src/profilers/file_profiler.py

"""
File profiler for detecting format characteristics.

Analyzes files to determine encoding, delimiter, column count,
schema version, and other structural properties.
"""

from pathlib import Path
from typing import Dict, Any, Optional, List
import pandas as pd

from ..core.base_classes import BaseProfiler
from ..core.config import Config
from ..core.exceptions import ProfilerError
from ..cleaners.encoding_cleaner import EncodingCleaner
from ..cleaners.delimiter_cleaner import DelimiterCleaner


class FileProfiler(BaseProfiler):
    """
    Profiles data files to detect their format characteristics.

    Creates a comprehensive profile including:
    - File size and line count
    - Encoding and confidence
    - Delimiter type and consistency
    - Column count and consistency
    - Sample data
    - Detected schema version
    """

    def __init__(self, config: Config):
        """
        Initialize file profiler.

        Args:
            config: Configuration object
        """
        super().__init__(config)

        # Initialize cleaners for detection
        self.encoding_cleaner = EncodingCleaner(config)
        self.delimiter_cleaner = DelimiterCleaner(config)

    def profile(self, file_path: str | Path) -> Dict[str, Any]:
        """
        Profile a file to detect its characteristics.

        Args:
            file_path: Path to file to profile

        Returns:
            Dictionary with profile information

        Raises:
            ProfilerError: If profiling fails
        """
        file_path = Path(file_path)

        if not file_path.exists():
            raise ProfilerError(f"File not found: {file_path}")

        profile = {
            'file_path': str(file_path),
            'file_name': file_path.name,
            'file_size_bytes': file_path.stat().st_size
        }

        try:
            # Detect encoding
            encoding, encoding_confidence = self.encoding_cleaner.detect_encoding(
                file_path
            )
            profile['encoding'] = encoding
            profile['encoding_confidence'] = encoding_confidence

            # Read file with detected encoding
            text = self.encoding_cleaner.clean(file_path)
            lines = text.strip().split('\n')
            profile['line_count'] = len(lines)

            if not lines:
                profile['status'] = 'empty'
                return profile

            # Detect delimiter
            delimiter, delimiter_consistency = self.delimiter_cleaner.detect_delimiter(
                text
            )
            profile['delimiter'] = delimiter
            profile['delimiter_consistency'] = delimiter_consistency

            # Analyze column structure
            column_counts = []
            for i, line in enumerate(lines[:100]):  # Sample first 100 lines
                if line.strip():
                    cols = line.split(delimiter)
                    column_counts.append(len(cols))

            if column_counts:
                profile['column_count_mode'] = max(set(column_counts), key=column_counts.count)
                profile['column_count_min'] = min(column_counts)
                profile['column_count_max'] = max(column_counts)
                profile['column_count_std'] = pd.Series(column_counts).std()

            # Get sample data
            profile['sample_lines'] = lines[:5]

            # Detect schema version based on column count
            profile['schema_version'] = self._detect_schema_version(
                profile.get('column_count_mode')
            )

            profile['status'] = 'success'

        except Exception as e:
            profile['status'] = 'failed'
            profile['error'] = str(e)

        return profile

    def _detect_schema_version(self, column_count: Optional[int]) -> Optional[str]:
        """
        Detect schema version based on column count.

        Args:
            column_count: Number of columns

        Returns:
            Schema version identifier or None
        """
        if column_count is None:
            return None

        # Schema version heuristics based on column count
        # These would be refined based on actual data analysis
        if column_count <= 10:
            return 'v1'
        elif column_count <= 13:
            return 'v2'
        elif column_count <= 15:
            return 'v3'
        elif column_count <= 18:
            return 'v4'
        else:
            return 'v5'

    def profile_batch(
        self,
        file_paths: List[str | Path],
        show_progress: bool = True
    ) -> List[Dict[str, Any]]:
        """
        Profile multiple files.

        Args:
            file_paths: List of file paths
            show_progress: Whether to show progress bar

        Returns:
            List of profile dictionaries
        """
        profiles = []

        if show_progress:
            try:
                from tqdm import tqdm
                iterator = tqdm(file_paths, desc="Profiling files")
            except ImportError:
                iterator = file_paths
        else:
            iterator = file_paths

        for file_path in iterator:
            try:
                profile = self.profile(file_path)
                profiles.append(profile)
            except ProfilerError as e:
                # Log error but continue
                profiles.append({
                    'file_path': str(file_path),
                    'status': 'failed',
                    'error': str(e)
                })

        return profiles

    def generate_summary_report(
        self,
        profiles: List[Dict[str, Any]]
    ) -> Dict[str, Any]:
        """
        Generate summary report from batch profiling results.

        Args:
            profiles: List of profile dictionaries

        Returns:
            Summary statistics dictionary
        """
        total_files = len(profiles)
        successful = sum(1 for p in profiles if p.get('status') == 'success')
        failed = sum(1 for p in profiles if p.get('status') == 'failed')

        # Encoding distribution
        encodings = [p.get('encoding') for p in profiles if 'encoding' in p]
        encoding_counts = pd.Series(encodings).value_counts().to_dict()

        # Delimiter distribution
        delimiters = [p.get('delimiter') for p in profiles if 'delimiter' in p]
        delimiter_counts = pd.Series(delimiters).value_counts().to_dict()

        # Schema version distribution
        versions = [p.get('schema_version') for p in profiles if 'schema_version' in p]
        version_counts = pd.Series(versions).value_counts().to_dict()

        # File sizes
        sizes = [p.get('file_size_bytes', 0) for p in profiles]
        total_size_bytes = sum(sizes)
        avg_size_bytes = total_size_bytes / len(sizes) if sizes else 0

        return {
            'total_files': total_files,
            'successful': successful,
            'failed': failed,
            'total_size_bytes': total_size_bytes,
            'total_size_gb': total_size_bytes / (1024**3),
            'avg_size_bytes': avg_size_bytes,
            'encodings': encoding_counts,
            'delimiters': delimiter_counts,
            'schema_versions': version_counts
        }