Fix concurrency issues by adding unique identifiers to all generated files

- Add uuid import for generating unique identifiers
- Generate unique ID for each request in score_and_create_matrix_all_singles
- Update create_scoring_matrix_visual to use unique file names with UUID
- Update comprehensive CSV generation to include UUID
- Prevents file overwrites when multiple users run predictions simultaneously

app.py

@@ -15,6 +15,7 @@ import gradio as gr
 from huggingface_hub import hf_hub_download
 import zipfile
 import shutil
+import uuid
 
 # Add current directory to path
 sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
@@ -77,13 +78,16 @@ def create_all_single_mutants(sequence,AA_vocab=AA_vocab,mutation_range_start=No
   all_single_mutants.columns = ['mutant','mutated_sequence']
   return all_single_mutants
 
-def create_scoring_matrix_visual(scores,sequence,image_index=0,mutation_range_start=None,mutation_range_end=None,AA_vocab=AA_vocab,annotate=True,fontsize=20):
+def create_scoring_matrix_visual(scores,sequence,image_index=0,mutation_range_start=None,mutation_range_end=None,AA_vocab=AA_vocab,annotate=True,fontsize=20,unique_id=None):
+  if unique_id is None:
+    unique_id = str(uuid.uuid4())
+    
   filtered_scores=scores.copy()
   filtered_scores=filtered_scores[filtered_scores.position.isin(range(mutation_range_start,mutation_range_end+1))]
   piv=filtered_scores.pivot(index='position',columns='target_AA',values='avg_score').round(4)
   
   # Save CSV file
-  csv_path = 'fitness_scoring_substitution_matrix_{}.csv'.format(image_index)
+  csv_path = 'fitness_scoring_substitution_matrix_{}_{}.csv'.format(unique_id, image_index)
   
   # Create a more detailed CSV with mutation info
   csv_data = []
@@ -147,7 +151,7 @@ def create_scoring_matrix_visual(scores,sequence,image_index=0,mutation_range_st
   # Set x-axis labels (amino acids) - ensuring correct number
   heat.set_xticklabels(list(AA_vocab), fontsize=fontsize)
   plt.tight_layout()
-  image_path = 'fitness_scoring_substitution_matrix_{}.png'.format(image_index)
+  image_path = 'fitness_scoring_substitution_matrix_{}_{}.png'.format(unique_id, image_index)
   plt.savefig(image_path,dpi=100)
   plt.close()
   return image_path, csv_path
@@ -191,6 +195,9 @@ def get_mutated_protein(sequence,mutant):
   return ''.join(mutated_sequence)
 
 def score_and_create_matrix_all_singles(sequence,mutation_range_start=None,mutation_range_end=None,model_type="Large",scoring_mirror=False,batch_size_inference=20,max_number_positions_per_heatmap=50,num_workers=0,AA_vocab=AA_vocab):
+  # Generate unique ID for this request
+  unique_id = str(uuid.uuid4())
+  
   if mutation_range_start is None: mutation_range_start=1
   if mutation_range_end is None: mutation_range_end=len(sequence)
   
@@ -254,14 +261,14 @@ def score_and_create_matrix_all_singles(sequence,mutation_range_start=None,mutat
   window_end = min(mutation_range_end,mutation_range_start+max_number_positions_per_heatmap-1)
   
   for image_index in range(number_heatmaps):
-    image_path, csv_path = create_scoring_matrix_visual(scores,sequence,image_index,window_start,window_end,AA_vocab)
+    image_path, csv_path = create_scoring_matrix_visual(scores,sequence,image_index,window_start,window_end,AA_vocab,unique_id=unique_id)
     score_heatmaps.append(image_path)
     csv_files.append(csv_path)
     window_start += max_number_positions_per_heatmap
     window_end = min(mutation_range_end,window_start+max_number_positions_per_heatmap-1)
   
   # Also save a comprehensive CSV with all mutations
-  comprehensive_csv_path = 'all_mutations_fitness_scores.csv'
+  comprehensive_csv_path = 'all_mutations_fitness_scores_{}.csv'.format(unique_id)
   scores_export = scores[['mutant', 'position', 'target_AA', 'avg_score', 'mutated_sequence']].copy()
   scores_export['original_AA'] = scores_export['mutant'].str[0]
   scores_export = scores_export.rename(columns={'avg_score': 'fitness_score'})