included precommit.

.pre-commit-config.yaml

@@ -0,0 +1,49 @@
+# .pre-commit-config.yaml
+
+repos:
+  - repo: https://github.com/pre-commit/pre-commit-hooks
+    rev: v4.5.0
+    hooks:
+      - id: trailing-whitespace
+      - id: end-of-file-fixer
+      - id: check-yaml
+      - id: check-added-large-files
+
+  - repo: https://github.com/psf/black
+    rev: 24.4.2  # Always check for the latest stable version
+    hooks:
+      - id: black
+        language_version: python3.10  # Match your project's Python version
+        args: ["--line-length", "120"]
+
+  - repo: https://github.com/PyCQA/flake8
+    rev: 7.0.0
+    hooks:
+      - id: flake8
+        additional_dependencies: ['flake8-bugbear']
+        args: [--max-line-length=120, "--ignore", "E203"]
+
+  - repo: https://github.com/pre-commit/mirrors-mypy
+    rev: v1.10.0
+    hooks:
+      - id: mypy
+        args: ["--ignore-missing-imports", "--install-types", "--non-interactive"]
+
+  - repo: https://github.com/PyCQA/isort
+    rev: 5.13.2
+    hooks:
+      - id: isort
+
+  - repo: https://github.com/PyCQA/bandit
+    rev: 1.7.7
+    hooks:
+      - id: bandit
+        args: ["-r", "src"]  # Adjust if your code is in a different directory
+
+  - repo: https://github.com/asottile/pyupgrade
+    rev: v3.15.0
+    hooks:
+      - id: pyupgrade
+        args: ["--py39-plus"]  # Set based on your minimum supported version
+
+default_stages: [pre-commit]

README.md

@@ -10,4 +10,4 @@ Upload an image. Agents apply multiple filters and pick the best one based on qu
 
 ### Credits
 
-The image in the repo originates from [toa-heftiba-Xmn-QXsVL4k-unsplash](https://unsplash.com/@heftiba)
+The image in the repo originates from [toa-heftiba-Xmn-QXsVL4k-unsplash](https://unsplash.com/@heftiba)

agents.py

@@ -1,14 +1,28 @@
-from smolagents import CodeAgent, HfApiModel, tool
-from filters import *
 import os
 
+from smolagents import CodeAgent, HfApiModel
+
+import filters as flt
+
 HUGGING_FACE_TOKEN = os.environ["HUGGING_FACE_TOKEN"]
 
 # Initialize the model and agent
 model = HfApiModel(token=HUGGING_FACE_TOKEN)
-agent = CodeAgent(tools=[adjust_contrast, load_image_as_bgr, save_image], model=model)
+picture_operator = CodeAgent(
+    tools=[flt.adjust_contrast, flt.load_image_as_bgr, flt.save_image],
+    model=model,
+    name="Picture Operator",
+    description=(
+        "Performs operations on images, such as adjusting contrast, loading images, and saving them. "
+        "Give it your query as an argument, as well as the path to the image and the output path."
+    ),
+)
 
 # Run the agent
-agent.run("Adjust the contrast of the image in image_path by a factor of 1.5. Save the image to output_path.", 
-          additional_args={'image_path': 'toa-heftiba-Xmn-QXsVL4k-unsplash.jpg',
-                           "output_path": 'adjusted_image.jpg'})
+picture_operator.run(
+    "Adjust the contrast of the image in image_path by a factor of 1.5. Save the image to output_path.",
+    additional_args={
+        "image_path": "toa-heftiba-Xmn-QXsVL4k-unsplash.jpg",
+        "output_path": "adjusted_image.jpg",
+    },
+)

app.py

@@ -1,22 +1,25 @@
 import gradio as gr
-from filters import apply_filters
+
 from evaluators import evaluate_filters
+from filters import apply_filters
+
 
 def process_image(image):
     filtered_images = apply_filters(image)
     best_index, reasons = evaluate_filters(filtered_images)
     return filtered_images, f"Winner: Filter {best_index + 1} — {reasons[best_index]}"
 
+
 demo = gr.Interface(
     fn=process_image,
     inputs=gr.Image(type="pil"),
     outputs=[
         gr.Gallery(label="Filtered Options"),
-        gr.Textbox(label="Critique Agent's Verdict")
+        gr.Textbox(label="Critique Agent's Verdict"),
     ],
     title="🧠 AI Art Director – Filter Showdown",
-    description="Upload an image and let AI agents apply, evaluate, and pick the best filter."
+    description="Upload an image and let AI agents apply, evaluate, and pick the best filter.",
 )
 
 if __name__ == "__main__":
-    demo.launch(server_name="0.0.0.0", server_port=7860)
+    demo.launch()

evaluators.py

@@ -1,13 +1,15 @@
-import numpy as np
 import cv2
+import numpy as np
+
 
 def score_image(pil_img):
     img = np.array(pil_img.convert("L"))
     variance = cv2.Laplacian(img, cv2.CV_64F).var()
     return variance
 
