use ChatGPT API in flask app

apps/file-q-and-a/nextjs-with-flask-server/server/answer_question.py

@@ -42,32 +42,34 @@ def get_answer_from_files(question, session_id, pinecone_index):
                 break
             files_string += file_string
         
-        prompt = f"Given a question, try to answer it using the content of the file extracts below, and if you cannot answer, or find " \
+        messages = [
-            f"a relevant file, just output \"I couldn't find the answer to that question in your files.\".\n\n" \
+            {
-            f"If the answer is not contained in the files or if there are no file extracts, respond with \"I couldn't find the answer " \
+                "role": "system",
-            f"to that question in your files.\" If the question is not actually a question, respond with \"That's not a valid question.\"\n\n" \
+                "content": f"Given a question, try to answer it using the content of the file extracts below, and if you cannot answer, or find " \
-            f"In the cases where you can find the answer, first give the answer. Then explain how you found the answer from the source or sources, " \
+                f"a relevant file, just output \"I couldn't find the answer to that question in your files.\".\n\n" \
-            f"and use the exact filenames of the source files you mention. Do not make up the names of any other files other than those mentioned "\
+                f"If the answer is not contained in the files or if there are no file extracts, respond with \"I couldn't find the answer " \
-            f"in the files context. Give the answer in markdown format." \
+                f"to that question in your files.\" If the question is not actually a question, respond with \"That's not a valid question.\"\n\n" \
-            f"Use the following format:\n\nQuestion: <question>\n\nFiles:\n<###\n\"filename 1\"\nfile text>\n<###\n\"filename 2\"\nfile text>...\n\n"\
+                f"In the cases where you can find the answer, first give the answer. Then explain how you found the answer from the source or sources, " \
-            f"Answer: <answer or \"I couldn't find the answer to that question in your files\" or \"That's not a valid question.\">\n\n" \
+                f"and use the exact filenames of the source files you mention. Do not make up the names of any other files other than those mentioned "\
-            f"Question: {question}\n\n" \
+                f"in the files context. Give the answer in markdown format." \
-            f"Files:\n{files_string}\n" \
+                f"Use the following format:\n\nQuestion: <question>\n\nFiles:\n<###\n\"filename 1\"\nfile text>\n<###\n\"filename 2\"\nfile text>...\n\n"\
-            f"Answer:"
+                f"Answer: <answer or \"I couldn't find the answer to that question in your files\" or \"That's not a valid question.\">\n\n" \
+                f"Question: {question}\n\n" \
+                f"Files:\n{files_string}\n" \
+                f"Answer:"
+            },
+        ]
 
-        logging.info(f"[get_answer_from_files] prompt: {prompt}")
+        response = openai.ChatCompletion.create(
-
+            messages=messages,
-        response = openai.Completion.create(
+            model=GENERATIVE_MODEL,
-            prompt=prompt,
-            temperature=0,
             max_tokens=1000,
-            top_p=1,
+            temperature=0,
-            frequency_penalty=0,
-            presence_penalty=0,
-            engine=GENERATIVE_MODEL,
         )
 
-        answer = response.choices[0].text.strip()
+        choices = response["choices"]  # type: ignore
+        answer = choices[0].message.content.strip()
+        
         logging.info(f"[get_answer_from_files] answer: {answer}")
 
         return jsonify({"answer": answer})

apps/file-q-and-a/nextjs-with-flask-server/server/config.yaml

@@ -8,7 +8,7 @@ SERVER_PORT: "8080"
 
 # ---- OPENAI CONFIG -----
 EMBEDDINGS_MODEL: "text-embedding-ada-002"
-GENERATIVE_MODEL: "text-davinci-003"
+GENERATIVE_MODEL: "gpt-3.5-turbo" # use gpt-4 for better results
 EMBEDDING_DIMENSIONS: 1536
 TEXT_EMBEDDING_CHUNK_SIZE: 200
 # This is the minimum cosine similarity score that a file must have with the search query to be considered relevant

apps/file-q-and-a/nextjs-with-flask-server/server/requirements.txt

@@ -1,11 +1,11 @@
-Flask-Cors==3.0.10
+Flask-Cors>=3.0.10
-openai==0.13.0
+openai>=0.27.2
-pinecone-client==2.0.13
+pinecone-client>=2.0.13
-PyPDF2==2.10.4
+PyPDF2>=2.10.4
-numpy==1.23.2
+numpy>=1.23.2
-scikit-learn==1.1.2
+scikit-learn>=1.1.2
-docx2txt==0.8
+docx2txt>=0.8
 flask>=1.1.4
-jinja2==3.0.1
+jinja2>=3.0.1
-PyYAML==6.0
+PyYAML>=6.0
-tiktoken==0.1.2
+tiktoken>=0.1.2