Update app.py

app.py

@@ -1,6 +1,7 @@
 import streamlit as st
 import cohere
 import os
+import base64
 
 st.set_page_config(page_title="Cohere Chat", layout="wide")
 
@@ -13,71 +14,13 @@ if not os.path.exists(AI_PFP) or not os.path.exists(USER_PFP):
     st.stop()
 
 model_info = {
-    "c4ai-aya-expanse-8b": {
-        "description": "Aya Expanse is a highly performant 8B multilingual model, designed to rival monolingual performance through innovations in instruction tuning with data arbitrage, preference training, and model merging. Serves 23 languages.",
-        "context": "4K",
-        "output": "4K"
-    },
-    "c4ai-aya-expanse-32b": {
-        "description": "Aya Expanse is a highly performant 32B multilingual model, designed to rival monolingual performance through innovations in instruction tuning with data arbitrage, preference training, and model merging. Serves 23 languages.",
-        "context": "128K",
-        "output": "4K"
-    },
-    "command-a-03-2025": {
-        "description": "Command A is our most performant model to date, excelling at tool use, agents, retrieval augmented generation (RAG), and multilingual use cases. Command A has a context length of 256K, only requires two GPUs to run, and has 150% higher throughput compared to Command R+ 08-2024.",
-        "context": "256K",
-        "output": "8K"
-    },
-    "command-r7b-12-2024": {
-        "description": "command-r7b-12-2024 is a small, fast update delivered in December 2024. It excels at RAG, tool use, agents, and similar tasks requiring complex reasoning and multiple steps.",
-        "context": "128K",
-        "output": "4K"
-    },
-    "command-r-plus-04-2024": {
-        "description": "Command R+ is an instruction-following conversational model that performs language tasks at a higher quality, more reliably, and with a longer context than previous models. It is best suited for complex RAG workflows and multi-step tool use.",
-        "context": "128K",
-        "output": "4K"
-    },
-    "command-r-plus": {
-        "description": "command-r-plus is an alias for command-r-plus-04-2024, so if you use command-r-plus in the API, that's the model you're pointing to.",
-        "context": "128K",
-        "output": "4K"
-    },
-    "command-r-08-2024": {
-        "description": "Updated Command R model from August 2024.",
-        "context": "128K",
-        "output": "4K"
-    },
-    "command-r-03-2024": {
-        "description": "Instruction-following model for code generation, RAG, and agents.",
-        "context": "128K",
-        "output": "4K"
-    },
-    "command-r": {
-        "description": "Alias for command-r-03-2024.",
-        "context": "128K",
-        "output": "4K"
-    },
-    "command": {
-        "description": "Conversational model with long context capabilities.",
-        "context": "4K",
-        "output": "4K"
-    },
-    "command-nightly": {
-        "description": "Experimental nightly build (not for production).",
-        "context": "128K",
-        "output": "4K"
-    },
-    "command-light": {
-        "description": "Faster lightweight version of command.",
-        "context": "4K",
-        "output": "4K"
-    },
-    "command-light-nightly": {
-        "description": "Experimental nightly build of command-light.",
-        "context": "128K",
-        "output": "4K"
-    },
+    "c4ai-aya-expanse-8b": {"description": "Aya Expanse is a highly performant 8B multilingual model, designed to rival monolingual performance through innovations in instruction tuning with data arbitrage, preference training, and model merging. Serves 23 languages.", "context": "4K", "output": "4K"},
+    "c4ai-aya-expanse-32b": {"description": "Aya Expanse is a highly performant 32B multilingual model, designed to rival monolingual performance through innovations in instruction tuning with data arbitrage, preference training, and model merging. Serves 23 languages.", "context": "128K", "output": "4K"},
+    "c4ai-aya-vision-8b": {"description": "Aya Vision is a state-of-the-art multimodal model excelling at a variety of critical benchmarks for language, text, and image capabilities. This 8 billion parameter variant is focused on low latency and best-in-class performance.", "context": "16K", "output": "4K"},
+    "c4ai-aya-vision-32b": {"description": "Aya Vision is a state-of-the-art multimodal model excelling at a variety of critical benchmarks for language, text, and image capabilities. Serves 23 languages. This 32 billion parameter variant is focused on state-of-art multilingual performance.", "context": "16k", "output": "4K"},
+    "command-a-03-2025": {"description": "Command A is our most performant model to date, excelling at tool use, agents, retrieval augmented generation (RAG), and multilingual use cases. Command A has a context length of 256K, only requires two GPUs to run, and has 150% higher throughput compared to Command R+ 08-2024.", "context": "256K", "output": "8K"},
+    "command-r7b-12-2024": {"description": "command-r7b-12-2024 is a small, fast update delivered in December 2024. It excels at RAG, tool use, agents, and similar tasks requiring complex reasoning and multiple steps.", "context": "128K", "output": "4K"},
+    "command-r-plus-04-2024": {"description": "Command R+ is an instruction-following conversational model that performs language tasks at a higher quality, more reliably, and with a longer context than previous models. It is best suited for complex RAG workflows and multi-step tool use.", "context": "128K", "output": "4K"},
 }
 
