File size: 7,570 Bytes
8b7ae7a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
import os
import re
import json
import base64
import openai
from pathlib import Path
import google.generativeai as genai
from app.utils.logging import get_logger
from config import make_path, OPENAI_API_KEY, GEMINI_API_KEY, DATA_DIR


class FrameAnalyzer:
    def __init__(self, video_path: str, openai_api_key: str = "", save_dir: str = 'processed/frame-analysis'):
        # βœ… Set OpenAI key (explicit or from environment)
        
        # print(openai_api_key)

        if openai_api_key:
            openai.api_key = openai_api_key
        else:
            import os
            openai.api_key = os.getenv("OPENAI_API_KEY")

        self.video_path = Path(video_path)
        self.frames_dir = DATA_DIR / 'interim' / 'frames' / f'{self.video_path.stem}_'
        self.save_path = make_path(save_dir, video_path, 'frame_analysis', 'json')
        self.save_path.parent.mkdir(parents=True, exist_ok=True)

        log_file = f'{self.video_path.stem}_log.txt'
        self.logger = get_logger('frame_analysis', log_file)

    @staticmethod
    def encode_image(path: Path) -> str:
        with open(path, 'rb') as f:
            return base64.b64encode(f.read()).decode('utf-8')

    @staticmethod
    def extract_json(text: str) -> dict:
        try:
            return json.loads(text)
        except json.JSONDecodeError:
            pass

        match = re.search(r'```json\s*(\{.*?\})\s*```', text, re.DOTALL)
        if match:
            return json.loads(match.group(1))

        match = re.search(r'(\{.*?\})', text, re.DOTALL)
        if match:
            return json.loads(match.group(1))

        raise ValueError('No valid JSON found in GPT response')

    def gpt_analyze(self, frame_path: Path, prev_path: Path, next_path: Path) -> dict:
        prompt = """
        You are an expert video content strategist. Analyze this video frame and surrounding context. 
        Determine if the lighting is poor or intentionally low for creative reasons. 

        Output JSON only:
        {
          lighting: 0-100,
          is_artistic_dark: true|false,
          composition: 0-100,
          has_text: true|false,
          text: "string",
          hook_strength: 0-100
        }
        """

        images = [
            {'type': 'image_url', 'image_url': {'url': f'data:image/jpeg;base64,{self.encode_image(p)}'}}
            for p in [prev_path, frame_path, next_path] if p.exists()
        ]

        response = openai.chat.completions.create(
            model='gpt-4o-mini',
            messages=[
                {'role': 'user', 'content': [{'type': 'text', 'text': prompt}] + images}
            ],
            temperature=0.2,
            max_tokens=400,
        )
        return self.extract_json(response.choices[0].message.content)

    def analyze(self) -> dict:
        results = {}
        all_frames = sorted(self.frames_dir.glob('*_scene_*.jpg'))
        center_frames = [f for f in all_frames if '_prev' not in f.name and '_next' not in f.name]

        for frame in center_frames:
            prev = frame.with_name(frame.name.replace('.jpg', '_prev.jpg'))
            next_ = frame.with_name(frame.name.replace('.jpg', '_next.jpg'))

            self.logger.info('Analyzing frame: %s', frame.name)
            try:
                result = self.gpt_analyze(frame, prev, next_)
                results[frame.name] = result
            except Exception as e:
                self.logger.error('LLM analysis failed on %s: %s', frame.name, e)
                results[frame.name] = {'error': str(e)}

        with open(self.save_path, 'w', encoding='utf-8') as f:
            json.dump(results, f, indent=2)

        self.logger.info('Frame analysis saved to %s', self.save_path)
        return results

class HookAnalyzer:
    def __init__(self, video_path: str, gemini_api_key: str = ""):
        self.video_path = Path(video_path)
        self.frames_dir = Path('data/interim/frames') / f'{self.video_path.stem}_'
        self.audio_json = make_path('processed/audio-analysis', video_path, 'audio_analysis', 'json')
        self.output_json = make_path('processed/hook-analysis', video_path, 'hook_analysis', 'json')
        self.logger = get_logger('hook_analysis', f'{self.video_path.stem}_log.txt')

        # βœ… Set Gemini key (explicit or from environment)
        if gemini_api_key:
            genai.configure(api_key=gemini_api_key)
        else:
            genai.configure(api_key=os.getenv("GEMINI_API_KEY", ""))
        self.model = genai.GenerativeModel('gemini-2.5-pro')

    def _encode_image(self, path: Path) -> bytes:
        with open(path, 'rb') as f:
            return f.read()

    def _load_audio_summary(self) -> dict:
        with open(self.audio_json, 'r', encoding='utf-8') as f:
            return json.load(f)

    def _gemini_hook_alignment(self, audio_summary: dict, frames: list[Path]) -> dict:
        parts = [{'mime_type': 'image/jpeg', 'data': self._encode_image(f)} for f in frames if f.exists()]
        text = f"""You are a virality analyst. Analyze the opening visuals and tone:
        - Does the audio mood match the expressions and visuals?
        - Are viewers likely to be hooked in the first few seconds?

        Audio Summary: {json.dumps(audio_summary)}

        Give JSON only:
        {{
        "hook_alignment_score": 0-100,
        "facial_sync": "good|ok|poor|none",
        "comment": "short summary"
        }}"""

        try:
            response = self.model.generate_content([text] + parts)
            raw_text = getattr(response, 'text', '').strip()
            self.logger.debug("Gemini raw response: %s", raw_text)
            if not raw_text:
                raise ValueError("Gemini response was empty.")
            
            raw_text = (
                raw_text
                .replace('```json\n', '')
                .replace('\n```', '')
                .replace('```json', '')
                .replace('```', '')
            )

            return json.loads(raw_text)
        except json.JSONDecodeError as e:
            self.logger.error("❌ Failed to parse Gemini response as JSON: %s", e)
            self.logger.debug("Gemini response was: %r", getattr(response, 'text', '<<NO TEXT>>'))
            return {
                "hook_alignment_score": -1,
                "facial_sync": "none",
                "comment": "Invalid JSON response from Gemini"
            }
        except Exception as e:
            error_msg = str(e)
            self.logger.error("❌ Gemini API call failed: %s", e)
            
            # Check if it's an API key error - if so, raise it to stop the pipeline
            if any(keyword in error_msg.lower() for keyword in ["api_key", "invalid", "401", "403", "authentication", "unauthorized"]):
                raise ValueError(f"Invalid Gemini API key: {error_msg}") from e
            
            # For other errors, return defaults
            return {
                "hook_alignment_score": -1,
                "facial_sync": "none",
                "comment": f"Gemini API error: {error_msg}"
            }

    def analyze(self) -> dict:
        audio_summary = self._load_audio_summary()
        frames = sorted(self.frames_dir.glob('*_scene_*.jpg'))[:3]
        result = self._gemini_hook_alignment(audio_summary, frames)

        self.output_json.parent.mkdir(parents=True, exist_ok=True)
        with open(self.output_json, 'w', encoding='utf-8') as f:
            json.dump(result, f, indent=2)

        self.logger.info('Hook analysis saved to %s', self.output_json)
        return result