File size: 7,568 Bytes
844cee8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
import os
import openai
import numpy as np
import time

import time
import pandas as pd

MODEL_AUGMENT = "gpt-3.5-turbo-16k"
MODEL_ANSWER = "gpt-3.5-turbo-16k"

def cosine_similarity(a, b):
    return np.dot(a, b) / (np.linalg.norm(a) * np.linalg.norm(b))

def get_embedding(text, model="text-embedding-ada-002"):
    try:
        text = text.replace("\n", " ")
    except:
        None
    try:
        return openai.Embedding.create(input = [text], model=model)['data'][0]['embedding']
    except:
        time.sleep(2)

def augment_query(query):
    SYS_PROMPT = """
        On [current date: 19 July], you'll receive a finance-related question from a sales manager, without direct interaction. Generate a JSON response with the following structure, considering the temporal aspect:

        {
        "timestamps": # Relevant timestamps to study corresponding tweets for a temporal dynamic aspect (e.g., topic drift). USE THE MINIMAL NUMBER OF TIMESTAMP POSSIBLE ALWAYS ALWAYS!,
        "query": # Repeat the user's query,
        "similarity_boilerplate": # Boilerplate of relevant documents for cosine similarity search after embedding (it could look like example of tweets that might help answer the query),
        }

        Allowed historical timestamps:
        ['2018-07-18', '2018-07-19', '2018-07-08', '2018-07-09', '2018-07-10', '2018-07-11', '2018-07-12', '2018-07-13', '2018-07-14', '2018-07-15', '2018-07-16', '2018-07-17']

        Ensure the output is always in JSON format and never provide any other response.
        """
    response = openai.chat.completions.create(
        model=MODEL_AUGMENT,
        messages=
        [
            {
            "role": "system",
            "content": SYS_PROMPT
             },
            {
            "role": "user",
            "content": query
            }
        ],
        temperature=1,
        max_tokens=1000,
        top_p=1,
        frequency_penalty=0,
        presence_penalty=0,
        ).choices[0].message.content
    return response


def semantic_search(df_loc, query,timestamp, nb_elements_to_consider=15):
    timestamp = str(timestamp).strip()
    embedding = get_embedding(query, model='text-embedding-ada-002')
    filtered_df = df_loc[df_loc["timestamp"]==timestamp].drop(columns=["url"])
    def wrap_cos(x,y):
        try:
            res = cosine_similarity(x,y)
        except:
            res = 0
        return res
    filtered_df['similarity']  = filtered_df.embedding.apply(lambda x: wrap_cos(x, embedding))

    results = filtered_df.sort_values('similarity', ascending=False).head(nb_elements_to_consider)

    return results

def condition_check(tweet, query):
    response = openai.chat.completions.create(model=MODEL_AUGMENT,messages=[    {
        "role": "system",
        "content": "Only answer with True or False no matter what"
        },
        {
        "role": "user",
        "content": f"Consider this tweet:\n\n{tweet}\n\nIs it relevant to the following query:\n\n\{query}"
        }
    ],
    temperature=1,
    max_tokens=1000,
    top_p=1,
    frequency_penalty=0,
    presence_penalty=0
    ).choices[0].message.content
    return bool(response)

def get_number_relevant_tweets(df,timestamp, query):
    sorted_df = semantic_search(df, str(str(query["query"]) + "\n"+  str(query["similarity_boilerplate"])),timestamp, nb_elements_to_consider=len(df))
    left, right = 0, len(sorted_df) - 1
    while left <= right:
        mid = (left + right) // 2
        print(f"Currently searching with max range at {mid}")
        if condition_check(sorted_df['text'].iloc[mid], query):
            left = mid + 1
        else:
            right = mid - 1
    print(f"Dichotomy done, found relevant tweets: {left}")
    return left



def get_relevant_documents(df, query,nb_elements_to_consider = 10):
    query = eval(query)
    all_retrieved = []
    for timestamp in query["timestamps"]:
        number_of_relevant_tweets = get_number_relevant_tweets(df,timestamp, query)
        all_retrieved.append({
            "timestamp" : timestamp,
            "number_of_relevant_tweets": str(number_of_relevant_tweets),
            "tweets" : semantic_search(df, str(str(query["query"]) + "\n"+  str(query["similarity_boilerplate"])),timestamp, nb_elements_to_consider=min(nb_elements_to_consider,number_of_relevant_tweets))
        })
    return all_retrieved

def get_final_answer(relevant_documents, query):
    context = ""
    for document in relevant_documents:
        print("TIMESTAMP: ", document["timestamp"] )
        tweet_entry = document["tweets"]
        context += "\nTimestamp: " + document["timestamp"] + " - Number of relevant tweets in database (EXACT VOLUME OF TWEETS): +"+ document["number_of_relevant_tweets"] + "\nList of tweets:\n" + str((tweet_entry["text"] + "   --- Tweeted by: @" +tweet_entry["source"] +  " \n").to_list()) + "\n---"


    SYS_PROMPT =  f"""
        You will be fed a list of tweets each at a specific timestamp and the number of relevant tweets. You need to take into account (if needed) the number of tweets relevant to the query and how this number evolved. Your task is to use those tweets to answer to the best of your knowledge the following question:

        QUESTION: {query}

        SPECIFIC INSTRUCTIONS AND SYSTEM WARNINGS: You redact a properly structured markdown string containing a professional report.
        You ALWAYS specify your sources by citing them (no urls though). Those tweets are samples from the data and are the closest to the query, you should also take into account the volume of tweets obtained.
        Otherwise, it will be considered highly misleading and harmful content.
        You should however always try your best to answer and you need to study in depth the historical relationship between the timestamps and how it answers the QUESTION.
        You never refer to yourself.
        Make it as if a real human provided a well constructed and structured report/answer extracting the best of the knowledge contained in the context."
        """
    response = openai.chat.completions.create(
                                                model=MODEL_ANSWER,
                                                messages=[
                                                    {
                                                    "role": "system",
                                                    "content": SYS_PROMPT
                                                            },
                                                    {
                                                    "role": "user",
                                                    "content": str(context)
                                                    }
                                                ],

                                                temperature=1,
                                                max_tokens=3000,
                                                top_p=1,
                                                frequency_penalty=0,
                                                presence_penalty=0,
                                                ).choices[0].message.content
    return response

def get_answer(query, df,api_key,nb_elements_to_consider=10):
    openai.api_key = api_key
    augmented_query = augment_query(query)

    relevant_documents = get_relevant_documents(df, augmented_query,nb_elements_to_consider=nb_elements_to_consider)

    response = get_final_answer(relevant_documents, augmented_query)
    print(response)


    return response