Update core/visual_engine.py
Browse files- core/visual_engine.py +190 -148
core/visual_engine.py
CHANGED
@@ -4,7 +4,7 @@ import base64
|
|
4 |
import mimetypes
|
5 |
import numpy as np
|
6 |
import os
|
7 |
-
import openai
|
8 |
import requests
|
9 |
import io
|
10 |
import time
|
@@ -15,17 +15,17 @@ from moviepy.editor import (ImageClip, VideoFileClip, concatenate_videoclips, Te
|
|
15 |
CompositeVideoClip, AudioFileClip)
|
16 |
import moviepy.video.fx.all as vfx
|
17 |
|
18 |
-
try:
|
19 |
-
if hasattr(Image, 'Resampling') and hasattr(Image.Resampling, 'LANCZOS'):
|
20 |
if not hasattr(Image, 'ANTIALIAS'): Image.ANTIALIAS = Image.Resampling.LANCZOS
|
21 |
-
elif hasattr(Image, 'LANCZOS'):
|
22 |
if not hasattr(Image, 'ANTIALIAS'): Image.ANTIALIAS = Image.LANCZOS
|
23 |
elif not hasattr(Image, 'ANTIALIAS'):
|
24 |
print("WARNING: Pillow version lacks common Resampling or ANTIALIAS. MoviePy effects might fail.")
|
25 |
except Exception as e_mp: print(f"WARNING: ANTIALIAS monkey-patch error: {e_mp}")
|
26 |
|
27 |
logger = logging.getLogger(__name__)
|
28 |
-
# logger.setLevel(logging.DEBUG)
|
29 |
|
30 |
ELEVENLABS_CLIENT_IMPORTED = False; ElevenLabsAPIClient = None; Voice = None; VoiceSettings = None
|
31 |
try:
|
@@ -42,163 +42,199 @@ try:
|
|
42 |
logger.info("RunwayML SDK imported.")
|
43 |
except Exception as e_rwy_imp: logger.warning(f"RunwayML SDK import failed: {e_rwy_imp}. RunwayML disabled.")
|
44 |
|
|
|
45 |
class VisualEngine:
|
46 |
DEFAULT_FONT_SIZE_PIL = 10; PREFERRED_FONT_SIZE_PIL = 20
|
47 |
VIDEO_OVERLAY_FONT_SIZE = 30; VIDEO_OVERLAY_FONT_COLOR = 'white'
|
48 |
DEFAULT_MOVIEPY_FONT = 'DejaVu-Sans-Bold'; PREFERRED_MOVIEPY_FONT = 'Liberation-Sans-Bold'
|
49 |
|
|
|
50 |
def __init__(self, output_dir="temp_cinegen_media", default_elevenlabs_voice_id="Rachel"):
|
51 |
self.output_dir = output_dir
|
52 |
try:
|
53 |
os.makedirs(self.output_dir, exist_ok=True)
|
54 |
logger.info(f"VisualEngine output directory set/ensured: {os.path.abspath(self.output_dir)}")
|
55 |
-
|
56 |
-
|
57 |
-
|
58 |
-
|
59 |
-
|
|
|
|
|
|
|
|
|
60 |
|
61 |
self.font_filename_pil_preference = "DejaVuSans-Bold.ttf"
|
62 |
-
|
63 |
-
self.resolved_font_path_pil = next((p for p in
|
64 |
-
|
|
|
|
|
|
|
|
|
65 |
if self.resolved_font_path_pil:
|
66 |
try:
|
67 |
self.active_font_pil = ImageFont.truetype(self.resolved_font_path_pil, self.PREFERRED_FONT_SIZE_PIL)
|
68 |
self.active_font_size_pil = self.PREFERRED_FONT_SIZE_PIL
|
69 |
logger.info(f"Pillow font loaded: {self.resolved_font_path_pil} at size {self.active_font_size_pil}.")
|
70 |
self.active_moviepy_font_name = 'DejaVu-Sans-Bold' if "dejavu" in self.resolved_font_path_pil.lower() else ('Liberation-Sans-Bold' if "liberation" in self.resolved_font_path_pil.lower() else self.DEFAULT_MOVIEPY_FONT)
|
71 |
-
except IOError as
|
72 |
-
else: logger.warning("Preferred Pillow font not found. Using default.")
|
73 |
|
74 |
self.openai_api_key = None; self.USE_AI_IMAGE_GENERATION = False; self.dalle_model = "dall-e-3"; self.image_size_dalle3 = "1792x1024"
|
75 |
self.video_frame_size = (1280, 720)
|
|
|
76 |
self.elevenlabs_api_key = None; self.USE_ELEVENLABS = False; self.elevenlabs_client_instance = None
|
77 |
-
self.elevenlabs_voice_id = default_elevenlabs_voice_id
|
78 |
logger.info(f"VisualEngine __init__: ElevenLabs Voice ID initially set to: {self.elevenlabs_voice_id}")
|
|
|
79 |
if VoiceSettings and ELEVENLABS_CLIENT_IMPORTED: self.elevenlabs_voice_settings_obj = VoiceSettings(stability=0.60, similarity_boost=0.80, style=0.15, use_speaker_boost=True)
|
80 |
else: self.elevenlabs_voice_settings_obj = None
|
|
|
81 |
self.pexels_api_key = None; self.USE_PEXELS = False
|
82 |
self.runway_api_key = None; self.USE_RUNWAYML = False; self.runway_ml_sdk_client_instance = None
|
|
|
83 |
if RUNWAYML_SDK_IMPORTED and RunwayMLAPIClientClass and os.getenv("RUNWAYML_API_SECRET"):
|
84 |
-
try: self.runway_ml_sdk_client_instance = RunwayMLAPIClientClass(); self.USE_RUNWAYML = True; logger.info("RunwayML Client
|
85 |
-
except Exception as
|
86 |
-
|
|
|
87 |
|
88 |
def set_openai_api_key(self, api_key_value): self.openai_api_key = api_key_value; self.USE_AI_IMAGE_GENERATION = bool(api_key_value); logger.info(f"DALL-E status: {'Ready' if self.USE_AI_IMAGE_GENERATION else 'Disabled'}")
|
|
|
89 |
def set_elevenlabs_api_key(self, api_key_value, voice_id_from_secret=None):
|
90 |
self.elevenlabs_api_key = api_key_value
|
91 |
if voice_id_from_secret: self.elevenlabs_voice_id = voice_id_from_secret; logger.info(f"11L Voice ID updated via set_elevenlabs_api_key to: {self.elevenlabs_voice_id}")
|
|
|
92 |
if api_key_value and ELEVENLABS_CLIENT_IMPORTED and ElevenLabsAPIClient:
|
93 |
-
try: self.elevenlabs_client_instance = ElevenLabsAPIClient(api_key=api_key_value); self.USE_ELEVENLABS = bool(self.elevenlabs_client_instance); logger.info(f"11L Client: {'Ready' if self.USE_ELEVENLABS else 'Failed'} (Using Voice: {self.elevenlabs_voice_id})")
|
94 |
-
except Exception as e_11l_setkey_init: logger.error(f"11L client
|
95 |
-
else: self.USE_ELEVENLABS = False; logger.info(f"11L Disabled (API key not provided or SDK issue).")
