modules/video_creator/services/libraries/pexels_client.py

import requests
import logging
from typing import List, Optional
from pathlib import Path
import random

logger = logging.getLogger(__name__)


class PexelsClient:
    """Client for Pexels API to fetch background videos"""
    
    def __init__(self, api_key: str):
        """
        Initialize Pexels client
        
        Args:
            api_key: Pexels API key
        """
        self.api_key = api_key
        self.base_url = "https://api.pexels.com/videos"
        self.headers = {"Authorization": api_key}
        self.joker_terms = ["nature", "globe", "space", "ocean"]
    
    def find_video(
        self,
        search_terms: List[str],
        duration: float,
        exclude_ids: Optional[List[int]] = None,
        orientation: str = "portrait"
    ) -> dict:
        """
        Find a suitable video from Pexels
        
        Args:
            search_terms: Keywords to search for
            duration: Required video duration in seconds
            exclude_ids: List of video IDs to exclude
            orientation: 'portrait' or 'landscape'
        
        Returns:
            Dict with 'id' and 'url' of the selected video
        """
        exclude_ids = exclude_ids or []
        
        # Try user-provided search terms first
        for term in search_terms:
            video = self._search_and_select(term, duration, exclude_ids, orientation)
            if video:
                return video
        
        # Fall back to joker terms
        logger.info(f"No videos found for {search_terms}, using joker terms")
        for term in self.joker_terms:
            video = self._search_and_select(term, duration, exclude_ids, orientation)
            if video:
                return video
        
        raise Exception("No suitable videos found on Pexels")
    
    def _search_and_select(
        self,
        query: str,
        min_duration: float,
        exclude_ids: List[int],
        orientation: str
    ) -> Optional[dict]:
        """Search for videos and select a suitable one"""
        try:
            logger.debug(f"Searching Pexels for: {query} ({orientation})")
            
            response = requests.get(
                f"{self.base_url}/search",
                headers=self.headers,
                params={
                    "query": query,
                    "orientation": orientation,
                    "per_page": 15,
                    "size": "medium"  # Good balance of quality and file size
                },
                timeout=10
            )
            
            if response.status_code != 200:
                logger.warning(f"Pexels API error: {response.status_code}")
                return None
            
            data = response.json()
            videos = data.get("videos", [])
            
            if not videos:
                logger.debug(f"No videos found for query: {query}")
                return None
            
            # Filter suitable videos
            suitable_videos = []
            for video in videos:
                if video["id"] in exclude_ids:
                    continue
                
                # Get video file URL (HD or SD)
                video_files = video.get("video_files", [])
                if not video_files:
                    continue
                
                # Sort by quality and find a good match
                video_files = sorted(
                    video_files,
                    key=lambda x: x.get("width", 0) * x.get("height", 0),
                    reverse=True
                )
                
                # Find appropriate quality based on orientation
                target_width = 1080 if orientation == "portrait" else 1920
                target_height = 1920 if orientation == "portrait" else 1080
                
                selected_file = None
                for vf in video_files:
                    # Look for files close to our target resolution
                    if vf.get("width") and vf.get("height"):
                        if (abs(vf["width"] - target_width) < 300 and 
                            abs(vf["height"] - target_height) < 300):
                            selected_file = vf
                            break
                
                # Fallback to highest quality if no exact match
                if not selected_file and video_files:
                    selected_file = video_files[0]
                
                if selected_file and selected_file.get("link"):
                    suitable_videos.append({
                        "id": video["id"],
                        "url": selected_file["link"],
                        "duration": video.get("duration", 0)
                    })
            
            if not suitable_videos:
                return None
            
            # Filter by duration if possible
            # Try to find videos that are at least 50% of the requested duration
            # to avoid stitching too many tiny clips
            duration_threshold = min(min_duration * 0.5, 15)  # Cap at 15s requirement
            long_enough_videos = [v for v in suitable_videos if v["duration"] >= duration_threshold]
            
            if long_enough_videos:
                # Select FIRST (most relevant) video instead of random
                selected = long_enough_videos[0]
                logger.info(f"Selected Pexels video ID {selected['id']} (duration: {selected['duration']}s) for query '{query}'")
                return selected
            
            # Fallback to first suitable video
            selected = suitable_videos[0]
            logger.info(f"Selected Pexels video ID {selected['id']} (duration: {selected['duration']}s) for query '{query}' (fallback)")
            return selected
            
        except Exception as e:
            logger.error(f"Error searching Pexels: {e}")
            return None

    def find_photo(
        self,
        query: str,
        orientation: str = "portrait"
    ) -> Optional[dict]:
        """
        Find a suitable photo from Pexels
        
        Args:
            query: Search term
            orientation: 'portrait' or 'landscape'
            
        Returns:
            Dict with 'id' and 'url' of the photo
        """
        try:
            logger.debug(f"Searching Pexels for photo: {query} ({orientation})")
            
            # Pexels Photo API endpoint
            url = "https://api.pexels.com/v1/search"
            
            response = requests.get(
                url,
                headers=self.headers,
                params={
                    "query": query,
                    "orientation": orientation,
                    "per_page": 15,
                    "size": "large"
                },
                timeout=10
            )
            
            if response.status_code != 200:
                logger.warning(f"Pexels Photo API error: {response.status_code}")
                return None
            
            data = response.json()
            photos = data.get("photos", [])
            
            if not photos:
                logger.debug(f"No photos found for query: {query}")
                return None
            
            # Select FIRST (most relevant) photo instead of random
            photo = photos[0]
            
            # Get URL (prefer original or large2x)
            src = photo.get("src", {})
            url = src.get("original") or src.get("large2x") or src.get("large")
            
            if not url:
                return None
                
            logger.info(f"Selected Pexels photo ID {photo['id']} for query '{query}'")
            return {
                "id": photo["id"],
                "url": url,
                "type": "photo"
            }
            
        except Exception as e:
            logger.error(f"Error searching Pexels photos: {e}")
            return None