#!/usr/bin/env python3 """ utils.background_factory ───────────────────────────────────────────────────────────────────────────── Generates professional backgrounds from presets **or** a user-supplied image. Public API ---------- create_professional_background(cfg_or_key, width, height) → np.ndarray (BGR) All lower-case helpers are considered private to this module. """ from __future__ import annotations from pathlib import Path from typing import Dict, Any, List, Tuple, Optional import logging, os, cv2, numpy as np from utils.background_presets import PROFESSIONAL_BACKGROUNDS log = logging.getLogger(__name__) __all__ = ["create_professional_background"] # ──────────────────────────────────────────────────────────────────────────── # Main entry # ──────────────────────────────────────────────────────────────────────────── def create_professional_background( bg_config: Dict[str, Any] | str, width: int, height: int, ) -> np.ndarray: """ Accepts either … • a **key** into PROFESSIONAL_BACKGROUNDS (e.g. "office_modern"), or • a **dict** (typically supplied by UI) that may include: ─ background_choice: "office_modern" ─ custom_path: "/path/to/image.png" ─ OR directly contain {type:"gradient", colors:[…]} Returns **BGR** uint8 image (OpenCV-ready). """ try: # ── Resolve input --------------------------------------------------- choice : str = "minimalist" custom_path : str | None = None direct_style : Dict[str, Any] | None = None if isinstance(bg_config, str): choice = bg_config.lower() elif isinstance(bg_config, dict): choice = bg_config.get("background_choice", bg_config.get("name", "minimalist")).lower() custom_path = bg_config.get("custom_path") if "type" in bg_config and "colors" in bg_config: direct_style = bg_config # full inline style # ── 1) Custom image? ---------------------------------------------- if custom_path and os.path.exists(custom_path): img = cv2.imread(custom_path, cv2.IMREAD_COLOR) if img is not None: img_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB) fitted = _fit_image_letterbox(img_rgb, width, height, fill=(32,32,32)) return cv2.cvtColor(fitted, cv2.COLOR_RGB2BGR) log.warning(f"Custom-background read failed: {custom_path}") # ── 2) Inline dict style? ----------------------------------------- if direct_style: if direct_style["type"] == "color": bg = _create_solid_background(direct_style, width, height) else: # gradient / image bg = _create_gradient_background(direct_style, width, height) return _apply_bg_adjustments(bg, direct_style) # ── 3) Preset dict lookup ----------------------------------------- preset = PROFESSIONAL_BACKGROUNDS.get(choice, PROFESSIONAL_BACKGROUNDS["minimalist"]) if preset["type"] == "color": bg = _create_solid_background(preset, width, height) elif preset["type"] == "image": path = Path(preset["path"]) if path.exists(): img_bgr = cv2.imread(str(path), cv2.IMREAD_COLOR) if img_bgr is not None: return cv2.resize(img_bgr, (width, height), interpolation=cv2.INTER_LANCZOS4) log.warning(f"Preset image not found: {path}; falling back to gradient") bg = _create_gradient_background( {**preset, "type": "gradient", "colors": ["#3a3a3a", "#2e2e2e"]}, width, height ) else: # gradient bg = _create_gradient_background(preset, width, height) return _apply_bg_adjustments(bg, preset) except Exception as e: log.error(f"create_professional_background: {e}") return np.full((height, width, 3), (128,128,128), np.uint8) # ──────────────────────────────────────────────────────────────────────────── # Letter-boxed