|
|
|
96 |
def set_pexels_api_key(self, api_key_value): self.pexels_api_key = api_key_value; self.USE_PEXELS = bool(api_key_value); logger.info(f"Pexels status: {'Ready' if self.USE_PEXELS else 'Disabled'}")
|
|
|
97 |
def set_runway_api_key(self, api_key_value):
|
98 |
self.runway_api_key = api_key_value
|
99 |
if api_key_value:
|
100 |
if RUNWAYML_SDK_IMPORTED and RunwayMLAPIClientClass:
|
101 |
-
if not self.runway_ml_sdk_client_instance:
|
102 |
try:
|
103 |
-
|
104 |
-
if not
|
105 |
self.runway_ml_sdk_client_instance = RunwayMLAPIClientClass(); self.USE_RUNWAYML = True; logger.info("RunwayML Client initialized successfully via set_runway_api_key.")
|
106 |
-
if not
|
107 |
-
except Exception as
|
108 |
else: self.USE_RUNWAYML = True; logger.info("RunwayML Client was already initialized (likely from environment variable). API key stored.")
|
109 |
else: logger.warning("RunwayML SDK not imported. API key stored, but current integration relies on SDK. Service effectively disabled."); self.USE_RUNWAYML = False
|
110 |
else: self.USE_RUNWAYML = False; self.runway_ml_sdk_client_instance = None; logger.info("RunwayML Service Disabled (no API key provided).")
|
111 |
|
112 |
-
|
|
|
|
|
113 |
try:
|
114 |
-
|
115 |
-
if not
|
116 |
-
|
117 |
-
|
118 |
-
|
119 |
-
|
120 |
-
|
121 |
-
|
|
|
|
|
|
|
122 |
|
123 |
-
def _map_resolution_to_runway_ratio(self,
|
124 |
-
|
125 |
-
if
|
126 |
-
logger.warning(f"Res {
|
127 |
|
128 |
-
def _get_text_dimensions(self,
|
129 |
-
|
130 |
-
if not
|
131 |
try:
|
132 |
-
if hasattr(
|
133 |
-
elif hasattr(
|
134 |
-
else: return int(len(
|
135 |
-
except Exception as
|
136 |
|
137 |
-
def _create_placeholder_image_content(self,
|
138 |
-
|
139 |
-
|
140 |
-
|
141 |
-
|
142 |
-
|
143 |
-
|
144 |
-
|
145 |
-
|
146 |
-
|
147 |
-
|
148 |
-
if
|
|
|
149 |
else:
|
150 |
-
if
|
151 |
-
|
152 |
-
if
|
153 |
-
if not
|
154 |
-
|
155 |
-
|
156 |
-
|
157 |
-
elif not
|
158 |
-
_,
|
159 |
-
|
160 |
-
|
161 |
-
for
|
162 |
-
|
163 |
-
if
|
164 |
-
|
165 |
-
try:
|
166 |
-
except Exception as
|
167 |
-
|
168 |
-
if
|
169 |
-
try:
|
170 |
-
except Exception as
|
171 |
-
|
172 |
-
try:
|
173 |
-
except Exception as
|
174 |
|
175 |
-
def _search_pexels_image(self,
|
176 |
if not self.USE_PEXELS or not self.pexels_api_key: return None
|
177 |
-
|
178 |
-
|
179 |
-
|
180 |
-
|
181 |
-
|
182 |
try:
|
183 |
-
logger.info(f"Pexels: Searching for '{
|
184 |
-
|
185 |
-
|
186 |
-
|
187 |
-
|
188 |
-
|
189 |
-
if
|
190 |
-
|
191 |
-
|
192 |
-
if not
|
193 |
-
|
194 |
-
|
195 |
-
if
|
196 |
-
|
197 |
-
else: logger.info(f"Pexels: No photos for '{
|
198 |
-
except requests.exceptions.RequestException as
|
199 |
-
except Exception as
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
200 |
|
201 |
def _generate_video_clip_with_runwayml(self, motion_prompt_rwy, input_img_path_rwy, scene_id_base_fn_rwy, duration_s_rwy=5):
|
|
|
202 |
if not self.USE_RUNWAYML or not self.runway_ml_sdk_client_instance: logger.warning("RunwayML skip: Not enabled/client not init."); return None
|
203 |
if not input_img_path_rwy or not os.path.exists(input_img_path_rwy): logger.error(f"Runway Gen-4 needs input img. Invalid: {input_img_path_rwy}"); return None
|
204 |
img_data_uri_rwy = self._image_to_data_uri(input_img_path_rwy)
|
@@ -228,18 +264,18 @@ class VisualEngine:
|
|
228 |
except AttributeError as e_rwy_sdk_attr: logger.error(f"RunwayML SDK AttrError:{e_rwy_sdk_attr}. SDK methods changed?",exc_info=True);return None
|
229 |
except Exception as e_rwy_general: logger.error(f"Runway Gen-4 API error:{e_rwy_general}",exc_info=True);return None
|
230 |
|
231 |
-
def _create_placeholder_video_content(self,
|
232 |
-
if
|
233 |
-
|
234 |
text_clip_object_ph = None
|
235 |
try:
|
236 |
-
text_clip_object_ph = TextClip(
|
237 |
-
bg_color='black', size=
|
238 |
-
text_clip_object_ph.write_videofile(
|
239 |
-
logger.info(f"Generic placeholder video created: {
|
240 |
-
return
|
241 |
except Exception as e_placeholder_video_creation:
|
242 |
-
logger.error(f"Failed to create generic placeholder video '{
|
243 |
return None
|
244 |
finally:
|
245 |
if text_clip_object_ph and hasattr(text_clip_object_ph, 'close'):
|
@@ -249,6 +285,7 @@ class VisualEngine:
|
|
249 |
def generate_scene_asset(self, image_generation_prompt_text, motion_prompt_text_for_video,
|
250 |
scene_data_dict, scene_identifier_fn_base,
|
251 |
generate_as_video_clip_flag=False, runway_target_dur_val=5):
|
|
|
252 |
base_name_current_asset, _ = os.path.splitext(scene_identifier_fn_base)
|
253 |
asset_info_return_obj = {'path': None, 'type': 'none', 'error': True, 'prompt_used': image_generation_prompt_text, 'error_message': 'Asset generation init failed'}
|
254 |
path_to_input_image_for_runway = None
|
@@ -286,33 +323,41 @@ class VisualEngine:
|
|
286 |
else:logger.warning("RunwayML selected but disabled. Use base img.");asset_info_return_obj['error']=True;asset_info_return_obj['error_message']=(asset_info_return_obj.get('error_message',"Base img ok.")+" RunwayML disabled; use base img.").strip();asset_info_return_obj['path']=path_to_input_image_for_runway;asset_info_return_obj['type']='image';asset_info_return_obj['prompt_used']=image_generation_prompt_text
|
287 |
return asset_info_return_obj
|
288 |
|
289 |
-
def generate_narration_audio(self,
|
290 |
-
|
291 |
-
|
292 |
-
|
293 |
-
|
294 |
-
|
295 |
-
|
296 |
-
|
297 |
-
|
298 |
-
|
299 |
-
|
300 |
-
|
301 |
-
|
302 |
-
|
303 |
-
|
304 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
