Initial commit

app.py

@@ -0,0 +1,114 @@
+import io
+import os
+import re
+from tempfile import NamedTemporaryFile
+
+import replicate
+import streamlit as st
+from PIL import Image
+from langchain.callbacks.base import BaseCallbackHandler
+from langchain.chat_models import ChatOpenAI
+from langchain.schema import ChatMessage
+from dotenv import load_dotenv
+
+load_dotenv()
+
+MODEL_NAME = os.getenv("MODEL_NAME")
+os.environ["REPLICATE_API_TOKEN"] = os.getenv("REPLICATE_API_TOKEN")
+openai_api_key = os.getenv("OPENAI_API_KEY")
+
+SYSTEM_PROMPT = (
+    "You're a chef and an helpful assistant who suggests delicious recipes that includes foods which are "
+    "nutritionally beneficial, easy & not time consuming, healthy yet economical."
+)
+
+
+class StreamHandler(BaseCallbackHandler):
+    def __init__(self, container, initial_text=""):
+        self.container = container
+        self.text = initial_text
+
+    def on_llm_new_token(self, token: str, **kwargs) -> None:
+        self.text += token
+        self.container.markdown(self.text)
+
+
+st.title('Chef:green[GPT] :cook:')
+
+
+with st.sidebar:
+    st.title("Add grocery items")
+
+    uploaded_files = st.file_uploader(
+        label="Upload image of each item",
+        key="files_upload",
+        accept_multiple_files=True,
+    )
+
+    cook = st.button("Let AI cook!")
+
+    if cook:
+        for uploaded_file in uploaded_files:
+            bytes_data = uploaded_file.getvalue()
+            image = Image.open(io.BytesIO(bytes_data))
+            st.image(image, width=100)
+
+            with NamedTemporaryFile(suffix=".jpg") as temp_file:
+                image.save(temp_file.name)
+                output = replicate.run(
+                    MODEL_NAME,
+                    input={"image": open(temp_file.name, "rb")},
+                )
+                pattern = r"Caption:\s(.*)"
+                match = re.search(pattern, output).group(1)
+
+                if "groceries" not in st.session_state:
+                    st.session_state["groceries"] = ""
+                st.session_state["groceries"] += f"{match},"
+
+
+if "messages" not in st.session_state:
+    st.session_state["messages"] = [
+        ChatMessage(role="system", content=SYSTEM_PROMPT),
+        ChatMessage(role="assistant", content="How can I help you?"),
+    ]
+
+for msg in st.session_state.messages:
+    st.chat_message(msg.role).write(msg.content)
+
+if prompt := st.chat_input():
+    st.session_state.messages.append(ChatMessage(role="user", content=prompt))
+    st.chat_message("user").write(prompt)
+
+    if not openai_api_key:
+        st.info("Please add your OpenAI API key to continue.")
+        st.stop()
+
+    with st.chat_message("assistant"):
+        stream_handler = StreamHandler(st.empty())
+        llm = ChatOpenAI(
+            openai_api_key=openai_api_key, streaming=True, callbacks=[stream_handler]
+        )
+        response = llm(st.session_state.messages)
+        st.session_state.messages.append(
+            ChatMessage(role="assistant", content=response.content)
+        )
+
+if "groceries" in st.session_state and st.session_state["groceries"]:
+    st.session_state.messages.append(
+        ChatMessage(
+            role="user",
+            content=f"Using these items: `{st.session_state['groceries']}` "
+            f"generate a recipe.",
+        )
+    )
+    with st.chat_message("assistant"):
+        stream_handler = StreamHandler(st.empty())
+        llm = ChatOpenAI(
+            openai_api_key=openai_api_key, streaming=True, callbacks=[stream_handler]
+        )
+        response = llm(st.session_state.messages)
+        st.session_state.messages.append(
+            ChatMessage(role="assistant", content=response.content)
+        )
+        st.session_state["groceries"] = ""

env.example

@@ -0,0 +1,3 @@
+MODEL_NAME=salesforce/blip:2e1dddc8621f72155f24cf2e0adbde548458d3cab9f00c0139eea840d0ac4746
+REPLICATE_API_TOKEN=
+OPENAI_API_KEY=

requirements.txt

@@ -0,0 +1,4 @@
+replicate
+streamlit
+langchain
+python-dotenv