[ENG-857] Filter large image messages (#190)

* Filter out large images from logs

* Filter out single images larger than 1MB

* Formatting

* Bump image size limit to 4MB

* Update magentic test to assert completion id in the platform

---------

Co-authored-by: Kim Tran <ksprtran@gmail.com>

Author: nqn

log10/load.py

@@ -473,6 +473,36 @@ def _process_chunk(self, chunk):
                 logger.error(f"Failed to insert in log10: {self.partial_log_row} with error {res.text}. Skipping")
 
 
+# Filter large images from messages, and replace with a text message saying "Image too large to display"
+def filter_large_images(messages):
+    for message in messages:
+        # Content may be an array of fragments, of text and images.
+        # If not, it's a single fragment.
+        if isinstance(message.get("content"), list):
+            new_content = []
+            for fragment in message.get("content", ""):
+                if fragment.get("type") == "image_url":
+                    # If image is more than 4MB, replace with a text message
+                    url = fragment.get("image_url", {}).get("url", "")
+                    if url.startswith("data:image"):
+                        if len(url) > 4e6:
+                            new_content.append(
+                                {
+                                    "type": "text",
+                                    "text": "Image too large to capture",
+                                }
+                            )
+                        else:
+                            new_content.append(fragment)
+                    else:
+                        new_content.append(fragment)
+                else:
+                    new_content.append(fragment)
+            message["content"] = new_content
+
+    return messages
+
+
 def flatten_messages(messages):
     flat_messages = []
     for message in messages:
@@ -524,7 +554,7 @@ def _init_log_row(func, *args, **kwargs):
     # We may have to flatten messages from their ChatCompletionMessage with nested ChatCompletionMessageToolCall to json serializable format
     # Rewrite in-place
     if "messages" in kwargs_copy:
-        kwargs_copy["messages"] = flatten_messages(kwargs_copy["messages"])
+        kwargs_copy["messages"] = filter_large_images(flatten_messages(kwargs_copy["messages"]))
 
     # kind and request are set based on the module and qualname
     # request is based on openai schema

tests/test_large_images.py

@@ -0,0 +1,76 @@
+import copy
+
+from log10.load import filter_large_images
+
+
+def test_empty_messages():
+    assert filter_large_images([]) == []
+
+
+# Test for regular messages without fragments i.e. content isn't a list.
+def test_non_fragment_messages():
+    messages = [
+        {"content": "This is a message.", "role": "system"},
+        {"content": "This is another message.", "role": "user"},
+    ]
+    assert filter_large_images(copy.deepcopy(messages)) == messages
+
+
+# Test for a message with a fragment that is not an image.
+def test_non_image_fragment():
+    messages = [
+        {"content": [{"type": "text", "text": "This is a message."}], "role": "system"},
+        {
+            "content": [{"type": "text", "text": "This is another message."}],
+            "role": "user",
+        },
+    ]
+    assert filter_large_images(copy.deepcopy(messages)) == messages
+
+
+def test_small_image_fragment():
+    messages = [
+        {"content": [{"type": "text", "text": "This is a message."}], "role": "system"},
+        {
+            "content": [
+                {
+                    "type": "image_url",
+                    "image_url": {"url": "https://example.com/image.png"},
+                }
+            ],
+            "role": "system",
+        },
+        {"content": [{"type": "text", "text": "This is a message."}], "role": "system"},
+    ]
+    assert filter_large_images(copy.deepcopy(messages)) == messages
+
+
+def test_large_image_fragment():
+    large_string = "a" * int(4e6)
+    before_messages = [
+        {"content": [{"type": "text", "text": "This is a message."}], "role": "system"},
+        {
+            "content": [
+                {
+                    "type": "image_url",
+                    "image_url": {"url": f"data:image/jpeg;base64,{large_string}"},
+                }
+            ],
+            "role": "system",
+        },
+        {"content": [{"type": "text", "text": "This is a message."}], "role": "system"},
+    ]
+    after_messages = [
+        {"content": [{"type": "text", "text": "This is a message."}], "role": "system"},
+        {
+            "content": [
+                {
+                    "type": "text",
+                    "text": "Image too large to capture",
+                },
+            ],
+            "role": "system",
+        },
+        {"content": [{"type": "text", "text": "This is a message."}], "role": "system"},
+    ]
+    assert filter_large_images(before_messages) == after_messages

tests/test_magentic.py

@@ -208,3 +208,4 @@ def _llm() -> str: ...
 
     output = _llm()
     assert isinstance(output, str)
+    _LogAssertion(completion_id=session.last_completion_id(), message_content=output).assert_chat_response()