sequence) // group_size + (len(protein_sequence) % group_size > 0)
+
+    # Initialize an array to hold the importance scores
     importance_scores = np.zeros((num_groups, len(original_probabilities)))
 
     # Initialize tqdm progress bar
@@ -428,7 +432,7 @@ def main_function_upload(protein_file): #, progress=gr.Progress()
 
 prediction_imagefam = gr.outputs.Image(type='pil', label="Family prediction graph")
 prediction_imagedonor = gr.outputs.Image(type='pil', label="Donor prediction graph")
-prediction_explain = gr.outputs.Image(type='pil', label="Donor prediction explaination")
+prediction_explain = gr.outputs.Image(type='pil', label="Donor prediction explanation")
 
 
 with gr.Blocks() as app: