Added ability to log pictures;

Introduced max size of 1024*1024 for images

analytics.py

@@ -2,6 +2,7 @@ import os
 import uuid
 import datetime as dt
 import sys
+from pathlib import Path
 from typing import Optional
 
 from supabase import create_client, Client
@@ -40,7 +41,18 @@ class AnalyticsLogger:
             print(f"[AnalyticsLogger] Failed to start session: {e}", file=sys.stderr)
             raise e
 
-    def log_interaction(self, user: str, answer: str, ts_iso: Optional[str] = None) -> None:
+    def _upload_image(self, image_path: str) -> Optional[str]:
+        try:
+            with open(image_path, "rb") as img_file:
+                image_name = f'{uuid.uuid4()}{Path(image_path).suffix}'
+                response = self.client.storage.from_("Images").upload(image_name, img_file, {"cacheControl": "3600", "upsert": "true"})
+
+                return response.full_path
+        except:
+            print(f"[AnalyticsLogger] Failed to upload image: {response['error']}", file=sys.stderr)
+            return None
+
+    def log_interaction(self, user: str | tuple[str, str], answer: str, ts_iso: Optional[str] = None) -> None:
         """
         Inserts a single chat interaction.
         """
@@ -48,12 +60,20 @@ class AnalyticsLogger:
             raise ValueError("Session not started. Call start_session() first.")
         session_id = self.session_id
 
+        image_handle: str | None = None
+
+        if isinstance(user, tuple): # (image_path, user_name)
+            image, user = user
+
+            image_handle = self._upload_image(image)
+
         chat_payload = {
             "id": str(uuid.uuid4()),
             "session_id": session_id,
             "timestamp": ts_iso or _utc_now_iso(),
             "user": user,
             "answer": answer,
+            "user_image_path": image_handle,
         }
         try:
             self.client.table("Chats").insert(chat_payload).execute()

app.py

@@ -11,6 +11,7 @@ import threading
 import spaces
 import gradio as gr
 import torch
+from PIL.Image import Image
 from transformers import AutoModelForCausalLM, AutoProcessor, AutoTokenizer, TextIteratorStreamer
 from analytics import AnalyticsLogger
 from kernels import get_kernel
@@ -36,6 +37,8 @@ MAX_TOKENS = 4096
 TEMPERATURE = 0.7
 TOP_P = 0.95
 
+IMAGE_MAX_SIZE = 1024
+
 logger = AnalyticsLogger()
 
 def _begin_analytics_session():
@@ -66,7 +69,7 @@ def load_model():
 model, tokenizer, processor, device = load_model()
 
 
-def user(user_message, image_data, history: list):
+def user(user_message, image_data: Image, history: list):
     """Format user message with optional image."""
     import io
 
@@ -78,6 +81,8 @@ def user(user_message, image_data, history: list):
 
     # If we have an image, save it to temp file for Gradio display
     if image_data is not None:
+        image_data.thumbnail((IMAGE_MAX_SIZE, IMAGE_MAX_SIZE))
+
         # Save to temp file for Gradio display
         fd, tmp_path = tempfile.mkstemp(suffix=".jpg")
         os.close(fd)
@@ -121,17 +126,14 @@ def append_example_message(x: gr.SelectData, history):
     return history
 
 
-def _extract_text_from_content(content: Any) -> str:
+def _extract_text_from_content(content: Any) -> str | tuple[str, str]:
     """Extract text from message content for logging."""
     if isinstance(content, str):
         return content
-    if isinstance(content, list):
-        text_parts = []
-        for item in content:
-            if isinstance(item, dict) and item.get("type") == "text":
-                text_parts.append(item.get("text", ""))
-        return " ".join(text_parts) if text_parts else "[Image]"
-    return str(content)
+    if isinstance(content, tuple) and len(content) == 2:
+        return content # (image_path, user_text)
+
+    raise ValueError(f"Unsupported content type for text extraction: {content}")
 
 
 def _clean_history_for_display(history: list[dict[str, Any]]) -> list[dict[str, Any]]: