CingenAI / core /visual_engine.py
mgbam's picture
Update core/visual_engine.py
b97795f verified
raw
history blame
6.46 kB
# core/visual_engine.py
from PIL import Image, ImageDraw, ImageFont
from moviepy.editor import ImageClip, concatenate_videoclips
import os
class VisualEngine:
def __init__(self, output_dir="generated_media"):
self.output_dir = output_dir
os.makedirs(self.output_dir, exist_ok=True)
try:
# Try to load Arial font. Ensure it's installed in your Docker image.
# Common paths: "arial.ttf", "/usr/share/fonts/truetype/msttcorefonts/Arial.ttf"
# If using a custom font, ensure you COPY it into the Docker image and provide the correct path.
self.font_path = "arial.ttf" # This assumes Arial is in a standard font path or current dir
self.font_size_pil = 24 # For Pillow text drawing
self.font = ImageFont.truetype(self.font_path, self.font_size_pil)
print(f"Successfully loaded font: {self.font_path} with size {self.font_size_pil}")
except IOError:
print(f"Warning: Could not load font '{self.font_path}'. Falling back to default font.")
self.font = ImageFont.load_default()
self.font_size_pil = 10 # Default font is smaller, adjust size for calculations
def _get_text_dimensions(self, text_content, font_obj):
"""
Gets the width and height of a single line of text with the given font.
Returns (width, height).
"""
if hasattr(font_obj, 'getbbox'): # For newer Pillow versions (>=8.0.0)
# getbbox returns (left, top, right, bottom)
bbox = font_obj.getbbox(text_content)
width = bbox[2] - bbox[0]
height = bbox[3] - bbox[1]
return width, height
elif hasattr(font_obj, 'getsize'): # For older Pillow versions
return font_obj.getsize(text_content)
else: # Fallback for very basic font objects (like default font)
# Estimate: average char width * num_chars, fixed height
# This is a rough estimate.
avg_char_width = self.font_size_pil * 0.6
height_estimate = self.font_size_pil * 1.2
return int(len(text_content) * avg_char_width), int(height_estimate)
def create_placeholder_image(self, text_description, filename, size=(1280, 720)):
img = Image.new('RGB', size, color=(73, 109, 137)) # Blueish background
draw = ImageDraw.Draw(img)
padding = 40
max_text_width = size[0] - (2 * padding)
lines = []
words = text_description.split()
current_line = ""
for word in words:
test_line = current_line + word + " "
line_width, _ = self._get_text_dimensions(test_line.strip(), self.font)
if line_width <= max_text_width:
current_line = test_line
else:
lines.append(current_line.strip())
current_line = word + " "
lines.append(current_line.strip()) # Add the last line
# Calculate starting y position to center the text block
# Use the height from the first line as an estimate for line height
_, single_line_height = self._get_text_dimensions("Tg", self.font) # Get height of a typical line
line_spacing_factor = 1.2 # Adjust for spacing between lines
total_text_height = len(lines) * single_line_height * line_spacing_factor
y_text = (size[1] - total_text_height) / 2
if y_text < padding: # Ensure text doesn't start too high if too much text
y_text = padding
for line in lines:
line_width, _ = self._get_text_dimensions(line, self.font)
x_text = (size[0] - line_width) / 2
# Corrected ImageDraw.text() call with explicit keyword arguments
draw.text(
xy=(x_text, y_text),
text=line,
fill=(255, 255, 0), # Yellow text
font=self.font
)
y_text += single_line_height * line_spacing_factor
filepath = os.path.join(self.output_dir, filename)
try:
img.save(filepath)
print(f"Placeholder image saved: {filepath}")
except Exception as e:
print(f"Error saving image {filepath}: {e}")
return None
return filepath
def create_video_from_images(self, image_paths, output_filename="final_video.mp4", fps=1, duration_per_image=3):
if not image_paths:
print("No images provided to create video.")
return None
valid_image_paths = [p for p in image_paths if p and os.path.exists(p)]
if not valid_image_paths:
print("No valid image paths found to create video.")
return None
print(f"Creating video from images: {valid_image_paths}")
try:
clips = [ImageClip(m).set_duration(duration_per_image) for m in valid_image_paths]
if not clips:
print("Could not create ImageClips from the provided image paths.")
return None
video_clip = concatenate_videoclips(clips, method="compose")
output_path = os.path.join(self.output_dir, output_filename)
# Using recommended parameters for write_videofile, especially in headless environments
video_clip.write_videofile(
output_path,
fps=fps,
codec='libx264', # A common and good codec
audio_codec='aac', # If you ever add audio
temp_audiofile='temp-audio.m4a', # For temporary audio processing
remove_temp=True, # Clean up temporary audio file
threads=os.cpu_count() or 2, # Use available CPUs or default to 2
logger=None # Suppress verbose ffmpeg output, use 'bar' for progress
)
print(f"Video successfully created: {output_path}")
return output_path
except Exception as e:
print(f"Error creating video: {e}")
# Print more details if it's an OSError, which can happen with ffmpeg issues
if isinstance(e, OSError):
print("OSError during video creation. This often indicates an issue with ffmpeg.")
print("Ensure ffmpeg is correctly installed and accessible in the environment PATH.")
return None