305 |
if self.elevenlabs_voice_settings_obj:
|
306 |
-
if hasattr(self.elevenlabs_voice_settings_obj,'model_dump'):
|
307 |
-
elif hasattr(self.elevenlabs_voice_settings_obj,'dict'):
|
308 |
-
else:
|
309 |
-
|
310 |
-
with open(
|
311 |
-
for
|
312 |
-
if
|
313 |
-
logger.info(f"
|
314 |
-
except AttributeError as
|
315 |
-
except Exception as
|
316 |
|
317 |
def assemble_animatic_from_assets(self, asset_data_list, overall_narration_path=None, output_filename="final_video.mp4", fps=24):
|
318 |
if not asset_data_list: logger.warning("No assets for animatic."); return None
|
@@ -340,7 +385,7 @@ class VisualEngine:
|
|
340 |
logger.debug(f"S{num_of_scene} (4-ToRGB): Final RGB. Mode:{final_rgb_image_for_pil.mode}, Size:{final_rgb_image_for_pil.size}")
|
341 |
debug_path_img_pre_numpy = os.path.join(self.output_dir,f"debug_4_PRE_NUMPY_RGB_S{num_of_scene}.png"); final_rgb_image_for_pil.save(debug_path_img_pre_numpy); logger.info(f"CRITICAL DEBUG: Saved PRE_NUMPY_RGB_S{num_of_scene} to {debug_path_img_pre_numpy}")
|
342 |
|
343 |
-
numpy_frame_arr = np.array(final_rgb_image_for_pil, dtype=np.uint8)
|
344 |
if not numpy_frame_arr.flags['C_CONTIGUOUS']: numpy_frame_arr = np.ascontiguousarray(numpy_frame_arr, dtype=np.uint8)
|
345 |
logger.debug(f"S{num_of_scene} (5-NumPy): Final NumPy. Shape:{numpy_frame_arr.shape}, DType:{numpy_frame_arr.dtype}, Flags:{numpy_frame_arr.flags}")
|
346 |
if numpy_frame_arr.size == 0 or numpy_frame_arr.ndim != 3 or numpy_frame_arr.shape[2] != 3: logger.error(f"S{num_of_scene}: Invalid NumPy shape/size ({numpy_frame_arr.shape}). Skipping."); continue
|
@@ -349,7 +394,7 @@ class VisualEngine:
|
|
349 |
logger.debug(f"S{num_of_scene} (6-ImageClip): Base ImageClip. Duration: {base_image_clip_mvpy.duration}")
|
350 |
|
351 |
debug_path_moviepy_frame = os.path.join(self.output_dir,f"debug_7_MOVIEPY_FRAME_S{num_of_scene}.png")
|
352 |
-
# <<< CORRECTED TRY-EXCEPT
|
353 |
try:
|
354 |
save_frame_time = min(0.1, base_image_clip_mvpy.duration / 2 if base_image_clip_mvpy.duration > 0 else 0.1)
|
355 |
base_image_clip_mvpy.save_frame(debug_path_moviepy_frame, t=save_frame_time)
|
@@ -358,14 +403,13 @@ class VisualEngine:
|
|
358 |
logger.error(f"DEBUG: Error saving frame FROM MOVIEPY ImageClip S{num_of_scene}: {e_save_mvpy_frame}", exc_info=True)
|
359 |
|
360 |
fx_image_clip_mvpy = base_image_clip_mvpy
|
361 |
-
try:
|
362 |
scale_end_kb_val = random.uniform(1.03, 1.08)
|
363 |
if duration_for_scene > 0: fx_image_clip_mvpy = base_image_clip_mvpy.fx(vfx.resize, lambda t_val: 1 + (scale_end_kb_val - 1) * (t_val / duration_for_scene)).set_position('center'); logger.debug(f"S{num_of_scene} (8-KenBurns): Ken Burns applied.")
|
364 |
else: logger.warning(f"S{num_of_scene}: Duration zero, skipping Ken Burns.")
|
365 |
except Exception as e_kb_fx_loop: logger.error(f"S{num_of_scene} Ken Burns error: {e_kb_fx_loop}", exc_info=False)
|
366 |
active_scene_clip = fx_image_clip_mvpy
|
367 |
elif type_of_asset == 'video':
|
368 |
-
# (Video processing logic as before)
|
369 |
source_video_clip_obj=None
|
370 |
try:
|
371 |
logger.debug(f"S{num_of_scene}: Loading VIDEO asset: {path_of_asset}")
|
@@ -385,8 +429,7 @@ class VisualEngine:
|
|
385 |
try: source_video_clip_obj.close()
|
386 |
except Exception as e_close_src_vid: logger.warning(f"S{num_of_scene}: Error closing source VideoFileClip: {e_close_src_vid}")
|
387 |
else: logger.warning(f"S{num_of_scene} Unknown asset type '{type_of_asset}'. Skipping."); continue
|
388 |
-
|
389 |
-
if active_scene_clip and action_in_key: # Text Overlay
|
390 |
try:
|
391 |
dur_text_overlay_val=min(active_scene_clip.duration-0.5,active_scene_clip.duration*0.8)if active_scene_clip.duration>0.5 else (active_scene_clip.duration if active_scene_clip.duration > 0 else 0)
|
392 |
start_text_overlay_val=0.25 if active_scene_clip.duration > 0.5 else 0
|
@@ -396,7 +439,6 @@ class VisualEngine:
|
|
396 |
logger.debug(f"S{num_of_scene}: Text overlay composited.")
|
397 |
else: logger.warning(f"S{num_of_scene}: Text overlay duration zero or negative ({dur_text_overlay_val}). Skipping text overlay.")
|
398 |
except Exception as e_txt_comp_loop:logger.error(f"S{num_of_scene} TextClip compositing error:{e_txt_comp_loop}. Proceeding without text for this scene.",exc_info=True)
|
399 |
-
|
400 |
if active_scene_clip: processed_moviepy_clips_list.append(active_scene_clip); logger.info(f"S{num_of_scene}: Asset successfully processed. Clip duration: {active_scene_clip.duration:.2f}s. Added to final list.")
|
401 |
except Exception as e_asset_loop_main_exc: logger.error(f"MAJOR UNHANDLED ERROR processing asset for S{num_of_scene} (Path: {path_of_asset}): {e_asset_loop_main_exc}", exc_info=True)
|
402 |
finally:
|
|
|
4 |
import mimetypes
|
5 |
import numpy as np
|
6 |
import os
|
7 |
+
import openai # OpenAI v1.x.x+
|
8 |
import requests
|
9 |
import io
|
10 |
import time
|
|
|
15 |
CompositeVideoClip, AudioFileClip)
|
16 |
import moviepy.video.fx.all as vfx
|
17 |
|
18 |
+
try: # MONKEY PATCH for Pillow/MoviePy compatibility
|
19 |
+
if hasattr(Image, 'Resampling') and hasattr(Image.Resampling, 'LANCZOS'): # Pillow 9+
|
20 |
if not hasattr(Image, 'ANTIALIAS'): Image.ANTIALIAS = Image.Resampling.LANCZOS
|
21 |
+
elif hasattr(Image, 'LANCZOS'): # Pillow 8
|
22 |
if not hasattr(Image, 'ANTIALIAS'): Image.ANTIALIAS = Image.LANCZOS
|
23 |
elif not hasattr(Image, 'ANTIALIAS'):
|
24 |
print("WARNING: Pillow version lacks common Resampling or ANTIALIAS. MoviePy effects might fail.")
|
25 |
except Exception as e_mp: print(f"WARNING: ANTIALIAS monkey-patch error: {e_mp}")
|
26 |
|
27 |
logger = logging.getLogger(__name__)
|
28 |
+
# logger.setLevel(logging.DEBUG) # Uncomment for maximum verbosity during active debugging
|
29 |
|
30 |
ELEVENLABS_CLIENT_IMPORTED = False; ElevenLabsAPIClient = None; Voice = None; VoiceSettings = None
|
31 |
try:
|
|
|
42 |
logger.info("RunwayML SDK imported.")
|
43 |
except Exception as e_rwy_imp: logger.warning(f"RunwayML SDK import failed: {e_rwy_imp}. RunwayML disabled.")
|
44 |
|
45 |
+
|
46 |
class VisualEngine:
|
47 |
DEFAULT_FONT_SIZE_PIL = 10; PREFERRED_FONT_SIZE_PIL = 20
|
48 |
VIDEO_OVERLAY_FONT_SIZE = 30; VIDEO_OVERLAY_FONT_COLOR = 'white'
|
49 |
DEFAULT_MOVIEPY_FONT = 'DejaVu-Sans-Bold'; PREFERRED_MOVIEPY_FONT = 'Liberation-Sans-Bold'
|
50 |
|
51 |
+
# <<< CRITICAL __init__ METHOD - ENSURE IT MATCHES THIS >>>
|
52 |
def __init__(self, output_dir="temp_cinegen_media", default_elevenlabs_voice_id="Rachel"):
|
53 |
self.output_dir = output_dir
|
54 |
try:
|
55 |
os.makedirs(self.output_dir, exist_ok=True)
|
56 |
logger.info(f"VisualEngine output directory set/ensured: {os.path.abspath(self.output_dir)}")
|
57 |
+
# Test writability immediately
|
58 |
+
test_file_path = os.path.join(self.output_dir, ".ve_write_test.txt")
|
59 |
+
with open(test_file_path, "w") as f_test:
|
60 |
+
f_test.write("VisualEngine write test OK")
|
61 |
+
os.remove(test_file_path)
|
62 |
+
logger.info(f"Write test to output directory '{self.output_dir}' successful.")
|
63 |
+
except Exception as e_mkdir_init: # More specific exception catching
|
64 |
+
logger.critical(f"CRITICAL FAILURE: Could not create or write to output directory '{os.path.abspath(self.output_dir)}': {e_mkdir_init}", exc_info=True)
|
65 |
+
raise OSError(f"VisualEngine failed to initialize output directory '{self.output_dir}'. Check permissions and path.") from e_mkdir_init
|
66 |
|
67 |
self.font_filename_pil_preference = "DejaVuSans-Bold.ttf"
|
68 |
+
font_paths_to_try = [ self.font_filename_pil_preference, f"/usr/share/fonts/truetype/dejavu/{self.font_filename_pil_preference}", f"/usr/share/fonts/truetype/liberation/LiberationSans-Bold.ttf", f"/System/Library/Fonts/Supplemental/Arial.ttf", f"C:/Windows/Fonts/arial.ttf", f"/usr/local/share/fonts/truetype/mycustomfonts/arial.ttf"]
|
69 |
+
self.resolved_font_path_pil = next((p for p in font_paths_to_try if os.path.exists(p)), None)
|
70 |
+
|
71 |
+
self.active_font_pil = ImageFont.load_default()
|
72 |
+
self.active_font_size_pil = self.DEFAULT_FONT_SIZE_PIL
|
73 |
+
self.active_moviepy_font_name = self.DEFAULT_MOVIEPY_FONT
|
74 |
+
|
75 |
if self.resolved_font_path_pil:
|
76 |
try:
|
77 |
self.active_font_pil = ImageFont.truetype(self.resolved_font_path_pil, self.PREFERRED_FONT_SIZE_PIL)
|
78 |
self.active_font_size_pil = self.PREFERRED_FONT_SIZE_PIL
|
79 |
logger.info(f"Pillow font loaded: {self.resolved_font_path_pil} at size {self.active_font_size_pil}.")
|
80 |
self.active_moviepy_font_name = 'DejaVu-Sans-Bold' if "dejavu" in self.resolved_font_path_pil.lower() else ('Liberation-Sans-Bold' if "liberation" in self.resolved_font_path_pil.lower() else self.DEFAULT_MOVIEPY_FONT)
|
81 |
+
except IOError as e_font_load: logger.error(f"Pillow font IOError for '{self.resolved_font_path_pil}': {e_font_load}. Using default.")
|
82 |
+
else: logger.warning("Preferred Pillow font not found in predefined paths. Using default.")
|
83 |
|
84 |
self.openai_api_key = None; self.USE_AI_IMAGE_GENERATION = False; self.dalle_model = "dall-e-3"; self.image_size_dalle3 = "1792x1024"
|
85 |
self.video_frame_size = (1280, 720)
|
86 |
+
|
87 |
self.elevenlabs_api_key = None; self.USE_ELEVENLABS = False; self.elevenlabs_client_instance = None
|
88 |
+
self.elevenlabs_voice_id = default_elevenlabs_voice_id
|
89 |
logger.info(f"VisualEngine __init__: ElevenLabs Voice ID initially set to: {self.elevenlabs_voice_id}")
|
90 |
+
|
91 |
if VoiceSettings and ELEVENLABS_CLIENT_IMPORTED: self.elevenlabs_voice_settings_obj = VoiceSettings(stability=0.60, similarity_boost=0.80, style=0.15, use_speaker_boost=True)
|
92 |
else: self.elevenlabs_voice_settings_obj = None
|
93 |
+
|
94 |
self.pexels_api_key = None; self.USE_PEXELS = False
|
95 |
self.runway_api_key = None; self.USE_RUNWAYML = False; self.runway_ml_sdk_client_instance = None
|
96 |
+
|
97 |
if RUNWAYML_SDK_IMPORTED and RunwayMLAPIClientClass and os.getenv("RUNWAYML_API_SECRET"):
|
98 |
+
try: self.runway_ml_sdk_client_instance = RunwayMLAPIClientClass(); self.USE_RUNWAYML = True; logger.info("RunwayML Client initialized using RUNWAYML_API_SECRET env var at startup.")
|
99 |
+
except Exception as e_rwy_init_constructor: logger.error(f"Initial RunwayML client initialization via env var failed: {e_rwy_init_constructor}"); self.USE_RUNWAYML = False
|
100 |
+
|
101 |
+
logger.info("VisualEngine __init__ sequence fully completed.")
