File size: 5,201 Bytes
9724ee5
 
 
2b8b510
 
 
 
 
8afd9fb
2b8b510
b3ad6ec
 
2b8b510
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6fd7c54
 
 
 
 
 
 
 
 
 
 
 
 
b628185
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2b8b510
 
8425393
9cadbe9
2b8b510
 
 
 
 
 
 
 
 
8425393
 
2b8b510
9cadbe9
 
 
 
5ec18f4
9cadbe9
 
b3ad6ec
3e6c28b
2b8b510
 
 
5ec18f4
1dc65be
e265b8a
3e45e67
2b8b510
 
88cec00
2b8b510
 
9724ee5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
23e4bc0
 
9724ee5
 
 
 
 
 
 
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
146
147
148
149
150
151
152
153
154
##Variables

import os
import streamlit as st
import pathlib

from langchain.embeddings import HuggingFaceEmbeddings,HuggingFaceInstructEmbeddings
from langchain.vectorstores import FAISS
from langchain.memory import ConversationBufferMemory
from langchain.chat_models.openai import ChatOpenAI
from langchain.callbacks.base import CallbackManager
from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler
from langchain import VectorDBQA
import pandas as pd

from langchain.chat_models import ChatOpenAI
from langchain.prompts.chat import (
    ChatPromptTemplate,
    SystemMessagePromptTemplate,
    AIMessagePromptTemplate,
    HumanMessagePromptTemplate,
)
from langchain.schema import (
    AIMessage,
    HumanMessage,
    SystemMessage
)

from optimum.onnxruntime import ORTModelForSequenceClassification
from transformers import pipeline, AutoTokenizer
from optimum.pipelines import pipeline
import tweepy
import pandas as pd
import numpy as np
import plotly_express as px
import plotly.graph_objects as go
from datetime import datetime as dt
from st_aggrid import GridOptionsBuilder, AgGrid, GridUpdateMode, DataReturnMode
from datasets import Dataset
from huggingface_hub import Repository

@st.experimental_singleton(suppress_st_warning=True)
def load_models():
    '''load sentimant and topic clssification models'''
    sent_pipe = pipeline(task,model=sent_model_id, tokenizer=sent_model_id)
    topic_pipe = pipeline(task, model=topic_model_id, tokenizer=topic_model_id)
    
    return sent_pipe, topic_pipe

@st.cache(allow_output_mutation=True, suppress_st_warning=True)
def process_tweets(df,df_users):
    '''process tweets into a dataframe'''
    
    df['author'] = df['author'].astype(np.int64)
    
    df_merged = df.merge(df_users, on='author')

    tweet_list = df_merged['tweet'].tolist()
    
    sentiment, topic = pd.DataFrame(sentiment_classifier(tweet_list)), pd.DataFrame(topic_classifier(tweet_list))
    
    sentiment.rename(columns={'score':'sentiment_confidence','label':'sentiment'}, inplace=True)
    
    topic.rename(columns={'score':'topic_confidence','label':'topic'}, inplace=True)
    
    df_group = pd.concat([df_merged,sentiment,topic],axis=1)

    df_group[['sentiment_confidence','topic_confidence']] = df_group[['sentiment_confidence','topic_confidence']].round(2).mul(100)

    df_tweets = df_group[['creation_time','username','tweet','sentiment','topic','sentiment_confidence','topic_confidence']]

    df_tweets = df_tweets.sort_values(by=['creation_time'],ascending=False)

    return df_tweets

@st.experimental_singleton(suppress_st_warning=True)
def create_vectorstore(texts,model,username,topic,creation_time):
    '''Create FAISS vectorstore'''

    if model == "hkunlp/instructor-large":
        emb = HuggingFaceInstructEmbeddings(model_name=model,
                                            query_instruction='Represent the Financial question for retrieving supporting documents: ',
                                            embed_instruction='Represent the Financial document for retrieval: ')
        
    elif model == "sentence-transformers/all-mpnet-base-v2":
        emb = HuggingFaceEmbeddings(model_name=model)

    docsearch = FAISS.from_texts(texts, emb,
                                metadatas=[{"source": user,"topic":top,"extraction_time":tme} for user,top,tme in zip(username,topic,creation_time)])

    return docsearch

    
@st.experimental_singleton(suppress_st_warning=True)
def embed_tweets(query,_prompt,_docsearch):
    '''Process file with latest tweets'''

    streaming_llm = ChatOpenAI(streaming=True, callback_manager=CallbackManager([StreamingStdOutCallbackHandler()]), verbose=True, temperature=0)
    chain_type_kwargs = {"prompt": _prompt}
    chain = VectorDBQA.from_chain_type(
    ChatOpenAI(temperature=0), 
    chain_type="stuff", 
    vectorstore=_docsearch,
    chain_type_kwargs=chain_type_kwargs,
    return_source_documents=True,
    k=5
    )

    result = chain({"query": query})

    return result

CONFIG = {
    "bearer_token": os.environ.get("bearer_token")
              }

sent_model_id = 'nickmuchi/optimum-finbert-tone-finetuned-fintwitter-classification'
topic_model_id = 'nickmuchi/optimum-finbert-tone-finetuned-finance-topic-classification'
task = 'text-classification'

sentiments = {"0": "Bearish", "1": "Bullish", "2": "Neutral"}

topics = {
    "0": "Analyst Update",
    "1": "Fed | Central Banks",
    "2": "Company | Product News",
    "3": "Treasuries | Corporate Debt",
    "4": "Dividend",
    "5": "Earnings",
    "6": "Energy | Oil",
    "7": "Financials",
    "8": "Currencies",
    "9": "General News | Opinion",
    "10": "Gold | Metals | Materials",
    "11": "IPO",
    "12": "Legal | Regulation",
    "13": "M&A | Investments",
    "14": "Macro",
    "15": "Markets",
    "16": "Politics",
    "17": "Personnel Change",
    "18": "Stock Commentary",
    "19": "Stock Movement",
}

sentiment_classifier, topic_classifier = load_models()

def convert_user_names(user_name: list):
    '''convert user_names to tweepy format'''
    users = []
    for user in user_name:
        users.append(f"from:{user}")
    
    return " OR ".join(users)