refactoring the code on the SOLID principles

chat.py

@@ -1,66 +0,0 @@
-import os
-import openai
-from dotenv import load_dotenv
-import jsonlines
-from pathlib import Path
-from utils import (
-    gpt3_embeddings,
-    gpt3_completion,
-    dot_similarity,
-    load_prompt,
-)
-
-load_dotenv()
-
-OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
-
-openai.api_key = OPENAI_API_KEY
-
-
-def search_index(question, indexes, count=4):
-    question_embedding = gpt3_embeddings(question)
-
-    simmilarities = []
-    for index in indexes:
-        embedding = index["embedding"]
-        score = dot_similarity(question_embedding, embedding)
-        simmilarities.append({"index": index, "score": score})
-
-        sorted_similarities = sorted(
-            simmilarities, key=lambda x: x["score"], reverse=True
-        )
-
-    return sorted_similarities[:count]
-
-
-if __name__ == "__main__":
-    with jsonlines.open(Path("./index") / "index.jsonl") as passages:
-        indexes = list(passages)
-
-    while True:
-        question = input("User >")
-
-        search_results = search_index(question=question, indexes=indexes, count=2)
-
-        answers = []
-        for result in search_results:
-            print("iterating over answering questions")
-
-            prompt = (
-                load_prompt("prompts\question_answering.txt")
-                .replace("<<PASSAGE>>", result["index"]["content"])
-                .replace("<<QUESTION>>", question)
-            )
-
-            answer = gpt3_completion(
-                prompt=prompt, max_tokens=80, model="text-curie-001"
-            )
-            answers.append(answer)
-
-            prompt = load_prompt("prompts\passage_summarization.txt").replace(
-                "<<PASSAGE>>", "\n".join(answers)
-            )
-
-            final_answer = gpt3_completion(prompt=prompt)
-
-        print(f"Bot: {final_answer}")

index/build_index.py

@@ -7,7 +7,7 @@ import openai
 import textwrap
 import jsonlines
 
-from utils import gpt3_embeddings
+from src.utils import gpt3_embeddings
 
 load_dotenv()
 

src/chat.py

@@ -0,0 +1,41 @@
+import os
+import openai
+from dotenv import load_dotenv
+from index import IndexSearchEngine
+from gpt_3_manager import Gpt3Manager
+from prompt import QuestionAnsweringPrompt, PassageSummarizationPrompt
+
+
+load_dotenv()
+
+OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
+
+openai.api_key = OPENAI_API_KEY
+
+
+class ChatBot:
+    def __init__(self, index_search_engine: IndexSearchEngine):
+        self.index_search_engine = index_search_engine
+
+    def ask(self, question):
+        search_result = self.index_search_engine.search(question=question)
+
+        answers = []
+        for result in search_result:
+            print("iterating over answering questions")
+
+            question_answering_prompt = QuestionAnsweringPrompt.load(
+                "prompts\question_answering.txt"
+            )
+
+            answer = Gpt3Manager.get_completion(
+                prompt=question_answering_prompt, max_tokens=80, model="text-curie-001"
+            )
+            answers.append(answer)
+
+        passage_summarization_prompt = PassageSummarizationPrompt.load(
+            "prompts\passage_summarization.txt"
+        )
+
+        final_answer = Gpt3Manager.get_completion(prompt=passage_summarization_prompt)
+        return final_answer

src/gpt_3_manager.py

@@ -0,0 +1,32 @@
+import openai
+
+
+class Gpt3Manager:
+    def __init__(self, api_key):
+        openai.api_key = api_key
+
+    def get_completion(prompt, max_tokens=128, model="text-davinci-003"):
+        response = None
+        try:
+            response = openai.Completion.create(
+                model=model,
+                prompt=prompt,
+                max_tokens=max_tokens,
+            )["choices"][0]["text"]
+
+        except Exception as err:
+            print(f"Sorry, There was a problem \n\n {err}")
+
+        return response
+
+    def get_embedding(text, model="text-similarity-ada-001"):
+        text = text.replace("\n", " ")
+        embedding = None
+        try:
+            embedding = openai.Embedding.create(input=[text], model=model)["data"][0][
+                "embedding"
+            ]
+        except Exception as err:
+            print(f"Sorry, There was a problem {err}")
+
+        return embedding

src/index.py

@@ -0,0 +1,40 @@
+from abc import ABC, abstractmethod
+import jsonlines
+from gpt_3_manager import Gpt3Manager
+from src.utils import dot_similarity
+
+
+class Index(ABC):
+    @abstractmethod
+    def load(self, path):
+        pass
+
+
+class JsonLinesIndex(Index):
+    def __init__(self):
+        pass
+
+    def load(self, path):
+        with jsonlines.open(path) as passages:
+            indexes = list(passages)
+        return indexes
+
+
+class IndexSearchEngine:
+    def __init__(self, index):
+        index = index
+
+    def search(self, question, indexes, count=4):
+        question_embedding = Gpt3Manager.get_embedding(question)
+
+        simmilarities = []
+        for index in indexes:
+            embedding = index["embedding"]
+            score = dot_similarity(question_embedding, embedding)
+            simmilarities.append({"index": index, "score": score})
+
+            sorted_similarities = sorted(
+                simmilarities, key=lambda x: x["score"], reverse=True
+            )
+
+        return sorted_similarities[:count]

src/prompt.py

@@ -0,0 +1,35 @@
+from abc import ABC, abstractmethod
+
+
+class Prompt(ABC):
+    def load_prompt(path):
+        with open(path) as f:
+            lines = f.readlines()
+            return "".join(lines)
+
+    @abstractmethod
+    def load(self, path):
+        pass
+
+
+class QuestionAnsweringPrompt(Prompt):
+    def __init__(self, result, question):
+        result = result
+        question = question
+
+    def load(self, path):
+        prompt = (
+            self.load_prompt(path)
+            .replace("<<PASSAGE>>", self.result["index"]["content"])
+            .replace("<<QUESTION>>", self.question)
+        )
+        return prompt
+
+
+class PassageSummarizationPrompt(Prompt):
+    def __init__(self, answers):
+        self.answers = answers
+
+    def load(self, path):
+        prompt = self.load_prompt(path).replace("<<PASSAGE>>", "\n".join(self.answers))
+        return prompt

utils.py → src/utils.py

@@ -1,35 +1,6 @@
-import openai
 import numpy as np
 
 
-def gpt3_embeddings(text, model="text-similarity-ada-001"):
-    text = text.replace("\n", " ")
-    embedding = None
-    try:
-        embedding = openai.Embedding.create(input=[text], model=model)["data"][0][
-            "embedding"
-        ]
-    except Exception as err:
-        print(f"Sorry, There was a problem {err}")
-
-    return embedding
-
-
-def gpt3_completion(prompt, max_tokens=128, model="text-davinci-003"):
-    response = None
-    try:
-        response = openai.Completion.create(
-            model=model,
-            prompt=prompt,
-            max_tokens=max_tokens,
-        )["choices"][0]["text"]
-
-    except Exception as err:
-        print(f"Sorry, There was a problem \n\n {err}")
-
-    return response
-
-
 def load_prompt(path):
     with open(path) as f:
         lines = f.readlines()