Update app.py

app.py

@@ -4,6 +4,17 @@ from mosestokenizer import *
 from indicnlp.tokenize import sentence_tokenize
 from docx import Document
 
+import os
+
+os.system('git clone https://github.com/TheAtticusProject/cuad.git')
+os.system('mv cuad cuad-training')
+os.system('unzip cuad-training/data.zip -d cuad-data/')
+os.system('mkdir cuad-models')
+os.system('curl https://zenodo.org/record/4599830/files/roberta-base.zip?download=1 --output cuad-models/roberta-base.zip')
+os.system('unzip cuad-models/roberta-base.zip -d cuad-models/')
+
+
+
 trans_tokenizer = AutoTokenizer.from_pretrained("facebook/nllb-200-distilled-600M" )
 trans_model = AutoModelForSeq2SeqLM.from_pretrained("facebook/nllb-200-distilled-600M")
 device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
@@ -487,7 +498,7 @@ def run_generate_questions(document_name, output_file, questions_file, delimiter
         return qg_output, q_output
         
         
- import docx
+import docx
 import random
 from docx.shared import RGBColor
 import time 
@@ -515,7 +526,7 @@ def run_redflags(filename, output_file):
         return output_file 
         
         
- import docx
+import docx
 import random
 from docx.shared import RGBColor
 import time 
@@ -637,7 +648,7 @@ def run_similar_clause(filename, output_file, clauses, source_language):
             return output_file, highlighted_paras
     
     
-  import gradio as gr
+import gradio as gr
 
 analysis_services = ['Translate Contract', 'Identify key Clauses', 'Red flag Identification', 'Similar Semantic Clause search', 'Generate Questions for Contract Template']
 analysis_label = 'Select Contract Analysis Service'