File size: 11,886 Bytes
f967233
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f4b4efa
7872237
 
 
 
 
 
 
 
 
 
 
 
f4b4efa
 
 
 
 
 
 
 
 
 
 
 
 
f967233
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9fa3305
7872237
 
 
 
 
 
 
 
 
 
 
 
 
9fa3305
 
7872237
 
 
9fa3305
 
 
7872237
 
 
 
 
 
 
 
 
 
 
 
 
9fa3305
 
 
 
7872237
9fa3305
 
 
 
7872237
9fa3305
 
 
 
 
f4b4efa
 
 
 
7872237
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f4b4efa
 
 
 
7872237
f4b4efa
7872237
 
f4b4efa
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f967233
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0abb0c9
 
2da8bfd
0abb0c9
f967233
 
 
 
 
 
 
 
f4b4efa
 
651347a
 
 
 
 
 
7872237
 
651347a
7872237
651347a
 
 
 
 
 
7872237
f4b4efa
651347a
7872237
651347a
 
 
 
 
 
 
7872237
 
651347a
7872237
651347a
 
 
 
 
 
 
 
ce44247
651347a
 
 
 
ce44247
651347a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
import os
import string
from typing import Any, Dict, List, Tuple, Union

import chromadb
import numpy as np
import openai
import pandas as pd
import requests
import streamlit as st
from datasets import load_dataset
from langchain.document_loaders import TextLoader
from langchain.embeddings.sentence_transformer import SentenceTransformerEmbeddings
from langchain.text_splitter import CharacterTextSplitter
from langchain.vectorstores import Chroma
from scipy.spatial.distance import cosine

openai.api_key = os.environ["OPENAI_API_KEY"]


def merge_dataframes(dataframes: List[pd.DataFrame]) -> pd.DataFrame:
    """
    Merges a list of pandas DataFrames into a single DataFrame.

    This function concatenates the given DataFrames and filters the resulting DataFrame to only include the columns 'context', 'questions', and 'answers'.

    Parameters:
        dataframes (List[pd.DataFrame]): A list of DataFrames to be merged.

    Returns:
        pd.DataFrame: The concatenated DataFrame containing only the specified columns.
    """

    # Concatenate the list of dataframes
    combined_dataframe = pd.concat(
        dataframes, ignore_index=True
    )  # Combine all dataframes into one

    # Ensure that the resulting dataframe only contains the columns "context", "questions", "answers"
    combined_dataframe = combined_dataframe[
        ["context", "questions", "answers"]
    ]  # Filter for specific columns

    return combined_dataframe  # Return the merged and filtered DataFrame


def call_chatgpt(prompt: str) -> str:
    """
    Uses the OpenAI API to generate an AI response to a prompt.

    Args:
        prompt: A string representing the prompt to send to the OpenAI API.

    Returns:
        A string representing the AI's generated response.

    """

    # Use the OpenAI API to generate a response based on the input prompt.
    response = openai.Completion.create(
        model="gpt-3.5-turbo-instruct",
        prompt=prompt,
        temperature=0.5,
        max_tokens=500,
        top_p=1,
        frequency_penalty=0,
        presence_penalty=0,
    )

    # Extract the text from the first (and only) choice in the response output.
    ans = response.choices[0]["text"]

    # Return the generated AI response.
    return ans


def openai_text_embedding(prompt: str) -> str:
    """
    Retrieves the text embedding for a given prompt using OpenAI's text-embedding model.

    This function utilizes OpenAI's API to generate an embedding for the input text. It specifically uses the "text-embedding-ada-002" model.

    Parameters:
        prompt (str): The text input for which to generate an embedding.

    Returns:
        str: A string representation of the text embedding.
    """

    # Call OpenAI API to create a text embedding
    return openai.Embedding.create(input=prompt, model="text-embedding-ada-002")[
        "data"
    ][0][
        "embedding"
    ]  # Retrieve the embedding from the response


def calculate_sts_openai_score(sentence1: str, sentence2: str) -> float:
    """
    Calculates the Semantic Textual Similarity (STS) between two sentences using OpenAI's text-embedding model.