fit for custom images # ──────────────────────────────────────────────────────────────────────────── def _fit_image_letterbox(img_rgb: np.ndarray, dst_w: int, dst_h: int, fill=(32,32,32)) -> np.ndarray: h, w = img_rgb.shape[:2] if h == 0 or w == 0: return np.full((dst_h, dst_w, 3), fill, np.uint8) src_a = w / h dst_a = dst_w / dst_h if src_a > dst_a: new_w, new_h = dst_w, int(dst_w / src_a) else: new_h, new_w = dst_h, int(dst_h * src_a) resized = cv2.resize(img_rgb, (new_w, new_h), interpolation=cv2.INTER_AREA) canvas = np.full((dst_h, dst_w, 3), fill, np.uint8) y0 = (dst_h-new_h)//2; x0 = (dst_w-new_w)//2 canvas[y0:y0+new_h, x0:x0+new_w] = resized return canvas # ──────────────────────────────────────────────────────────────────────────── # Background builders # ──────────────────────────────────────────────────────────────────────────── def _create_solid_background(style: Dict[str,Any], w: int, h: int) -> np.ndarray: clr_hex = style["colors"][0].lstrip("#") rgb = tuple(int(clr_hex[i:i+2],16) for i in (0,2,4)) return np.full((h,w,3), rgb[::-1], np.uint8) # BGR def _create_gradient_background(style: Dict[str,Any], w:int, h:int) -> np.ndarray: cols = [hex.lstrip("#") for hex in style["colors"]] rgbs = [tuple(int(c[i:i+2],16) for i in (0,2,4)) for c in cols] dirn = style.get("direction","vertical") if dirn=="vertical": grad = _grad_vertical(rgbs, w, h) elif dirn=="horizontal": grad = _grad_horizontal(rgbs, w, h) elif dirn=="diagonal": grad = _grad_diagonal(rgbs, w, h) else: grad = _grad_radial(rgbs, w, h, soft=(dirn=="soft_radial")) return cv2.cvtColor(grad, cv2.COLOR_RGB2BGR) # --- gradient helpers ------------------------------------------------------- def _grad_vertical(colors, w, h): g = np.zeros((h, w, 3), np.uint8) for y in range(h): g[y, :] = _interp_multi(colors, y/h) return g def _grad_horizontal(colors, w, h): g = np.zeros((h, w, 3), np.uint8) for x in range(w): g[:, x] = _interp_multi(colors, x/w) return g def _grad_diagonal(colors, w, h): y,x = np.mgrid[0:h, 0:w] prog = np.clip((x+y)/(h+w), 0, 1) g = np.zeros((h,w,3), np.uint8) for c in range(3): g[:,:,c] = _vector_interp(colors, prog, c) return g def _grad_radial(colors, w, h, soft=False): cx, cy = w/2, h/2 maxd = np.hypot(cx, cy) y,x = np.mgrid[0:h, 0:w] prog = np.clip(np.hypot(x-cx, y-cy)/maxd, 0, 1) if soft: prog = prog**0.7 g = np.zeros((h,w,3), np.uint8) for c in range(3): g[:,:,c] = _vector_interp(colors, prog, c) return g def _vector_interp(cols, prog, chan): if len(cols)==1: return np.full_like(prog, cols[0][chan], np.uint8) segs = len(cols)-1 seg_prog = prog*segs idx = np.clip(np.floor(seg_prog).astype(int), 0, segs-1) local = seg_prog - idx start = np.take([c[chan] for c in cols], idx) end = np.take([c[chan] for c in cols[1:]+[cols[-1]]], idx) return (start + (end-start)*local).astype(np.uint8) def _interp_multi(cols, p): # cols length 1..n p ∈[0,1] if len(cols)==1: return cols[0] seg = p*(len(cols)-1) i = int(seg) l = seg - i c1, c2 = cols[i], cols[min(i+1, len(cols)-1)] return tuple(int(c1[c]+(c2[c]-c1[c])*l) for c in range(3)) # ──────────────────────────────────────────────────────────────────────────── # Post-adjust # ──────────────────────────────────────────────────────────────────────────── def _apply_bg_adjustments(bg: np.ndarray, cfg: Dict[str,Any]) -> np.ndarray: bright = cfg.get("brightness",1.0) contrast = cfg.get("contrast",1.0) if bright==1.0 and contrast==1.0: return bg out = bg.astype(np.float32)*contrast*bright return np.clip(out,0,255).astype(np.uint8)