Support for sending images into OpenAI chat API (#4827)

Author: kabachuha

extensions/multimodal/README.md

@@ -67,8 +67,56 @@ This extension uses the following parameters (from `settings.json`):
 
 ## Usage through API
 
+### Chat completions endpoint
+
+#### With an image URL
+
+```shell
+curl http://127.0.0.1:5000/v1/chat/completions \
+  -H "Content-Type: application/json" \
+  -d '{
+    "messages": [
+      {
+        "role": "user",
+        "image_url": "https://avatars.githubusercontent.com/u/112222186?v=4"
+      },
+      {
+        "role": "user",
+        "content": "What is unusual about this image?"
+      }
+    ]
+  }'
+```
+
+#### With a Base64 image
+
+```python
+import base64
+import json
+import requests
+
+img = open('image.jpg', 'rb')
+img_bytes = img.read()
+img_base64 = base64.b64encode(img_bytes).decode('utf-8')
+data = { "messages": [
+        {
+            "role": "user",
+            "image_url": f"data:image/jpeg;base64,{img_base64}"
+        },
+        {
+            "role": "user",
+            "content": "what is unusual about this image?"
+        }
+    ]
+}
+response = requests.post('http://127.0.0.1:5000/v1/chat/completions', json=data)
+print(response.text)
+```
+
 You can run the multimodal inference through API, by inputting the images to prompt. Images are embedded like so: `f'<img src="data:image/jpeg;base64,{img_str}">'`, where `img_str` is base-64 jpeg data. Note that you will need to launch `server.py` with the arguments `--api --extensions multimodal`. 
 
+### Completions endpoint
+
 Python example:
 
 ```Python

extensions/openai/completions.py

@@ -1,10 +1,15 @@
+import base64
 import copy
+import re
 import time
 from collections import deque
+from io import BytesIO
 
+import requests
 import tiktoken
 import torch
 import torch.nn.functional as F
+from PIL import Image
 from transformers import LogitsProcessor, LogitsProcessorList
 
 from extensions.openai.errors import InvalidRequestError
@@ -140,7 +145,25 @@ def convert_history(history):
     system_message = ""
 
     for entry in history:
-        content = entry["content"]
+        if "image_url" in entry:
+            image_url = entry['image_url']
+            if "base64" in image_url:
+                image_url = re.sub('^data:image/.+;base64,', '', image_url)
+                img = Image.open(BytesIO(base64.b64decode(image_url)))
+            else:
+                try:
+                    my_res = requests.get(image_url)
+                    img = Image.open(BytesIO(my_res.content))
+                except Exception:
+                    raise 'Image cannot be loaded from the URL!'
+
+            buffered = BytesIO()
+            img.save(buffered, format="JPEG")
+            img_str = base64.b64encode(buffered.getvalue()).decode('utf-8')
+            content = f'<img src="data:image/jpeg;base64,{img_str}">'
+        else:
+            content = entry["content"]
+
         role = entry["role"]
 
         if role == "user":
@@ -182,7 +205,8 @@ def chat_completions_common(body: dict, is_legacy: bool = False, stream=False) -
             raise InvalidRequestError(message="messages: missing role", param='messages')
         elif m['role'] == 'function':
             raise InvalidRequestError(message="role: function is not supported.", param='messages')
-        if 'content' not in m:
+
+        if 'content' not in m and "image_url" not in m:
             raise InvalidRequestError(message="messages: missing content", param='messages')
 
     # Chat Completions