|
102 |
|
103 |
def set_openai_api_key(self, api_key_value): self.openai_api_key = api_key_value; self.USE_AI_IMAGE_GENERATION = bool(api_key_value); logger.info(f"DALL-E status: {'Ready' if self.USE_AI_IMAGE_GENERATION else 'Disabled'}")
|
104 |
+
|
105 |
def set_elevenlabs_api_key(self, api_key_value, voice_id_from_secret=None):
|
106 |
self.elevenlabs_api_key = api_key_value
|
107 |
if voice_id_from_secret: self.elevenlabs_voice_id = voice_id_from_secret; logger.info(f"11L Voice ID updated via set_elevenlabs_api_key to: {self.elevenlabs_voice_id}")
|
108 |
+
|
109 |
if api_key_value and ELEVENLABS_CLIENT_IMPORTED and ElevenLabsAPIClient:
|
110 |
+
try: self.elevenlabs_client_instance = ElevenLabsAPIClient(api_key=api_key_value); self.USE_ELEVENLABS = bool(self.elevenlabs_client_instance); logger.info(f"11L Client: {'Ready' if self.USE_ELEVENLABS else 'Failed Initialization'} (Using Voice: {self.elevenlabs_voice_id})")
|
111 |
+
except Exception as e_11l_setkey_init: logger.error(f"11L client initialization error: {e_11l_setkey_init}. Service Disabled.", exc_info=True); self.USE_ELEVENLABS=False; self.elevenlabs_client_instance=None
|
112 |
+
else: self.USE_ELEVENLABS = False; logger.info(f"11L Service Disabled (API key not provided or SDK component issue).")
|
113 |
+
|
114 |
def set_pexels_api_key(self, api_key_value): self.pexels_api_key = api_key_value; self.USE_PEXELS = bool(api_key_value); logger.info(f"Pexels status: {'Ready' if self.USE_PEXELS else 'Disabled'}")
|
115 |
+
|
116 |
def set_runway_api_key(self, api_key_value):
|
117 |
self.runway_api_key = api_key_value
|
118 |
if api_key_value:
|
119 |
if RUNWAYML_SDK_IMPORTED and RunwayMLAPIClientClass:
|
120 |
+
if not self.runway_ml_sdk_client_instance: # If not already initialized by env var
|
121 |
try:
|
122 |
+
original_env_secret_val = os.getenv("RUNWAYML_API_SECRET") # Renamed
|
123 |
+
if not original_env_secret_val: os.environ["RUNWAYML_API_SECRET"] = api_key_value; logger.info("Temporarily set RUNWAYML_API_SECRET from provided key for SDK client init.")
|
124 |
self.runway_ml_sdk_client_instance = RunwayMLAPIClientClass(); self.USE_RUNWAYML = True; logger.info("RunwayML Client initialized successfully via set_runway_api_key.")
|
125 |
+
if not original_env_secret_val: del os.environ["RUNWAYML_API_SECRET"]; logger.info("Cleared temporary RUNWAYML_API_SECRET environment variable.")
|
126 |
+
except Exception as e_runway_setkey_init_local: logger.error(f"RunwayML Client initialization in set_runway_api_key failed: {e_runway_setkey_init_local}", exc_info=True); self.USE_RUNWAYML=False;self.runway_ml_sdk_client_instance=None # Renamed
|
127 |
else: self.USE_RUNWAYML = True; logger.info("RunwayML Client was already initialized (likely from environment variable). API key stored.")
|
128 |
else: logger.warning("RunwayML SDK not imported. API key stored, but current integration relies on SDK. Service effectively disabled."); self.USE_RUNWAYML = False
|
129 |
else: self.USE_RUNWAYML = False; self.runway_ml_sdk_client_instance = None; logger.info("RunwayML Service Disabled (no API key provided).")
|
130 |
|
131 |
+
# --- Helper Methods (_image_to_data_uri, _map_resolution_to_runway_ratio, etc.) ---
|
132 |
+
# (These should be the corrected versions from previous iterations)
|
133 |
+
def _image_to_data_uri(self, image_path_in): # Renamed image_path
|
134 |
try:
|
135 |
+
mime_type_val, _ = mimetypes.guess_type(image_path_in) # Renamed
|
136 |
+
if not mime_type_val:
|
137 |
+
file_ext = os.path.splitext(image_path_in)[1].lower() # Renamed
|
138 |
+
mime_type_map = {".png": "image/png", ".jpg": "image/jpeg", ".jpeg": "image/jpeg", ".webp": "image/webp"} # Renamed
|
139 |
+
mime_type_val = mime_type_map.get(file_ext, "application/octet-stream")
|
140 |
+
if mime_type_val == "application/octet-stream": logger.warning(f"Could not determine MIME type for {image_path_in} from ext '{file_ext}', using default {mime_type_val}.")
|
141 |
+
with open(image_path_in, "rb") as img_file_handle: img_binary_data = img_file_handle.read() # Renamed
|
142 |
+
encoded_b64_str = base64.b64encode(img_binary_data).decode('utf-8') # Renamed
|
143 |
+
final_data_uri = f"data:{mime_type_val};base64,{encoded_b64_str}"; logger.debug(f"Data URI for {os.path.basename(image_path_in)} (MIME:{mime_type_val}): {final_data_uri[:100]}..."); return final_data_uri # Renamed
|
144 |
+
except FileNotFoundError: logger.error(f"Img not found {image_path_in} for data URI."); return None
|
145 |
+
except Exception as e_to_data_uri: logger.error(f"Error converting {image_path_in} to data URI:{e_to_data_uri}", exc_info=True); return None # Renamed
|
146 |
|
147 |
+
def _map_resolution_to_runway_ratio(self, width_in, height_in): # Renamed
|
148 |
+
ratio_string = f"{width_in}:{height_in}"; supported_ratios = ["1280:720","720:1280","1104:832","832:1104","960:960","1584:672"]; # Renamed
|
149 |
+
if ratio_string in supported_ratios: return ratio_string
|
150 |
+
logger.warning(f"Res {ratio_string} not in Gen-4 list. Default 1280:720 for Runway."); return "1280:720"
|
151 |
|
152 |
+
def _get_text_dimensions(self, text_str, font_pil_obj): # Renamed
|
153 |
+
def_h = getattr(font_pil_obj, 'size', self.active_font_size_pil);
|
154 |
+
if not text_str: return 0, def_h
|
155 |
try:
|
156 |
+
if hasattr(font_pil_obj,'getbbox'): box = font_pil_obj.getbbox(text_str); w_val=box[2]-box[0]; h_val=box[3]-box[1]; return w_val, h_val if h_val > 0 else def_h # Renamed
|
157 |
+
elif hasattr(font_pil_obj,'getsize'): w_val,h_val=font_pil_obj.getsize(text_str); return w_val, h_val if h_val > 0 else def_h # Renamed
|
158 |
+
else: return int(len(text_str)*def_h*0.6), int(def_h*1.2)
|
159 |
+
except Exception as e_get_dim: logger.warning(f"Error in _get_text_dimensions: {e_get_dim}"); return int(len(text_str)*self.active_font_size_pil*0.6),int(self.active_font_size_pil*1.2) # Renamed
