2125_GCE/gx_compiler/segmenter.py

from typing import List
from dataclasses import dataclass


@dataclass
class Segment:
    segment_id: str
    start_line: int
    end_line: int
    start_byte: int
    end_byte: int
    content: str
    symbolic_lane: int = 1  # Default lane


class SourceSegmenter:
    def __init__(self, code: str):
        self.code = code
        self.lines = code.split('\n')

    @staticmethod
    def segment(code: str) -> List[Segment]:
        segmenter = SourceSegmenter(code)
        return segmenter._find_segments()

    def _find_segments(self) -> List[Segment]:
        boundaries = self._find_boundaries()
        segments = []

        for i, (start, end) in enumerate(boundaries):
            segment_id = f"seg_{i}"
            start_byte = self._line_offset(start)
            end_byte = self._line_offset(end)
            content = '\n'.join(self.lines[start:end])
            
            # Analyze content to determine lane
            symbolic_lane = self._infer_lane_from_content(content, i)
            
            segments.append(Segment(
                segment_id=segment_id,
                start_line=start,
                end_line=end,
                start_byte=start_byte,
                end_byte=end_byte,
                content=content,
                symbolic_lane=symbolic_lane
            ))

        return segments
    
    def _infer_lane_from_content(self, content: str, index: int) -> int:
        """Infer lane from content analysis (not just ID patterns).
        
        Returns lane 0-7 based on content characteristics.
        """
        lines = content.split('\n')
        stripped_lines = [l.strip() for l in lines if l.strip()]
        
        if not stripped_lines:
            return 1  # Default semantic flow
        
        # Check for structural markers (lane 0)
        has_control_flow = any(
            any(x in l for x in ['if ', 'for ', 'while ', 'return ', 'try:', 'except', 'with '])
            for l in stripped_lines
        )
        if has_control_flow:
            return 0
        
        # Check for comments/annotations (lane 3)
        has_comments = any(
            l.startswith('#') or l.startswith('//') or '/*' in l or '*/' in l
            for l in lines
        )
        if has_comments:
            return 3
        
        # Check for execution hints (lane 4)
        has_hints = any(
            any(x in l.lower() for x in ['hint', 'note', 'todo', 'fixme', 'warning', 'danger'])
            for l in lines
        )
        if has_hints:
            return 4
        
        # Check for metadata/annotations (lane 3)
        has_metadata = any(
            any(x in l.lower() for x in ['<glyph:', 'metadata', 'tag', 'annotation', 'symbolic'])
            for l in lines
        )
        if has_metadata:
            return 3
        
        # Check for execution hints (dangerous commands) (lane 4)
        has_execution_hints = any(
            any(x in l for x in ['rm -rf', 'del ', 'os.system', 'subprocess', 'exec(', 'eval('])
            for l in lines
        )
        if has_execution_hints:
            return 4
        
        # Check for outlines/templates (lane 5)
        has_template = any(
            any(x in l.lower() for x in ['step ', 'todo:', 'placeholder', 'fill-in', 'template', 'outline'])
            for l in lines
        )
        if has_template:
            return 5
        
        # Check for author/contributor markers (lane 6)
        has_contributor = any(
            any(x in l.lower() for x in ['author', 'contributor', 'copyright', '@', 'by '])
            for l in lines
        )
        if has_contributor:
            return 6
        
        # Check for epoch/time markers (lane 7)
        has_epoch = any(
            any(x in l.lower() for x in ['epoch', 'timestamp', 'date', 'time', 'version', 'build'])
            for l in lines
        )
        if has_epoch:
            return 7
        
        # Default: semantic flow (lane 1)
        return 1

    def _line_offset(self, line_num: int) -> int:
        total = 0
        for i in range(min(line_num, len(self.lines))):
            total += len(self.lines[i]) + 1
        return total

    def _find_boundaries(self) -> List[tuple]:
        boundaries = []
        start = 0
        
        # First pass: find def/class boundaries (code structure)
        code_boundaries = []
        for i, line in enumerate(self.lines):
            stripped = line.lstrip()
            if stripped.startswith('def ') or stripped.startswith('class '):
                if start < i and any(self.lines[start:i]):
                    code_boundaries.append((start, i))
                start = i
        if start < len(self.lines):
            code_boundaries.append((start, len(self.lines)))
        
        # Use code boundaries if present
        if code_boundaries:
            return code_boundaries
        
        # For non-code content, split by content type changes OR blank lines
        current_start = 0
        current_type = None
        blank_count = 0
        
        for i, line in enumerate(self.lines):
            stripped = line.lstrip()
            
            # Track blank lines
            if not stripped:
                blank_count += 1
                continue
            
            # Detect content type
            if stripped.startswith('#') or stripped.startswith('//'):
                new_type = 'comment'
            elif any(x in stripped.lower() for x in ['step ', 'todo:', 'placeholder', 'outline']):
                new_type = 'template'
            elif any(x in stripped.lower() for x in ['author', 'contributor', 'copyright', '@']):
                new_type = 'author'
            elif any(x in stripped.lower() for x in ['epoch', 'timestamp', 'date', 'version']):
                new_type = 'epoch'
            elif any(x in stripped for x in ['if ', 'for ', 'while ', 'return ', 'try:', 'except', 'with ']):
                new_type = 'control'
            else:
                new_type = 'content'
            
            # Split on blank line (2+ consecutive) or type change
            if blank_count >= 2 or (current_type is not None and new_type != current_type):
                if current_start < i - blank_count:
                    boundaries.append((current_start, i - blank_count))
                current_start = i - blank_count if blank_count > 0 else i
                current_type = new_type
                blank_count = 0
            
            # If no blank line but type changed, split here
            if blank_count == 0 and current_type is not None and new_type != current_type:
                boundaries.append((current_start, i))
                current_start = i
                current_type = new_type
        
        # Final boundary
        if current_start < len(self.lines):
            boundaries.append((current_start, len(self.lines)))
        
        return boundaries if boundaries else [(0, len(self.lines))]