    This function computes embeddings for each sentence and then calculates the cosine similarity between these embeddings. A higher score indicates greater similarity.

    Parameters:
        sentence1 (str): The first sentence for similarity comparison.
        sentence2 (str): The second sentence for similarity comparison.

    Returns:
        float: The STS score representing the similarity between sentence1 and sentence2.
    """

    # Compute sentence embeddings
    embedding1 = openai_text_embedding(sentence1)  # Flatten the embedding array
    embedding2 = openai_text_embedding(sentence2)  # Flatten the embedding array

    # Convert embeddings to NumPy arrays
    embedding1 = np.asarray(embedding1)
    embedding2 = np.asarray(embedding2)

    # Calculate cosine similarity between the embeddings
    # Since 'cosine' returns the distance, 1 - distance is used to get similarity
    similarity_score = 1 - cosine(embedding1, embedding2)

    return similarity_score


def add_dist_score_column(
    dataframe: pd.DataFrame,
    sentence: str,
) -> pd.DataFrame:
    """
    Adds a new column to the provided DataFrame with STS (Semantic Textual Similarity) scores,
    calculated between a given sentence and each question in the 'questions' column of the DataFrame.
    The DataFrame is then sorted by this new column in descending order and the top 5 rows are returned.

    Parameters:
        dataframe (pd.DataFrame): A pandas DataFrame containing a 'questions' column.
        sentence (str): The sentence against which to compute STS scores for each question in the DataFrame.

    Returns:
        pd.DataFrame: A DataFrame containing the original data along with the new 'stsopenai' column,
                       sorted by the 'stsopenai' column, and limited to the top 5 entries with the highest scores.
    """

    # Calculate the STS score between `sentence` and each row's `question`
    dataframe["stsopenai"] = dataframe["questions"].apply(
        lambda x: calculate_sts_openai_score(str(x), sentence)
    )

    # Sort the dataframe by the newly added 'stsopenai' column in descending order
    sorted_dataframe = dataframe.sort_values(by="stsopenai", ascending=False)

    # Return the top 5 rows from the sorted dataframe
    return sorted_dataframe.iloc[:5, :]


def convert_to_list_of_dict(df: pd.DataFrame) -> List[Dict[str, str]]:
    """
    Reads in a pandas DataFrame and produces a list of dictionaries with two keys each, 'question' and 'answer.'

    Args:
        df: A pandas DataFrame with columns named 'questions' and 'answers'.

    Returns:
        A list of dictionaries, with each dictionary containing a 'question' and 'answer' key-value pair.
    """

    # Initialize an empty list to store the dictionaries
    result = []

    # Loop through each row of the DataFrame
    for index, row in df.iterrows():
        # Create a dictionary with the current question and answer
        qa_dict_quest = {"role": "user", "content": row["questions"]}
        qa_dict_ans = {"role": "assistant", "content": row["answers"]}

        # Add the dictionary to the result list
        result.append(qa_dict_quest)
        result.append(qa_dict_ans)

    # Return the list of dictionaries
    return result


def query(payload: Dict[str, Any]) -> Dict[str, Any]:
    """
    Sends a JSON payload to a predefined API URL and returns the JSON response.
    Args:
        payload (Dict[str, Any]): The JSON payload to be sent to the API.
    Returns:
        Dict[str, Any]: The JSON response received from the API.
    """

    # API endpoint URL
    API_URL = "https://sks7h7h5qkhoxwxo.us-east-1.aws.endpoints.huggingface.cloud"

    # Headers to indicate both the request and response formats are JSON
    headers = {"Accept": "application/json", "Content-Type": "application/json"}

    # Sending a POST request with the JSON payload and headers
    response = requests.post(API_URL, headers=headers, json=payload)