|
160 |
|
161 |
+
def _create_placeholder_image_content(self,text_desc_val, filename_val, size_val=None): # Renamed
|
162 |
+
# (Corrected version from previous responses)
|
163 |
+
if size_val is None: size_val = self.video_frame_size
|
164 |
+
placeholder_img = Image.new('RGB', size_val, color=(20, 20, 40)); placeholder_draw = ImageDraw.Draw(placeholder_img); ph_padding = 25 # Renamed
|
165 |
+
ph_max_w = size_val[0] - (2 * ph_padding); ph_lines = []
|
166 |
+
if not text_desc_val: text_desc_val = "(Placeholder Image)"
|
167 |
+
ph_words = text_desc_val.split(); ph_current_line = ""
|
168 |
+
for ph_word_idx, ph_word in enumerate(ph_words):
|
169 |
+
ph_prosp_add = ph_word + (" " if ph_word_idx < len(ph_words) - 1 else "")
|
170 |
+
ph_test_line = ph_current_line + ph_prosp_add
|
171 |
+
ph_curr_w, _ = self._get_text_dimensions(ph_test_line, self.active_font_pil)
|
172 |
+
if ph_curr_w == 0 and ph_test_line.strip(): ph_curr_w = len(ph_test_line) * (self.active_font_size_pil * 0.6)
|
173 |
+
if ph_curr_w <= ph_max_w: ph_current_line = ph_test_line
|
174 |
else:
|
175 |
+
if ph_current_line.strip(): ph_lines.append(ph_current_line.strip())
|
176 |
+
ph_current_line = ph_prosp_add
|
177 |
+
if ph_current_line.strip(): ph_lines.append(ph_current_line.strip())
|
178 |
+
if not ph_lines and text_desc_val:
|
179 |
+
ph_avg_char_w, _ = self._get_text_dimensions("W", self.active_font_pil); ph_avg_char_w = ph_avg_char_w or (self.active_font_size_pil * 0.6)
|
180 |
+
ph_chars_line = int(ph_max_w / ph_avg_char_w) if ph_avg_char_w > 0 else 20
|
181 |
+
ph_lines.append(text_desc_val[:ph_chars_line] + ("..." if len(text_desc_val) > ph_chars_line else ""))
|
182 |
+
elif not ph_lines: ph_lines.append("(Placeholder Error)")
|
183 |
+
_, ph_single_h = self._get_text_dimensions("Ay", self.active_font_pil); ph_single_h = ph_single_h if ph_single_h > 0 else self.active_font_size_pil + 2
|
184 |
+
ph_max_l = min(len(ph_lines), (size_val[1] - (2 * ph_padding)) // (ph_single_h + 2)) if ph_single_h > 0 else 1; ph_max_l = max(1, ph_max_l)
|
185 |
+
ph_y_pos = ph_padding + (size_val[1] - (2 * ph_padding) - ph_max_l * (ph_single_h + 2)) / 2.0
|
186 |
+
for ph_i_line in range(ph_max_l):
|
187 |
+
ph_line_txt = ph_lines[ph_i_line]; ph_line_w, _ = self._get_text_dimensions(ph_line_txt, self.active_font_pil)
|
188 |
+
if ph_line_w == 0 and ph_line_txt.strip(): ph_line_w = len(ph_line_txt) * (self.active_font_size_pil * 0.6)
|
189 |
+
ph_x_pos = (size_val[0] - ph_line_w) / 2.0
|
190 |
+
try: placeholder_draw.text((ph_x_pos, ph_y_pos), ph_line_txt, font=self.active_font_pil, fill=(200, 200, 180))
|
191 |
+
except Exception as e_ph_draw: logger.error(f"Pillow d.text error: {e_ph_draw} for '{ph_line_txt}'")
|
192 |
+
ph_y_pos += ph_single_h + 2
|
193 |
+
if ph_i_line == 6 and ph_max_l > 7:
|
194 |
+
try: placeholder_draw.text((ph_x_pos, ph_y_pos), "...", font=self.active_font_pil, fill=(200, 200, 180))
|
195 |
+
except Exception as e_ph_elip: logger.error(f"Pillow d.text ellipsis error: {e_ph_elip}"); break
|
196 |
+
ph_filepath = os.path.join(self.output_dir, filename_val)
|
197 |
+
try: placeholder_img.save(ph_filepath); return ph_filepath
|
198 |
+
except Exception as e_ph_save: logger.error(f"Saving placeholder image '{ph_filepath}' error: {e_ph_save}", exc_info=True); return None
|
199 |
|
200 |
+
def _search_pexels_image(self, pexels_query, pexels_output_fn_base): # Renamed
|
201 |
if not self.USE_PEXELS or not self.pexels_api_key: return None
|
202 |
+
pexels_headers = {"Authorization": self.pexels_api_key}
|
203 |
+
pexels_params = {"query": pexels_query, "per_page": 1, "orientation": "landscape", "size": "large2x"}
|
204 |
+
pexels_base_name, _ = os.path.splitext(pexels_output_fn_base)
|
205 |
+
pexels_output_filename = pexels_base_name + f"_pexels_{random.randint(1000,9999)}.jpg"
|
206 |
+
pexels_filepath = os.path.join(self.output_dir, pexels_output_filename)
|
207 |
try:
|
208 |
+
logger.info(f"Pexels: Searching for '{pexels_query}'")
|
209 |
+
pexels_eff_query = " ".join(pexels_query.split()[:5])
|
210 |
+
pexels_params["query"] = pexels_eff_query
|
211 |
+
pexels_response = requests.get("https://api.pexels.com/v1/search", headers=pexels_headers, params=pexels_params, timeout=20)
|
212 |
+
pexels_response.raise_for_status()
|
213 |
+
pexels_data = pexels_response.json()
|
214 |
+
if pexels_data.get("photos") and len(pexels_data["photos"]) > 0:
|
215 |
+
pexels_photo_details = pexels_data["photos"][0]
|
216 |
+
pexels_photo_url = pexels_photo_details.get("src", {}).get("large2x")
|
217 |
+
if not pexels_photo_url: logger.warning(f"Pexels: 'large2x' URL missing for '{pexels_eff_query}'. Details: {pexels_photo_details}"); return None
|
218 |
+
pexels_image_response = requests.get(pexels_photo_url, timeout=60); pexels_image_response.raise_for_status()
|
219 |
+
pexels_img_pil_data = Image.open(io.BytesIO(pexels_image_response.content))
|
220 |
+
if pexels_img_pil_data.mode != 'RGB': pexels_img_pil_data = pexels_img_pil_data.convert('RGB')
|
221 |
+
pexels_img_pil_data.save(pexels_filepath); logger.info(f"Pexels: Image saved to {pexels_filepath}"); return pexels_filepath
|
222 |
+
else: logger.info(f"Pexels: No photos for '{pexels_eff_query}'."); return None
|
223 |
+
except requests.exceptions.RequestException as e_pexels_req: logger.error(f"Pexels: RequestException for '{pexels_query}': {e_pexels_req}", exc_info=False); return None
|
224 |
+
except Exception as e_pexels_general: logger.error(f"Pexels: General error for '{pexels_query}': {e_pexels_general}", exc_info=True); return None
|
225 |
+
|
226 |
+
# ... (Rest of methods: _generate_video_clip_with_runwayml, generate_scene_asset, generate_narration_audio, assemble_animatic_from_assets)
|
227 |
+
# Ensure these are taken from the last fully corrected versions provided, paying close attention to their specific fixes.
