Vision tool

tools/vision.py

@@ -0,0 +1,57 @@
+import os
+import base64
+from tools.coder import get_current_workspace
+
+MIME_MAP = {
+    ".jpg":  "image/jpeg",
+    ".jpeg": "image/jpeg",
+    ".png":  "image/png",
+    ".gif":  "image/gif",
+    ".webp": "image/webp",
+    ".bmp":  "image/bmp",
+}
+
+MAX_SIZE = 10 * 1024 * 1024  # 10MB
+
+class ImagePayload:
+    def __init__(self, data, mime, path):
+        self.data = data
+        self.mime = mime
+        self.path = path
+
+schema_describe_image = {
+    "type": "function",
+    "function": {
+        "name": "describe_image",
+        "description": "Load an image file and send it to the model for visual analysis. Use this when the user wants to look at, describe, or analyze an image file.",
+        "parameters": {
+            "type": "object",
+            "properties": {
+                "path": {"type": "string", "description": "Path to the image file (absolute or relative to workspace)"}
+            },
+            "required": ["path"]
+        }
+    }
+}
+
+def describe_image(path):
+    full_path = os.path.join(get_current_workspace(), path) if not os.path.isabs(path) else path
+
+    if not os.path.isfile(full_path):
+        return f"Error: file not found: {full_path}"
+
+    ext = os.path.splitext(full_path)[1].lower()
+    if ext not in MIME_MAP:
+        return f"Error: unsupported image format '{ext}'. Supported: {', '.join(MIME_MAP.keys())}"
+
+    size = os.path.getsize(full_path)
+    if size > MAX_SIZE:
+        return f"Error: image too large ({size / 1024 / 1024:.1f}MB). Max: {MAX_SIZE / 1024 / 1024:.0f}MB"
+
+    with open(full_path, "rb") as f:
+        raw = f.read()
+
+    encoded = base64.b64encode(raw).decode("ascii")
+    mime = MIME_MAP[ext]
+
+    return ImagePayload(data=encoded, mime=mime, path=full_path)