    # Returning the JSON response
    return response.json()


def llama2_7b_ysa(prompt: str) -> str:
    """
    Queries a model and retrieves the generated text based on the given prompt.
    This function sends a prompt to a model (presumably named 'llama2_7b') and extracts
    the generated text from the model's response. It's tailored for handling responses
    from a specific API or model query structure where the response is expected to be
    a list of dictionaries, with at least one dictionary containing a key 'generated_text'.
    Parameters:
    - prompt (str): The text prompt to send to the model.
    Returns:
    - str: The generated text response from the model.
    Note:
    - The function assumes that the 'query' function is previously defined and accessible
      within the same scope or module. It should send a request to the model and return
      the response in a structured format.
    - The 'parameters' dictionary is passed empty but can be customized to include specific
      request parameters as needed by the model API.
    """

    # Define the query payload with the prompt and any additional parameters
    query_payload: Dict[str, Any] = {
        "inputs": prompt,
        "parameters": {"max_new_tokens": 20},
    }

    # Send the query to the model and store the output response
    output = query(query_payload)

    # Extract the 'generated_text' from the first item in the response list
    response: str = output[0]["generated_text"]

    return response


def quantize_to_kbit(arr: Union[np.ndarray, Any], k: int = 16) -> np.ndarray:
    """Converts an array to a k-bit representation by normalizing and scaling its values.
    
    Args:
        arr (Union[np.ndarray, Any]): The input array to be quantized.
        k (int): The number of levels to quantize to. Defaults to 16 for 4-bit quantization.

    Returns:
        np.ndarray: The quantized array with values scaled to 0 to k-1.
    """
    if not isinstance(arr, np.ndarray):  # Check if input is not a numpy array
        arr = np.array(arr)  # Convert input to a numpy array
    arr_min = arr.min()  # Calculate the minimum value in the array
    arr_max = arr.max()  # Calculate the maximum value in the array
    normalized_arr = (arr - arr_min) / (arr_max - arr_min)  # Normalize array values to [0, 1]
    return np.round(normalized_arr * (k - 1)).astype(int)  # Scale normalized values to 0-(k-1) and convert to integer


def quantized_influence(arr1: np.ndarray, arr2: np.ndarray, k: int = 16, use_dagger: bool = False) -> Tuple[float, List[float]]:
    """
    Calculates a weighted measure of influence based on quantized version of input arrays and optionally applies a transformation.
    
    Args:
        arr1 (np.ndarray): First input array to be quantized and analyzed.
        arr2 (np.ndarray): Second input array to be quantized and used for influence measurement.
        k (int): The quantization level, defaults to 16 for 4-bit quantization.
        use_dagger (bool): Flag to apply a transformation based on local averages, defaults to False.

    Returns:
        Tuple[float, List[float]]: A tuple containing the quantized influence measure and an optional list of transformed values based on local estimates.
    """
    # Quantize both arrays to k levels
    arr1_quantized = quantize_to_kbit(arr1, k)
    arr2_quantized = quantize_to_kbit(arr2, k)
    
    # Find unique quantized values in arr1
    unique_values = np.unique(arr1_quantized)
    
    # Compute the global average of quantized arr2
    total_samples = len(arr2_quantized)
    y_bar_global = np.mean(arr2_quantized)  
    
    # Compute weighted local averages and normalize
    weighted_local_averages = [(np.mean(arr2_quantized[arr1_quantized == val]) - y_bar_global)**2 * len(arr2_quantized[arr1_quantized == val])**2 for val in unique_values]
    qim = np.sum(weighted_local_averages) / (total_samples * np.std(arr2_quantized))  # Calculate the quantized influence measure

    if use_dagger:
        # If use_dagger is True, compute local estimates and map them to unique quantized values
        local_estimates = [np.mean(arr2_quantized[arr1_quantized == val]) for val in unique_values]
        daggers = {unique_values[i]: v for i, v in enumerate(local_estimates)}  # Map unique values to local estimates

        def find_val_(i: int) -> float:
            """Helper function to map quantized values to their local estimates."""
            return daggers[i]

        # Apply transformation based on local estimates
        daggered_values = list(map(find_val_, arr1_quantized))
    else:
        # If use_dagger is False, return the original quantized arr1 values
        daggered_values = arr1_quantized.tolist()

    return qim, daggered_values