|
228 |
+
# For example, generate_narration_audio had its own try-except fix.
|
229 |
+
# assemble_animatic_from_assets had extensive debugging for image corruption.
|
230 |
+
|
231 |
+
# For brevity, I will paste the corrected generate_narration_audio and the
|
232 |
+
# structure for generate_scene_asset and assemble_animatic_from_assets.
|
233 |
+
# You MUST ensure the internal logic of generate_scene_asset and assemble_animatic_from_assets
|
234 |
+
# matches the last "expert" versions that included detailed debugging for image/video issues.
|
235 |
|
236 |
def _generate_video_clip_with_runwayml(self, motion_prompt_rwy, input_img_path_rwy, scene_id_base_fn_rwy, duration_s_rwy=5):
|
237 |
+
# (Keep robust RunwayML logic from before, with proper SDK client instance: self.runway_ml_sdk_client_instance)
|
238 |
if not self.USE_RUNWAYML or not self.runway_ml_sdk_client_instance: logger.warning("RunwayML skip: Not enabled/client not init."); return None
|
239 |
if not input_img_path_rwy or not os.path.exists(input_img_path_rwy): logger.error(f"Runway Gen-4 needs input img. Invalid: {input_img_path_rwy}"); return None
|
240 |
img_data_uri_rwy = self._image_to_data_uri(input_img_path_rwy)
|
|
|
264 |
except AttributeError as e_rwy_sdk_attr: logger.error(f"RunwayML SDK AttrError:{e_rwy_sdk_attr}. SDK methods changed?",exc_info=True);return None
|
265 |
except Exception as e_rwy_general: logger.error(f"Runway Gen-4 API error:{e_rwy_general}",exc_info=True);return None
|
266 |
|
267 |
+
def _create_placeholder_video_content(self, text_desc_ph_vid, filename_ph_vid, duration_ph_vid=4, size_ph_vid=None):
|
268 |
+
if size_ph_vid is None: size_ph_vid = self.video_frame_size
|
269 |
+
filepath_ph_vid_out = os.path.join(self.output_dir, filename_ph_vid)
|
270 |
text_clip_object_ph = None
|
271 |
try:
|
272 |
+
text_clip_object_ph = TextClip(text_desc_ph_vid, fontsize=50, color='white', font=self.video_overlay_font,
|
273 |
+
bg_color='black', size=size_ph_vid, method='caption').set_duration(duration_ph_vid)
|
274 |
+
text_clip_object_ph.write_videofile(filepath_ph_vid_out, fps=24, codec='libx264', preset='ultrafast', logger=None, threads=2)
|
275 |
+
logger.info(f"Generic placeholder video created: {filepath_ph_vid_out}")
|
276 |
+
return filepath_ph_vid_out
|
277 |
except Exception as e_placeholder_video_creation:
|
278 |
+
logger.error(f"Failed to create generic placeholder video '{filepath_ph_vid_out}': {e_placeholder_video_creation}", exc_info=True)
|
279 |
return None
|
280 |
finally:
|
281 |
if text_clip_object_ph and hasattr(text_clip_object_ph, 'close'):
|
|
|
285 |
def generate_scene_asset(self, image_generation_prompt_text, motion_prompt_text_for_video,
|
286 |
scene_data_dict, scene_identifier_fn_base,
|
287 |
generate_as_video_clip_flag=False, runway_target_dur_val=5):
|
288 |
+
# (Corrected DALL-E loop from previous response)
|
289 |
base_name_current_asset, _ = os.path.splitext(scene_identifier_fn_base)
|
290 |
asset_info_return_obj = {'path': None, 'type': 'none', 'error': True, 'prompt_used': image_generation_prompt_text, 'error_message': 'Asset generation init failed'}
|
291 |
path_to_input_image_for_runway = None
|
|
|
323 |
else:logger.warning("RunwayML selected but disabled. Use base img.");asset_info_return_obj['error']=True;asset_info_return_obj['error_message']=(asset_info_return_obj.get('error_message',"Base img ok.")+" RunwayML disabled; use base img.").strip();asset_info_return_obj['path']=path_to_input_image_for_runway;asset_info_return_obj['type']='image';asset_info_return_obj['prompt_used']=image_generation_prompt_text
|
324 |
return asset_info_return_obj
|
325 |
|
326 |
+
def generate_narration_audio(self, text_to_narrate, output_filename="narration_overall.mp3"):
|
327 |
+
# <<< CORRECTED METHOD with try/except >>>
|
328 |
+
if not self.USE_ELEVENLABS or not self.elevenlabs_client_instance or not text_to_narrate:
|
329 |
+
logger.info("ElevenLabs conditions not met (service disabled, client not init, or no text). Skipping audio generation.")
|
330 |
+
return None
|
331 |
+
audio_filepath_narration = os.path.join(self.output_dir, output_filename)
|
332 |
+
try: # Main try block for the entire operation
|
333 |
+
logger.info(f"Generating ElevenLabs audio (Voice ID: {self.elevenlabs_voice_id}) for text: \"{text_to_narrate[:70]}...\"")
|
334 |
+
audio_stream_method_11l = None
|
335 |
+
if hasattr(self.elevenlabs_client_instance, 'text_to_speech') and hasattr(self.elevenlabs_client_instance.text_to_speech, 'stream'):
|
336 |
+
audio_stream_method_11l = self.elevenlabs_client_instance.text_to_speech.stream; logger.info("Using ElevenLabs SDK method: client.text_to_speech.stream()")
|
337 |
+
elif hasattr(self.elevenlabs_client_instance, 'generate_stream'):
|
338 |
+
audio_stream_method_11l = self.elevenlabs_client_instance.generate_stream; logger.info("Using ElevenLabs SDK method: client.generate_stream()")
|
339 |
+
elif hasattr(self.elevenlabs_client_instance, 'generate'):
|
340 |
+
logger.info("Using ElevenLabs SDK method: client.generate() (non-streaming).")
