Initial Scheenshot tool

Author: yorevs

dependencies.hspd

@@ -52,6 +52,7 @@ package: SpeechRecognition, version: 3.10.4, mode: ge
 
 /* Video */
 package: opencv-python, version: 4.10.0.84, mode: ge
+package: pyautogui, version: 0.9.54, mode: ge
 
 /* Pre-Trained and AI */
 package: torch, version: 2.2.0, mode: ge

src/demo/others/screenshot_demo.py

@@ -1,18 +1,8 @@
-import pyautogui
-
-
-def take_screenshot(file_path):
-    # Take a screenshot using pyautogui
-    screenshot = pyautogui.screenshot()
-
-    # Check if saving as JPEG and convert to RGB if necessary
-    if file_path.lower().endswith(".jpg") or file_path.lower().endswith(".jpeg"):
-        screenshot = screenshot.convert("RGB")
-
-    # Save the screenshot to the specified file path
-    screenshot.save(file_path)
-    print(f"Screenshot saved at {file_path}")
+from askai.core.features.router.tools.vision import take_screenshot
+from hspylib.core.tools.commons import sysout
 
+from utils import init_context
 
 if __name__ == "__main__":
-    take_screenshot("gabiroba.jpeg")
+    init_context(log_name="camera-demo")
+    sysout(take_screenshot("gabiroba.jpeg"))

src/demo/utils.py

@@ -34,7 +34,7 @@ def init_context(
     KeyboardInput.preload_history(cache.load_input_history(commands()))
     shared.create_engine(engine_name=engine_name, model_name=model_name)
     shared.create_context(context_size)
-    events.reply.subscribe(cb_event_handler=lambda ev: display_text(ev.args.message))
+    events.reply.subscribe(cb_event_handler=lambda ev: display_text(ev.args.reply))
     atexit.register(cache.save_input_history)
 
 

src/main/askai/core/askai_messages.py

@@ -117,14 +117,17 @@ def press_esc_enter(self) -> str:
     def device_switch(self, device_info: str) -> str:
         return f"\nSwitching to Audio Input device: `{device_info}`\n"
 
+    # Debug messages
+
     def photo_captured(self, photo: str) -> str:
         return f"~~[DEBUG]~~ WebCam photo captured: `{photo}`"
 
+    def screenshot_saved(self, screenshot: str) -> str:
+        return f"~~[DEBUG]~~ Screenshot saved: `{screenshot}`"
+
     def executing(self, command_line: str) -> str:
         return f"~~[DEBUG]~~ Executing: `{command_line}`…"
 
-    # Debug messages
-
     def analysis(self, result: str) -> str:
         return f"~~[DEBUG]~~ Analysis result => {result}"
 

src/main/askai/core/features/router/tools/vision.py

@@ -1,3 +1,9 @@
+import os
+from textwrap import indent
+
+import pyautogui
+import torch
+from PIL import Image
 from askai.core.askai_events import events
 from askai.core.askai_messages import msg
 from askai.core.component.cache_service import PICTURE_DIR
@@ -9,13 +15,9 @@
 from hspylib.core.config.path_object import PathObject
 from hspylib.core.enums.enumeration import Enumeration
 from hspylib.core.metaclass.classpath import AnyPath
-from PIL import Image
-from textwrap import indent
+from hspylib.core.preconditions import check_argument
 from transformers import BlipForConditionalGeneration, BlipProcessor
 
-import os
-import torch
-
 
 class HFModel(Enumeration):
     """Available Hugging Face models"""
@@ -111,3 +113,24 @@ def parse_caption(image_caption: str) -> str:
         )  # fmt: on
 
     return msg.no_caption()
+
+
+def take_screenshot(path_name: AnyPath, load_dir: AnyPath | None = None) -> str:
+    """Takes a screenshot and saves it to the specified path.
+    :param path_name: The path where the screenshot will be saved.
+    :param load_dir: Optional directory to save the screenshot.
+    :return: The path to the saved screenshot.
+    """
+
+    posix_path: PathObject = PathObject.of(path_name)
+    check_argument(os.path.exists(posix_path.abs_dir))
+    screenshot = pyautogui.screenshot()
+    _, ext = os.path.splitext(posix_path.filename)
+    if ext.casefold().endswith((".jpg", ".jpeg")):
+        screenshot = screenshot.convert("RGB")
+    final_path: str = os.path.join(load_dir or posix_path.abs_dir or PICTURE_DIR, posix_path.filename)
+    screenshot.save(final_path)
+    events.reply.emit(reply=AIReply.full(msg.screenshot_saved(final_path)))
+    desktop_caption = image_captioner(final_path, load_dir)
+
+    return desktop_caption