+
 def evaluate_filters(images):
     scores = [score_image(img) for img in images]
     best_index = int(np.argmax(scores))
     reasons = [f"Sharpness score: {s:.2f}" for s in scores]
-    return best_index, reasons
+    return best_index, reasons

filters.py

@@ -1,9 +1,11 @@
+import os
+import tempfile
+
 import cv2
 import numpy as np
 from PIL import Image
 from smolagents import tool
-import tempfile
-import os
+
 
 def apply_filters(image: np.ndarray) -> list[np.ndarray]:
     """Applies a series of filters to the input image.
@@ -33,6 +35,7 @@ def apply_filters(image: np.ndarray) -> list[np.ndarray]:
 
     return filtered_images
 
+
 @tool
 def adjust_contrast(image: np.ndarray, alpha: float = 1.5) -> np.ndarray:
     """Adjusts the contrast of the image.
@@ -46,6 +49,7 @@ def adjust_contrast(image: np.ndarray, alpha: float = 1.5) -> np.ndarray:
     """
     return cv2.convertScaleAbs(image, alpha=alpha, beta=0)
 
+
 @tool
 def adjust_saturation(image: np.ndarray, saturation_scale: float = 1.0) -> np.ndarray:
     """Adjusts the saturation of the image.
@@ -62,19 +66,21 @@ def adjust_saturation(image: np.ndarray, saturation_scale: float = 1.0) -> np.nd
     hsv_img[:, :, 1] = np.clip(hsv_img[:, :, 1], 0, 255)
     return cv2.cvtColor(hsv_img.astype(np.uint8), cv2.COLOR_HSV2BGR)
 
+
 @tool
 def adjust_exposure(image: np.ndarray, beta: int = 50) -> np.ndarray:
     """Adjusts the exposure (brightness) of the image.
 
     Args:
         image (np.ndarray): Input image in BGR format.
-        beta (int, optional): Brightness control. Positive values increase brightness, negative decrease. Defaults to 50.
+        beta (int, optional): Brightness control. Positive values increase brightness, negative decrease. Defaults 50.
 
     Returns:
         np.ndarray: Exposure adjusted image in BGR format.
     """
     return cv2.convertScaleAbs(image, alpha=1.0, beta=beta)
 
+
 @tool
 def denoise_image(image: np.ndarray, h: int = 10) -> np.ndarray:
     """Denoises the image using Non-local Means Denoising algorithm.
@@ -88,6 +94,7 @@ def denoise_image(image: np.ndarray, h: int = 10) -> np.ndarray:
     """
     return cv2.fastNlMeansDenoisingColored(image, None, h, h, 7, 21)
 
+
 @tool
 def crop_image(image: np.ndarray, x: int, y: int, width: int, height: int) -> np.ndarray:
     """Crops the image to the specified rectangle.
@@ -102,7 +109,8 @@ def crop_image(image: np.ndarray, x: int, y: int, width: int, height: int) -> np
     Returns:
         np.ndarray: Cropped image in BGR format.
     """
-    return image[y:y+height, x:x+width]
+    return image[y : y + height, x : x + width]
+
 
 @tool
 def apply_vignette(image: np.ndarray, level: int = 2) -> np.ndarray:
@@ -116,8 +124,8 @@ def apply_vignette(image: np.ndarray, level: int = 2) -> np.ndarray:
         np.ndarray: Image with vignette effect applied in BGR format.
     """
     rows, cols = image.shape[:2]
-    kernel_x = cv2.getGaussianKernel(cols, cols/level)
-    kernel_y = cv2.getGaussianKernel(rows, rows/level)
+    kernel_x = cv2.getGaussianKernel(cols, cols / level)
+    kernel_y = cv2.getGaussianKernel(rows, rows / level)
     kernel = kernel_y * kernel_x.T
     mask = kernel / kernel.max()
     vignette = np.copy(image)
@@ -125,13 +133,14 @@ def apply_vignette(image: np.ndarray, level: int = 2) -> np.ndarray:
         vignette[:, :, i] = vignette[:, :, i] * mask
     return vignette
 
+
 @tool
 def load_image_as_bgr(image_path: str) -> np.ndarray:
     """Loads an image from path and converts it to BGR format for OpenCV.
-    
+
     Args:
         image_path (str): Path to the image file.
-        
+
     Returns:
         np.ndarray: Image in BGR format as numpy array.
     """
@@ -139,7 +148,8 @@ def load_image_as_bgr(image_path: str) -> np.ndarray:
     image_np = np.array(image)
     return cv2.cvtColor(image_np, cv2.COLOR_RGB2BGR)
 
-@tool 
+
+@tool
 def save_image(image: np.ndarray, image_path: str) -> None:
     """Saves an image to the specified path.
 
@@ -153,14 +163,14 @@ def save_image(image: np.ndarray, image_path: str) -> None:
 if __name__ == "__main__":
     # Load a test image
     test_image_np = load_image_as_bgr("test_image.jpg")
-    
+
     # Apply all filters
     filtered_images = apply_filters(test_image_np)
-    
+
     # Save results
     dir = tempfile.mkdtemp()
     for i, filtered_img in enumerate(filtered_images):
         output_path = os.path.join(dir, f"filter_{i+1}.jpg")
         rgb_img = cv2.cvtColor(filtered_img, cv2.COLOR_BGR2RGB)
         Image.fromarray(rgb_img).save(output_path)
-        print(f"Saved {output_path}")
+        print(f"Saved {output_path}")