Update app.py

app.py

@@ -20,7 +20,9 @@ from tenacity import retry
 from tqdm import tqdm
 import tiktoken
 import scipy.stats
+import inseq
 import torch
+from transformers import AutoModelForCausalLM
 from transformers import GPT2LMHeadModel
 import tiktoken
 import seaborn as sns
@@ -28,6 +30,24 @@ from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
 # from colorama import Fore, Style
 import openai  # for OpenAI API calls
 
+######################################
+def find_indices(arr, target):
+    indices = []
+    start_index = None
+
+    for i, element in enumerate(arr):
+        if target in element:
+            if start_index is None:
+                start_index = i
+            else:
+                indices.append((start_index, i - 1))
+                start_index = i
+
+    if start_index is not None:
+        indices.append((start_index, len(arr) - 1))
+
+    return indices
+
 ######################################
 import streamlit as st
 def colorize_tokens(token_data, sentence):
@@ -50,16 +70,16 @@ def colorize_tokens(token_data, sentence):
 
 # Define colors for the tags
 tag_colors = {
-    "ADJP": "#0000FF",  # Blue
-    "ADVP": "#008000",  # Green
-    "CONJP": "#FF0000",  # Red
-    "INTJ": "#00FFFF",  # Cyan
-    "LST": "#FF00FF",  # Magenta
-    "NP": "#FFFF00",  # Yellow
-    "PP": "#800080",  # Purple
-    "PRT": "#00008B",  # Dark Blue
-    "SBAR": "#006400",  # Dark Green
-    "VP": "#008B8B",  # Dark Cyan
+    "ADJP": "#8F6B9F",  # Blue
+    "ADVP": "#7275A7",  # Green
+    "CONJP": "#5BA4BB",  # Red
+    "INTJ": "#95CA73",  # Cyan
+    "LST": "#DFDA70",  # Magenta
+    "NP": "#EFBC65",  # Yellow
+    "PP": "#FC979B",  # Purple
+    "PRT": "#F1C5C1",  # Dark Blue
+    "SBAR": "#FAEBE8",  # Dark Green
+    "VP": "#90DFD2",  # Dark Cyan
 }
 ##################