Feature 1/clip work

app/integrations/diffusers.py

@@ -95,6 +95,7 @@ def image_to_image(diffusion_model_predictor, image_url: str, prompt: str,  user
         return decode_and_show(sdxl_response, filename)
     except Exception as e:
         # Handle general exceptions
+        print(f"Error generating image: {str(e)}")
         raise HTTPException(status_code=500, detail=str(e))
 
 def download_profile_image(url, username):

app/integrations/twitch.py

@@ -31,14 +31,6 @@ async def stop_bot():
     await bot.close()
     bot = None
 
-# def get_user_profile_url(client, username):
-#     users = client.fetch_users(names=[username])
-#     if users:
-#         user = users[0]
-#         return user.profile_image
-#     else:
-#         return None
-
 class TwitchBot(commands.Bot):
     def __init__(self, token, channels):
         print(f"Initializing TwitchBot with channels {channels}")
@@ -79,15 +71,26 @@ async def llm_command(self, ctx):
             await ctx.send(llm.help_text)
             return
 
-        temp_addon = " return the result in under 500 characters as a response to a user in chat" 
-        prompt = " ".join(split[2:])+temp_addon # "explain my next step"
+        temp_addon = "Respond in less than 500 characters as a chatbot would: " 
+        prompt = temp_addon+" ".join(split[2:]) # "explain my next step"
         print(f"Received command: {ctx.message.content}") # !llm mistral explain my next step
         print(f"Prompt: {prompt}") # explain my next step
 
         response = llm.query(llm_type, prompt) # "mistral", "explain my next step"
-        answer = response.split(temp_addon)[-1]
+        answer = response.split(prompt)[-1]
         print(f"Response: {answer}") 
+
+        # Considering the length of username and additional characters in the message
+        max_length = 500 - len(ctx.author.name) - 4 # 4 for " @: "
+        if len(answer) > max_length:
+            answer = answer[:max_length] + ".."
+
         await ctx.send(" @" + ctx.author.name+ ": " + answer)
+        # Splitting the answer into chunks of 500 characters
+        # chunks = [answer[i:i+500] for i in range(0, len(answer), 500)]
+
+        # for chunk in chunks:
+        #     await ctx.send(" @" + ctx.author.name+ ": " + chunk)
 
     @commands.command(name="diffuse")
     async def diffuse_command(self, ctx):
@@ -125,7 +128,11 @@ async def welcome_command(self, ctx):
             "image_url": profile_image_url,
             "title": "welcome image"
         }
-        diffusers.prompt_diffuser_image_to_image(prompt_map)
+        try:
+            diffusers.prompt_diffuser_image_to_image(prompt_map)
+        except Exception as e:
+            print(f"Error: {e}")
+            ctx.send(f"Error: {e}")
 
     @commands.command(name="commands")
     async def commands_command(self, ctx):

app/requirements.txt

@@ -4,4 +4,9 @@ uvicorn
 sagemaker
 boto3
 IPython
-python-dotenv
+python-dotenv
+lanchain-experimental
+pillow
+open_clip_torch
+torch
+matplotlib