 with st.sidebar:
@@ -86,12 +29,10 @@ with st.sidebar:
     st.title("Settings")
     api_key = st.text_input("Cohere API Key", type="password")
     selected_model = st.selectbox("Model", options=list(model_info.keys()))
-    
     if st.button("Clear Chat"):
         st.session_state.messages = []
         st.session_state.first_message_sent = False
         st.rerun()
-        
     st.divider()
     st.image(AI_PFP, width=60)
     st.subheader(selected_model)
@@ -102,44 +43,49 @@ with st.sidebar:
 
 if "messages" not in st.session_state:
     st.session_state.messages = []
-    
 if "first_message_sent" not in st.session_state:
     st.session_state.first_message_sent = False
 
 if not st.session_state.first_message_sent:
     st.markdown("<h1 style='text-align: center; color: #4a4a4a; margin-top: 100px;'>How can Cohere help you today?</h1>", unsafe_allow_html=True)
-
 for msg in st.session_state.messages:
     with st.chat_message(msg["role"], avatar=USER_PFP if msg["role"] == "user" else AI_PFP):
         st.markdown(msg["content"])
 
-if prompt := st.chat_input("Message..."):
+col1, col2 = st.columns([1, 4])
+with col1:
+    if selected_model.startswith("c4ai-aya-vision"):
+        uploaded = st.file_uploader("Upload image", type=["png", "jpg", "jpeg"], key="image_uploader")
+    else:
+        uploaded = None
+with col2:
+    prompt = st.chat_input("Message...")
+
+if prompt:
     if not api_key:
         st.error("API key required")
         st.stop()
-
     st.session_state.first_message_sent = True
-    
     st.session_state.messages.append({"role": "user", "content": prompt})
     with st.chat_message("user", avatar=USER_PFP):
         st.markdown(prompt)
 
     try:
         co = cohere.ClientV2(api_key)
+        user_message = [{"type": "text", "text": prompt}]
+        if uploaded:
+            raw = uploaded.read()
+            b64 = base64.b64encode(raw).decode("utf-8")
+            data_url = f"data:image/jpeg;base64,{b64}"
+            user_message.append({"type": "image_url", "image_url": {"url": data_url}})
+        response = co.chat(
+            model=selected_model,
+            messages=[{"role": "user", "content": user_message}]
+        )
+        content_items = response.message.content
+        reply = "".join(getattr(item, 'text', '') for item in content_items)
         with st.chat_message("assistant", avatar=AI_PFP):
-            response = co.chat(
-                model=selected_model,
-                messages=st.session_state.messages
-            )
-            if hasattr(response, "message") and hasattr(response.message, "content"):
-                content_items = response.message.content
-                reply = "".join(getattr(item, 'text', '') for item in content_items)
-            else:
-                st.write(response)
-                reply = "❗️Couldn't extract reply from the Cohere response."
             st.markdown(reply)
-
         st.session_state.messages.append({"role": "assistant", "content": reply})
-
     except Exception as e:
-        st.error(f"Error: {str(e)}")
+        st.error(f"Error: {str(e)}")