updated openAI embeddings.

src/embeddings/openai/remote/model.py

@@ -1,52 +1,27 @@
-import os
-import openai
-from openai.embeddings_utils import get_embedding
-from cache import AsyncTTL
-from request import ModelRequest
-import numpy as np
-import pandas as pd
-import tiktoken
-import ast
-from sklearn.metrics.pairwise import cosine_similarity
-
-openai.api_key = os.getenv("OPENAI_API_KEY")
+from openai import OpenAI
 
 
 class Model:
-    embedding_df = None
     embedding_model = "text-embedding-ada-002"
-    embedding_encoding = "cl100k_base"  # this the encoding for text-embedding-ada-002
-    max_tokens = 8000  # the maximum for text-embedding-ada-002 is 8191
 
     def __new__(cls, context):
         cls.context = context
         if not hasattr(cls, 'instance'):
-            cls.embedding_df = pd.read_csv('src/embeddings/openai/remote/akai.csv')
-            cls.embedding_df['embedding'] = cls.embedding_df['embedding'].apply(ast.literal_eval)
+            cls.instance = client = OpenAI(
+                api_key=os.getenv("OPENAI_API_KEY"),
+            )
             cls.instance = super(Model, cls).__new__(cls)
         return cls.instance
 
-    @AsyncTTL(time_to_live=600000, maxsize=1024)
     async def inference(self, request: ModelRequest):
-        print("request.prompt", request.prompt)
-        new_prompt_embedding = get_embedding(request.prompt, engine=self.embedding_model)
-        similarity_scores = cosine_similarity(
-            [new_prompt_embedding], np.stack(self.embedding_df['embedding'], axis=0))[0]
-        most_similar_indices = np.argsort(similarity_scores)[::-1]
-        most_similar_prompts = self.embedding_df.loc[most_similar_indices, ['combined_prompt', 'combined_content']]
-        most_similar_prompts['similarity_score'] = np.sort(similarity_scores)[::-1]
-        similar_content = most_similar_prompts.iloc[0:20]
-        sim_cutoff_range = np.max(similar_content['similarity_score']) - request.similarity_score_range
-        similar_content_df = similar_content.loc[similar_content['similarity_score'] >= sim_cutoff_range, :]
-        similar_content_df1 = similar_content_df.drop(columns='similarity_score')
-        similar_content_dict = similar_content_df1.to_dict('records')
-        # modified_content_dict = remove_content_tags_from_dic(similar_content_dict)
-        print("similar_content_dict", similar_content_dict)
-        return (similar_content_dict)
+    # Modify this function according to model requirements such that inputs and output remains the same
+        query = request.query
 
-    async def create_embeddings(self, embedding_df):
-        encoding = tiktoken.get_encoding(self.embedding_encoding)
-        embedding_df["n_tokens"] = embedding_df.combined_prompt.apply(lambda x: len(encoding.encode(x)))
-        embedding_df["embedding"] = embedding_df.combined_prompt.apply(
-            lambda x: get_embedding(x, engine=self.embedding_model))
-        return embedding_df
+        if(query != None):
+            embedding = client.embeddings.create(
+                input=query,
+                model=self.embedding_model,
+            ).data[0].embedding
+            return embedding
+
+        return "Invalid input"

src/embeddings/openai/remote/request.py

@@ -1,11 +1,14 @@
 import json
+import pandas as pd
 
 
 class ModelRequest():
-    def __init__(self, prompt, similarity_score_range=0):
-        self.prompt = prompt
-        self.similarity_score_range = similarity_score_range
+    def __init__(self, query=None, df = pd.DataFrame(), query_type =  None):
+         # Url to download csv file
+        self.query = query # String
+        self.query_type =  query_type
+        self.df = df
 
     def to_json(self):
         return json.dumps(self, default=lambda o: o.__dict__,
-                          sort_keys=True, indent=4)
+                          sort_keys=True, indent=4)