File size: 5,519 Bytes
bae2915 a38e99b bae2915 004c842 f2200d9 bbb5c60 19b5c0c bbb5c60 87b4fba bbb5c60 7aa681f 698d115 7aa681f 19b5c0c 004c842 f87ddc7 3c76f86 d54014f 004c842 09b233c 903df6b 004c842 7e31ed4 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 |
from transformers import BlenderbotTokenizer, BlenderbotForConditionalGeneration
import torch
import gradio as gr
from datasets import load_dataset
# PersistDataset -----
import os
import csv
from gradio import inputs, outputs
import huggingface_hub
from huggingface_hub import Repository, hf_hub_download, upload_file
from datetime import datetime
#fastapi is where its at: share your app, share your api
import fastapi
from typing import List, Dict
import httpx
import pandas as pd
import datasets as ds
UseMemory=True
HF_TOKEN=os.environ.get("HF_TOKEN")
def SaveResult(text, outputfileName):
basedir = os.path.dirname(__file__)
savePath = outputfileName
print("Saving: " + text + " to " + savePath)
from os.path import exists
file_exists = exists(savePath)
if file_exists:
with open(outputfileName, "a") as f: #append
f.write(str(text.replace("\n"," ")))
f.write('\n')
else:
with open(outputfileName, "w") as f: #write
f.write(str("time, message, text\n")) # one time only to get column headers for CSV file
f.write(str(text.replace("\n"," ")))
f.write('\n')
return
def store_message(name: str, message: str, outputfileName: str):
basedir = os.path.dirname(__file__)
savePath = outputfileName
# if file doesnt exist, create it with labels
from os.path import exists
file_exists = exists(savePath)
if (file_exists==False):
with open(savePath, "w") as f: #write
f.write(str("time, message, text\n")) # one time only to get column headers for CSV file
if name and message:
writer = csv.DictWriter(f, fieldnames=["time", "message", "name"])
writer.writerow(
{"time": str(datetime.now()), "message": message.strip(), "name": name.strip() }
)
df = pd.read_csv(savePath)
df = df.sort_values(df.columns[0],ascending=False)
else:
if name and message:
with open(savePath, "a") as csvfile:
writer = csv.DictWriter(csvfile, fieldnames=[ "time", "message", "name", ])
writer.writerow(
{"time": str(datetime.now()), "message": message.strip(), "name": name.strip() }
)
df = pd.read_csv(savePath)
df = df.sort_values(df.columns[0],ascending=False)
return df
mname = "facebook/blenderbot-400M-distill"
model = BlenderbotForConditionalGeneration.from_pretrained(mname)
tokenizer = BlenderbotTokenizer.from_pretrained(mname)
def take_last_tokens(inputs, note_history, history):
if inputs['input_ids'].shape[1] > 128:
inputs['input_ids'] = torch.tensor([inputs['input_ids'][0][-128:].tolist()])
inputs['attention_mask'] = torch.tensor([inputs['attention_mask'][0][-128:].tolist()])
note_history = ['</s> <s>'.join(note_history[0].split('</s> <s>')[2:])]
history = history[1:]
return inputs, note_history, history
def add_note_to_history(note, note_history):# good example of non async since we wait around til we know it went okay.
note_history.append(note)
note_history = '</s> <s>'.join(note_history)
return [note_history]
title = "💬ChatBack🧠💾"
description = """Chatbot With persistent memory dataset allowing multiagent system AI to access a shared dataset as memory pool with stored interactions.
Current Best SOTA Chatbot: https://huggingface.co/facebook/blenderbot-400M-distill?text=Hey+my+name+is+ChatBack%21+Are+you+ready+to+rock%3F """
def get_base(filename):
basedir = os.path.dirname(__file__)
print(basedir)
#loadPath = basedir + "\\" + filename # works on windows
loadPath = basedir + filename
print(loadPath)
return loadPath
def chat(message, history):
history = history or []
if history:
history_useful = ['</s> <s>'.join([str(a[0])+'</s> <s>'+str(a[1]) for a in history])]
else:
history_useful = []
history_useful = add_note_to_history(message, history_useful)
inputs = tokenizer(history_useful, return_tensors="pt")
inputs, history_useful, history = take_last_tokens(inputs, history_useful, history)
reply_ids = model.generate(**inputs)
response = tokenizer.batch_decode(reply_ids, skip_special_tokens=True)[0]
history_useful = add_note_to_history(response, history_useful)
list_history = history_useful[0].split('</s> <s>')
history.append((list_history[-2], list_history[-1]))
df=pd.DataFrame()
if UseMemory:
#outputfileName = 'ChatbotMemory.csv'
outputfileName = 'ChatbotMemory3.csv' # Test first time file create
df = store_message(message, response, outputfileName) # Save to dataset
basedir = get_base(outputfileName)
return history, df, basedir
with gr.Blocks() as demo:
gr.Markdown("<h1><center>🍰Gradio chatbot backed by dataframe CSV memory🎨</center></h1>")
with gr.Row():
t1 = gr.Textbox(lines=1, default="", label="Chat Text:")
b1 = gr.Button("Respond and Retrieve Messages")
with gr.Row(): # inputs and buttons
s1 = gr.State([])
df1 = gr.Dataframe(wrap=True, max_rows=1000, overflow_row_behaviour= "paginate")
with gr.Row(): # inputs and buttons
file = gr.File(label="File")
s2 = gr.Markdown()
b1.click(fn=chat, inputs=[t1, s1], outputs=[s1, df1, file])
demo.launch(debug=True, show_error=True)
|