|
341 |
+
voice_param_11l = str(self.elevenlabs_voice_id)
|
342 |
+
if Voice and self.elevenlabs_voice_settings_obj: voice_param_11l = Voice(voice_id=str(self.elevenlabs_voice_id), settings=self.elevenlabs_voice_settings_obj)
|
343 |
+
audio_bytes_data = self.elevenlabs_client_instance.generate(text=text_to_narrate, voice=voice_param_11l, model="eleven_multilingual_v2")
|
344 |
+
with open(audio_filepath_narration, "wb") as audio_file_out: audio_file_out.write(audio_bytes_data)
|
345 |
+
logger.info(f"ElevenLabs audio (non-streamed) saved successfully to: {audio_filepath_narration}"); return audio_filepath_narration
|
346 |
+
else: logger.error("No recognized audio generation method found on the ElevenLabs client instance."); return None
|
347 |
+
|
348 |
+
if audio_stream_method_11l:
|
349 |
+
params_for_voice_stream = {"voice_id": str(self.elevenlabs_voice_id)}
|
350 |
if self.elevenlabs_voice_settings_obj:
|
351 |
+
if hasattr(self.elevenlabs_voice_settings_obj, 'model_dump'): params_for_voice_stream["voice_settings"] = self.elevenlabs_voice_settings_obj.model_dump()
|
352 |
+
elif hasattr(self.elevenlabs_voice_settings_obj, 'dict'): params_for_voice_stream["voice_settings"] = self.elevenlabs_voice_settings_obj.dict()
|
353 |
+
else: params_for_voice_stream["voice_settings"] = self.elevenlabs_voice_settings_obj
|
354 |
+
audio_data_iterator_11l = audio_stream_method_11l(text=text_to_narrate, model_id="eleven_multilingual_v2", **params_for_voice_stream)
|
355 |
+
with open(audio_filepath_narration, "wb") as audio_file_out_stream:
|
356 |
+
for audio_chunk_data in audio_data_iterator_11l:
|
357 |
+
if audio_chunk_data: audio_file_out_stream.write(audio_chunk_data)
|
358 |
+
logger.info(f"ElevenLabs audio (streamed) saved successfully to: {audio_filepath_narration}"); return audio_filepath_narration
|
359 |
+
except AttributeError as ae_11l_sdk: logger.error(f"AttributeError with ElevenLabs SDK client: {ae_11l_sdk}. SDK version/methods might differ.", exc_info=True); return None
|
360 |
+
except Exception as e_11l_general_audio: logger.error(f"General error during ElevenLabs audio generation: {e_11l_general_audio}", exc_info=True); return None
|
361 |
|
362 |
def assemble_animatic_from_assets(self, asset_data_list, overall_narration_path=None, output_filename="final_video.mp4", fps=24):
|
363 |
if not asset_data_list: logger.warning("No assets for animatic."); return None
|
|
|
385 |
logger.debug(f"S{num_of_scene} (4-ToRGB): Final RGB. Mode:{final_rgb_image_for_pil.mode}, Size:{final_rgb_image_for_pil.size}")
|
386 |
debug_path_img_pre_numpy = os.path.join(self.output_dir,f"debug_4_PRE_NUMPY_RGB_S{num_of_scene}.png"); final_rgb_image_for_pil.save(debug_path_img_pre_numpy); logger.info(f"CRITICAL DEBUG: Saved PRE_NUMPY_RGB_S{num_of_scene} to {debug_path_img_pre_numpy}")
|
387 |
|
388 |
+
numpy_frame_arr = np.array(final_rgb_image_for_pil, dtype=np.uint8)
|
389 |
if not numpy_frame_arr.flags['C_CONTIGUOUS']: numpy_frame_arr = np.ascontiguousarray(numpy_frame_arr, dtype=np.uint8)
|
390 |
logger.debug(f"S{num_of_scene} (5-NumPy): Final NumPy. Shape:{numpy_frame_arr.shape}, DType:{numpy_frame_arr.dtype}, Flags:{numpy_frame_arr.flags}")
|
391 |
if numpy_frame_arr.size == 0 or numpy_frame_arr.ndim != 3 or numpy_frame_arr.shape[2] != 3: logger.error(f"S{num_of_scene}: Invalid NumPy shape/size ({numpy_frame_arr.shape}). Skipping."); continue
|
|
|
394 |
logger.debug(f"S{num_of_scene} (6-ImageClip): Base ImageClip. Duration: {base_image_clip_mvpy.duration}")
|
395 |
|
396 |
debug_path_moviepy_frame = os.path.join(self.output_dir,f"debug_7_MOVIEPY_FRAME_S{num_of_scene}.png")
|
397 |
+
# <<< THIS IS THE CORRECTED TRY-EXCEPT BLOCK >>>
|
398 |
try:
|
399 |
save_frame_time = min(0.1, base_image_clip_mvpy.duration / 2 if base_image_clip_mvpy.duration > 0 else 0.1)
|
400 |
base_image_clip_mvpy.save_frame(debug_path_moviepy_frame, t=save_frame_time)
|
|
|
403 |
logger.error(f"DEBUG: Error saving frame FROM MOVIEPY ImageClip S{num_of_scene}: {e_save_mvpy_frame}", exc_info=True)
|
404 |
|
405 |
fx_image_clip_mvpy = base_image_clip_mvpy
|
406 |
+
try:
|
407 |
scale_end_kb_val = random.uniform(1.03, 1.08)
|
408 |
if duration_for_scene > 0: fx_image_clip_mvpy = base_image_clip_mvpy.fx(vfx.resize, lambda t_val: 1 + (scale_end_kb_val - 1) * (t_val / duration_for_scene)).set_position('center'); logger.debug(f"S{num_of_scene} (8-KenBurns): Ken Burns applied.")
|
409 |
else: logger.warning(f"S{num_of_scene}: Duration zero, skipping Ken Burns.")
|
410 |
except Exception as e_kb_fx_loop: logger.error(f"S{num_of_scene} Ken Burns error: {e_kb_fx_loop}", exc_info=False)
|
411 |
active_scene_clip = fx_image_clip_mvpy
|
412 |
elif type_of_asset == 'video':
|
|
|
413 |
source_video_clip_obj=None
|
414 |
try:
|
415 |
logger.debug(f"S{num_of_scene}: Loading VIDEO asset: {path_of_asset}")
|
|
|
429 |
try: source_video_clip_obj.close()
|
430 |
except Exception as e_close_src_vid: logger.warning(f"S{num_of_scene}: Error closing source VideoFileClip: {e_close_src_vid}")
|
431 |
else: logger.warning(f"S{num_of_scene} Unknown asset type '{type_of_asset}'. Skipping."); continue
|
432 |
+
if active_scene_clip and action_in_key:
|
|
|
433 |
try:
|
434 |
dur_text_overlay_val=min(active_scene_clip.duration-0.5,active_scene_clip.duration*0.8)if active_scene_clip.duration>0.5 else (active_scene_clip.duration if active_scene_clip.duration > 0 else 0)
|
435 |
start_text_overlay_val=0.25 if active_scene_clip.duration > 0.5 else 0
|
|
|
439 |
logger.debug(f"S{num_of_scene}: Text overlay composited.")
|
440 |
else: logger.warning(f"S{num_of_scene}: Text overlay duration zero or negative ({dur_text_overlay_val}). Skipping text overlay.")
|
441 |
except Exception as e_txt_comp_loop:logger.error(f"S{num_of_scene} TextClip compositing error:{e_txt_comp_loop}. Proceeding without text for this scene.",exc_info=True)
|
|
|
442 |
if active_scene_clip: processed_moviepy_clips_list.append(active_scene_clip); logger.info(f"S{num_of_scene}: Asset successfully processed. Clip duration: {active_scene_clip.duration:.2f}s. Added to final list.")
|
443 |
except Exception as e_asset_loop_main_exc: logger.error(f"MAJOR UNHANDLED ERROR processing asset for S{num_of_scene} (Path: {path_of_asset}): {e_asset_loop_main_exc}", exc_info=True)
|
444 |
finally:
|