Update app.py

app.py

@@ -1,22 +1,16 @@
 import gradio as gr
 import base64
 import random
-import gradio as gr
-#import urllib.request
 import requests
 import bs4
 import lxml
 import os
-#import subprocess
 from huggingface_hub import InferenceClient,HfApi
 import random
 import json
 import datetime
 from pypdf import PdfReader
 import uuid
-#from query import tasks
-from gradio_client import Client
-import cairosvg
 
 from agent import (
     PREFIX,
@@ -27,19 +21,9 @@ from agent import (
     LOG_RESPONSE,
 )
 api=HfApi()
-
 client = InferenceClient("mistralai/Mixtral-8x7B-Instruct-v0.1")
 
-
-def sort_fn(inp):
-
-    client_sort = Client("Omnibus/sort_document")
-    sen,nouns = client_sort.predict(
-    		f"{inp}",	# str  in 'Paste Text' Textbox component
-    		api_name="/sort_doc"
-    )
-    return nouns
-
+############  Document Functions #################
 def find_all(url):
     return_list=[]
     print (url)
@@ -68,51 +52,8 @@ def find_all(url):
     except Exception as e:
         print (e)
         return False, f'Error: {e}'
-
-        #else:
-    #    history = "observation: The search query I used did not return a valid response"
-        
     return "MAIN", None, history, task
 
-FIND_KEYWORDS="""Find keywords from the dictionary of provided keywords that are relevant to the users query.
-Return the keyword:value pairs from the list in the form of a JSON file output.
-dictionary:
-{keywords}
-user query:
-"""
-
-def find_keyword_fn(c,inp,data):
-
-    data=str(data)
-    seed=random.randint(1,1000000000)
-    divr=int(c)/20000
-    divi=int(divr)+1 if divr != int(divr) else int(divr)
-    chunk = int(int(c)/divr)
-    out = []
-    s=0
-    e=chunk
-    print(f'e:: {e}')
-    #task = f'Compile this data to fulfill the task: {task}, and complete the purpose: {purpose}\n'
-    for z in range(divi):
-        print(f's:e :: {s}:{e}')
-        
-        hist = data[s:e]
-        resp = run_gpt(
-            FIND_KEYWORDS,
-            stop_tokens=[],
-            max_tokens=8192,
-            seed=seed,
-            keywords=data,
-        ).strip("\n")    
-        out.append(resp)
-        #new_history = resp
-        print (resp)
-        #out+=resp
-        e=e+chunk
-        s=s+chunk    
-    return out 
-    
-
 def read_txt(txt_path):
     text=""
     with open(txt_path,"r") as f:
@@ -138,9 +79,6 @@ def read_pdf_online(url):
     response = requests.get(url, stream=True)
     print(response.status_code)
     text=""
-#################
-    
-#####################
     try:
         if response.status_code == 200:
             with open("test.pdf", "wb") as f:
@@ -300,60 +238,7 @@ def compress_data(c, instruct, history):
         s=s+chunk
     return out
 
-    
-def compress_data_og(c, instruct, history):
-    seed=random.randint(1,1000000000)
-    
-    print (c)
-    #tot=len(purpose)
-    #print(tot)
-    divr=int(c)/MAX_DATA
-    divi=int(divr)+1 if divr != int(divr) else int(divr)
-    chunk = int(int(c)/divr)
-    print(f'chunk:: {chunk}')
-    print(f'divr:: {divr}')
-    print (f'divi:: {divi}')
-    out = []
-    #out=""
-    s=0
-    e=chunk
-    print(f'e:: {e}')
-    new_history=""
-    #task = f'Compile this data to fulfill the task: {task}, and complete the purpose: {purpose}\n'
-    for z in range(divi):
-        print(f's:e :: {s}:{e}')
-        
-        hist = history[s:e]
-        
-        resp = run_gpt(
-            COMPRESS_DATA_PROMPT,
-            stop_tokens=["observation:", "task:", "action:", "thought:"],
-            max_tokens=8192,
-            seed=seed,
-            direction=instruct,
-            knowledge=new_history,
-            history=hist,
-        ).strip("\n")
-        
-        new_history = resp
-        print (resp)
-        out+=resp
-        e=e+chunk
-        s=s+chunk
-    '''
-    resp = run_gpt(
-        COMPRESS_DATA_PROMPT,
-        stop_tokens=["observation:", "task:", "action:", "thought:"],
-        max_tokens=8192,
-        seed=seed,
-        direction=instruct,
-        knowledge=new_history,
-        history="All data has been recieved.",
-    )'''
-    print ("final" + resp)
-    #history = "observation: {}\n".format(resp)
-    return resp
-
+  
 def get_chart(inp):
     seed=random.randint(1,1000000000)
     try: