From 3fe88f7725aa7fa83a5ec4bc548cdf7e7d5ca054 Mon Sep 17 00:00:00 2001
From: sauravniraula <developmentsaurav@gmail.com>
Date: Tue, 24 Jun 2025 14:52:16 +0545
Subject: [PATCH] feat(fastapi): use openai instead of langchain

---
 Dockerfile.dev                           |  2 +-
 servers/fastapi/api/models.py            |  7 ++++
 servers/fastapi/api/utils/model_utils.py | 52 ++++++++++++++++++++++++
 servers/fastapi/api/utils/utils.py       | 37 -----------------
 4 files changed, 60 insertions(+), 38 deletions(-)
 create mode 100644 servers/fastapi/api/utils/model_utils.py

diff --git a/Dockerfile.dev b/Dockerfile.dev
index 4c8d2c2b..58c401ef 100644
--- a/Dockerfile.dev
+++ b/Dockerfile.dev
@@ -37,7 +37,7 @@ RUN npx puppeteer browsers install chrome --install-deps
 COPY nginx.conf /etc/nginx/nginx.conf
 
 # Expose the port
-EXPOSE 80 3000 8000 6379
+EXPOSE 80 3000 8000
 
 # Start the servers
 CMD ["/bin/bash", "-c", "ollama serve & service nginx start & service redis-server start && node /app/start.js"]
\ No newline at end of file
diff --git a/servers/fastapi/api/models.py b/servers/fastapi/api/models.py
index 15d38ae4..84cda2df 100644
--- a/servers/fastapi/api/models.py
+++ b/servers/fastapi/api/models.py
@@ -1,3 +1,4 @@
+from enum import Enum
 import json
 from typing import Optional
 from pydantic import BaseModel
@@ -73,3 +74,9 @@ class OllamaModelMetadata(BaseModel):
     icon: str
     size: str
     supports_graph: bool
+
+
+class SelectedLLMProvider(Enum):
+    OLLAMA = "ollama"
+    OPENAI = "openai"
+    GOOGLE = "google"
diff --git a/servers/fastapi/api/utils/model_utils.py b/servers/fastapi/api/utils/model_utils.py
new file mode 100644
index 00000000..e642e066
--- /dev/null
+++ b/servers/fastapi/api/utils/model_utils.py
@@ -0,0 +1,52 @@
+import os
+
+from api.models import SelectedLLMProvider
+
+
+def is_ollama_selected() -> bool:
+    return get_selected_llm_provider() == SelectedLLMProvider.OLLAMA
+
+
+def get_selected_llm_provider() -> SelectedLLMProvider:
+    return SelectedLLMProvider(os.getenv("LLM"))
+
+
+def get_model_base_url():
+    selected_llm = get_selected_llm_provider()
+
+    if selected_llm == SelectedLLMProvider.OLLAMA:
+        return "http://localhost:11434"
+    elif selected_llm == SelectedLLMProvider.OPENAI:
+        return "https://api.openai.com/v1"
+    else:
+        return "https://generativelanguage.googleapis.com/v1beta/openai"
+
+
+def get_large_model():
+    selected_llm = get_selected_llm_provider()
+    if selected_llm == SelectedLLMProvider.OPENAI:
+        return ChatOpenAI(model="gpt-4.1")
+    elif selected_llm == SelectedLLMProvider.GOOGLE:
+        return ChatGoogleGenerativeAI(model="gemini-2.0-flash")
+    else:
+        return ChatOllama(model=os.getenv("OLLAMA_MODEL"), temperature=0.8)
+
+
+def get_small_model():
+    selected_llm = os.getenv("LLM")
+    if selected_llm == "openai":
+        return ChatOpenAI(model="gpt-4.1-mini")
+    elif selected_llm == "google":
+        return ChatGoogleGenerativeAI(model="gemini-2.0-flash")
+    else:
+        return ChatOllama(model=os.getenv("OLLAMA_MODEL"), temperature=0.8)
+
+
+def get_nano_model():
+    selected_llm = os.getenv("LLM")
+    if selected_llm == "openai":
+        return ChatOpenAI(model="gpt-4.1-nano")
+    elif selected_llm == "google":
+        return ChatGoogleGenerativeAI(model="gemini-2.0-flash")
+    else:
+        return ChatOllama(model=os.getenv("OLLAMA_MODEL"), temperature=0.8)
diff --git a/servers/fastapi/api/utils/utils.py b/servers/fastapi/api/utils/utils.py
index 90882961..16b434d0 100644
--- a/servers/fastapi/api/utils/utils.py
+++ b/servers/fastapi/api/utils/utils.py
@@ -9,48 +9,11 @@ from typing import List, Optional
 import aiohttp
 from fastapi import HTTPException, UploadFile
 from fastapi.responses import StreamingResponse
-from langchain_google_genai import ChatGoogleGenerativeAI
-from langchain_ollama import ChatOllama
-from langchain_openai import ChatOpenAI
 
 from api.models import LogMetadata, UserConfig
 from api.services.logging import LoggingService
 
 
-def is_ollama_selected() -> bool:
-    return os.getenv("LLM") == "ollama"
-
-
-def get_large_model():
-    selected_llm = os.getenv("LLM")
-    if selected_llm == "openai":
-        return ChatOpenAI(model="gpt-4.1")
-    elif selected_llm == "google":
-        return ChatGoogleGenerativeAI(model="gemini-2.0-flash")
-    else:
-        return ChatOllama(model=os.getenv("OLLAMA_MODEL"), temperature=0.8)
-
-
-def get_small_model():
-    selected_llm = os.getenv("LLM")
-    if selected_llm == "openai":
-        return ChatOpenAI(model="gpt-4.1-mini")
-    elif selected_llm == "google":
-        return ChatGoogleGenerativeAI(model="gemini-2.0-flash")
-    else:
-        return ChatOllama(model=os.getenv("OLLAMA_MODEL"), temperature=0.8)
-
-
-def get_nano_model():
-    selected_llm = os.getenv("LLM")
-    if selected_llm == "openai":
-        return ChatOpenAI(model="gpt-4.1-nano")
-    elif selected_llm == "google":
-        return ChatGoogleGenerativeAI(model="gemini-2.0-flash")
-    else:
-        return ChatOllama(model=os.getenv("OLLAMA_MODEL"), temperature=0.8)
-
-
 def get_presentation_dir(presentation_id: str) -> str:
     presentation_dir = os.path.join(os.getenv("APP_DATA_DIRECTORY"), presentation_id)
     os.makedirs(presentation_dir, exist_ok=True)