chore: add more Config params

Author: lpm0073

Makefile

@@ -3,7 +3,7 @@ SHELL := /bin/bash
 ifneq ("$(wildcard .env)","")
     include .env
 else
-    $(shell echo -e "OPENAI_API_ORGANIZATION=PLEASE-ADD-ME\nOPENAI_API_KEY=PLEASE-ADD-ME\nPINECONE_API_KEY=PLEASE-ADD-ME\nPINECONE_ENVIRONMENT=gcp-starter\nPINECONE_INDEX_NAME=netec-ssm\nDEBUG_MODE=True\n" >> .env)
+    $(shell echo -e "OPENAI_API_ORGANIZATION=PLEASE-ADD-ME\nOPENAI_API_KEY=PLEASE-ADD-ME\nPINECONE_API_KEY=PLEASE-ADD-ME\nPINECONE_ENVIRONMENT=gcp-starter\nPINECONE_INDEX_NAME=hsr\nOPENAI_CHAT_MODEL_NAME=gpt-3.5-turbo\nOPENAI_PROMPT_MODEL_NAME=text-davinci-003\nOPENAI_CHAT_TEMPERATURE=0.0\nOPENAI_CHAT_MAX_RETRIES=3\nDEBUG_MODE=True\n" >> .env)
 endif
 
 .PHONY: analyze init activate test lint clean

models/const.py

@@ -18,15 +18,21 @@
     PINECONE_INDEX_NAME = os.environ.get("PINECONE_INDEX_NAME", "hsr")
     OPENAI_CHAT_MODEL_NAME = os.environ.get("OPENAI_CHAT_MODEL_NAME", "gpt-3.5-turbo")
     OPENAI_PROMPT_MODEL_NAME = os.environ.get("OPENAI_PROMPT_MODEL_NAME", "text-davinci-003")
+    OPENAI_CHAT_TEMPERATURE = float(os.environ.get("OPENAI_CHAT_TEMPERATURE", 0.0))
+    OPENAI_CHAT_MAX_RETRIES = int(os.environ.get("OPENAI_CHAT_MAX_RETRIES", 3))
+    OPENAI_CHAT_CACHE = bool(os.environ.get("OPENAI_CHAT_CACHE", True))
 else:
     raise FileNotFoundError("No .env file found in root directory of repository")
 
 
 class Config:
     """Configuration parameters."""
 
-    OPENAI_CHAT_MODEL_NAME = OPENAI_CHAT_MODEL_NAME
-    OPENAI_PROMPT_MODEL_NAME = OPENAI_PROMPT_MODEL_NAME
+    OPENAI_CHAT_MODEL_NAME: str = OPENAI_CHAT_MODEL_NAME
+    OPENAI_PROMPT_MODEL_NAME: str = OPENAI_PROMPT_MODEL_NAME
+    OPENAI_CHAT_TEMPERATURE: float = OPENAI_CHAT_TEMPERATURE
+    OPENAI_CHAT_MAX_RETRIES: int = OPENAI_CHAT_MAX_RETRIES
+    OPENAI_CHAT_CACHE: bool = OPENAI_CHAT_CACHE
 
 
 class Credentials:

models/hybrid_search_retreiver.py

@@ -20,7 +20,6 @@
 import glob
 import os
 import textwrap
-from typing import List
 
 # pinecone integration
 import pinecone
@@ -78,10 +77,10 @@ class HybridSearchRetriever:
     chat = ChatOpenAI(
         api_key=Credentials.OPENAI_API_KEY,
         organization=Credentials.OPENAI_API_ORGANIZATION,
-        cache=True,
-        max_retries=3,
+        cache=Config.OPENAI_CHAT_CACHE,
+        max_retries=Config.OPENAI_CHAT_MAX_RETRIES,
         model=Config.OPENAI_CHAT_MODEL_NAME,
-        temperature=0.0,
+        temperature=Config.OPENAI_CHAT_TEMPERATURE,
     )
 
     # embeddings
@@ -110,22 +109,6 @@ def prompt_with_template(self, prompt: PromptTemplate, concept: str, model: str
         retval = llm(prompt.format(concept=concept))
         return retval
 
-    def fit_tf_idf_values(self, corpus: List[str]):
-        """Fit TF-IDF values.
-        1. Fit the BM25 encoder on the corpus
-        2. Encode the corpus
-        3. Store the encoded corpus in Pinecone
-        """
-        corpus = ["foo", "bar", "world", "hello"]
-
-        # fit tf-idf values on your corpus
-        self.bm25_encoder.fit(corpus)
-
-        # persist the values to a json file
-        self.bm25_encoder.dump("bm25_values.json")
-        self.bm25_encoder = BM25Encoder().load("bm25_values.json")
-        self.bm25_encoder.fit(corpus)
-
     def load(self, filepath: str):
         """
         Embed PDF.