feat: basic image upload to AI model (still no context window)

Dockerfile

@@ -17,6 +17,7 @@ COPY aw.py .
 COPY LICENSE .
 
 ENV BOT_TOKEN=""
+ENV GOOGLE_API_KEY=""
 
 # Where the database will be stored
 ENV BOT_DATA_DIR=""

bot/ai/handle_request.py

@@ -1,20 +1,46 @@
 import os
+import requests
+from google.genai import types
 from google import genai
 
 API_KEY = os.getenv("GOOGLE_API_KEY")
 
 
-async def forward_to_google_api(prompt):
+async def forward_to_google_api(prompt, image_object=None):
     """
-    Forwards the message content and optional image URL to a Google API.
+    Forwards the message content and optional image object to a Google API.
 
     Args:
         prompt (discord.Message): The message object to forward.
+        image_object (tuple, optional): A tuple containing the image URL and its MIME type (e.g., ("url", "image/jpeg")).
     """
+    if not API_KEY:
+        await prompt.reply(
+            "Google API key is not set. Please contact the administrator.",
+            mention_author=True,
+        )
+        return
+
     client = genai.Client(api_key=API_KEY)
 
+    input = [prompt.content]
+    if image_object:
+        try:
+            image_url, mime_type = image_object
+            image = requests.get(image_url)
+            image.raise_for_status()
+            input.append(types.Part.from_bytes(data=image.content, mime_type=mime_type))
+        except requests.RequestException:
+            await prompt.reply(f"Failed to fetch the image", mention_author=True)
+            return
+
     response = client.models.generate_content(
-        model="gemini-2.0-flash", contents=prompt.content
+        model="gemini-2.0-flash",
+        contents=input,
+        config=types.GenerateContentConfig(
+            max_output_tokens=400,
+            system_instruction="You are a Discord chat bot named 'AlterWare' who helps users. You should limit your answers to be less than 2000 characters.",
+        ),
     )
 
     await prompt.reply(

bot/events_handlers/message_events.py

@@ -317,7 +317,19 @@ async def handle_message(message, bot):
         staff_role = message.guild.get_role(STAFF_ROLE_ID)
         member = message.guild.get_member(message.author.id)
         if staff_role in member.roles:
-            await forward_to_google_api(message)
+            image_object = None
+
+            for attachment in message.attachments:
+                if attachment.filename.lower().endswith(
+                    ".jpg"
+                ) or attachment.filename.lower().endswith(".jpeg"):
+                    image_object = (attachment.url, "image/jpeg")
+                    break
+                elif attachment.filename.lower().endswith(".png"):
+                    image_object = (attachment.url, "image/png")
+                    break
+
+            await forward_to_google_api(message, image_object)
             return
 
     # Too many mentions