Load and initialize the caption model.

Load and initialize the YOLO model.

Convert an agent ToolResult to an API ToolResultBlockParam.

With the assumption that images are screenshots that are of diminishing value as the conversation progresses, remove all but the final `image

With the assumption that images are screenshots that are of diminishing value as the conversation progresses, remove all but the final `image

With the assumption that images are screenshots that are of diminishing value as the conversation progresses, remove all but the final `image

Setup directories needed for one system run.

Add recent info to memory.

Align coordinates for nearly horizontal and nearly vertical walls. For walls that are nearly horizontal, the average y coordinate (round

Draws bounding boxes on the frame using the detections provided. Args: scene (np.ndarray): The image on which the boundi

Visualize a list of walls (and optionally openings) for debugging. Parameters: step_name: A string title for the plot. walls

Extract and parse the floorplan room data. Args: text (str): The input text containing floorplan data Returns:

Finds the largest changed region between two screenshots and writes an image that shows only that region (everything else is white).

Return the CLI argument parser.

check overlap of text and background detection box, and get_optimal_label_pos, pos: str, position of the text, must be one of 'top left', 't

Return width and height of the screen

Extract the wall ID (e.g., Wall1) from the wall string.

Extract the wall ID (e.g., Wall1) from the wall string.

Handle file uploads and store them in the upload directory

Resize an image using PIL. Expects the image in numpy array format.

Convert an indexed image to an RGB image using the given color map.

Initialize session state variables

Check if two points are close enough to be considered the same.

Computes the intersection point between two line segments p1-p2 and p3-p4. Returns the intersection point as integers if it lies stri

Load environment-specific configuration from a JSON file.

Load the persisted Chroma DB; raise if not present.

Determine and run the environment‑specific runner module.

Map the entire floorplan to the new bounding box. Args: floorplan_data: The original floorplan data (dictionary) origina

Merge walls that are collinear and adjacent, overlapping, or nearly touching.

Pad the screenshot to 16:10 aspect ratio, when the aspect ratio is not 16:10.

Extract start and end coordinates from a wall string like 'Wall1: (x1, y1) to (x2, y2)'.

Extract start and end coordinates from a wall string like 'Wall1: (x1, y1) to (x2, y2)'.

Use huggingface model to replace the original model

Processes the floorplan image with step-by-step visualization.

Project point P onto the infinite line defined by points A and B. Returns the projected point.

Project point P onto the line segment AB. If the projection falls outside AB, return the closest endpoint.

Interactively obtain a task description.

POST the payload to the HF endpoint, retrying automatically on network problems or >=500 server errors. Parameters ---------- pa

ocr_bbox format: [{'type': 'text', 'bbox':[x,y], 'interactivity':False, 'content':str }, ...] boxes format: [{'type': 'icon', 'bbox':[x,y], '

Compute the intersection point between two line segments AB and CD. Returns the intersection point as (x, y) if they intersect; otherwise, re

Recursively convert non-serializable items in the dictionary.