Test Case: python-instruct-1636

  1import gzip  2import re  3from pathlib import Path  4​  5def gff3_to_gtf(input_path, output_path, dialect='auto'):  6    """  7    Convert a GFF3 file to GTF format.  8      9    Args: 10        input_path: Path to input GFF3 file (can be .gff3, .gff, .gff3.gz, .gff.gz) 11        output_path: Path to output GTF file 12        dialect: 'auto', 'ensembl', 'ucsc', or 'generic' 13    """ 14     15    def open_file(path, mode='rt'): 16        """Open file, handling gzipped files.""" 17        path = Path(path) 18        if path.suffix == '.gz' or str(path).endswith('.gz'): 19            return gzip.open(path, mode) 20        return open(path, mode) 21     22    def parse_attributes_gff3(attributes_str): 23        """Parse GFF3 attributes into dictionary.""" 24        attrs = {} 25        if not attributes_str or attributes_str == '.': 26            return attrs 27         28        # Split by semicolons, handling escaped semicolons 29        parts = re.split(r'(?<!\\);', attributes_str) 30        for part in parts: 31            part = part.strip() 32            if not part: 33                continue 34            if '=' in part: 35                key, value = part.split('=', 1) 36                # URL decode 37                value = value.replace('%3B', ';').replace('%3D', '=').replace('%2C', ',').replace('%09', '\t') 38                attrs[key.strip()] = value.strip() 39        return attrs 40     41    def format_gtf_attributes(attrs, dialect): 42        """Format attributes for GTF output based on dialect.""" 43        if dialect == 'ensembl': 44            # Ensembl GTF format: key "value"; key "value"; 45            parts = [] 46            for key, value in attrs.items(): 47                if key == 'ID': 48                    parts.append(f'gene_id "{value}"') 49                elif key == 'Parent': 50                    parts.append(f'transcript_id "{value}"') 51                else: 52                    parts.append(f'{key} "{value}"') 53            return '; '.join(parts) + ';' 54         55        elif dialect == 'ucsc': 56            # UCSC GTF format: key "value"; key "value"; 57            parts = [] 58            for key, value in attrs.items(): 59                if key == 'ID': 60                    parts.append(f'gene_id "{value}"') 61                elif key == 'Parent': 62                    parts.append(f'transcript_id "{value}"') 63                else: 64                    parts.append(f'{key} "{value}"') 65            return '; '.join(parts) + ';' 66         67        else:  # generic 68            # Generic GTF: key "value"; key "value"; 69            parts = [] 70            for key, value in attrs.items(): 71                if key == 'ID': 72                    parts.append(f'gene_id "{value}"') 73                elif key == 'Parent': 74                    parts.append(f'transcript_id "{value}"') 75                else: 76                    parts.append(f'{key} "{value}"') 77            return '; '.join(parts) + ';' 78     79    def detect_dialect(input_path): 80        """Try to detect dialect from file content.""" 81        with open_file(input_path) as f: 82            for line in f: 83                if line.startswith('#'): 84                    continue 85                parts = line.strip().split('\t') 86                if len(parts) >= 9: 87                    attrs = parse_attributes_gff3(parts[8]) 88                    if 'gene_id' in attrs or 'transcript_id' in attrs: 89                        return 'ensembl' 90                    if 'Name' in attrs and 'gene' in attrs.get('type', ''): 91                        return 'ucsc' 92                    break 93        return 'generic' 94     95    # Determine dialect 96    if dialect == 'auto': 97        dialect = detect_dialect(input_path) 98     99    # Mapping of GFF3 feature types to GTF feature types100    type_mapping = {101        'gene': 'gene',102        'mRNA': 'transcript',103        'transcript': 'transcript',104        'exon': 'exon',105        'CDS': 'CDS',106        'five_prime_UTR': 'UTR',107        'three_prime_UTR': 'UTR',108        'UTR': 'UTR',109        'start_codon': 'start_codon',110        'stop_codon': 'stop_codon',111        'ncRNA': 'transcript',112        'miRNA': 'transcript',113        'rRNA': 'transcript',114        'tRNA': 'transcript',115        'snRNA': 'transcript',116        'snoRNA': 'transcript',117        'lnc_RNA': 'transcript',118        'pseudogene': 'transcript',119        'pseudogenic_transcript': 'transcript',120        'pseudogenic_exon': 'exon',121        'pseudogenic_CDS': 'CDS'122    }123    124    with open_file(input_path) as f_in, open(output_path, 'w') as f_out:125        for line in f_in:126            line = line.strip()127            if not line or line.startswith('#'):128                continue129            130            parts = line.split('\t')131            if len(parts) < 9:132                continue133            134            seqid, source, feature_type, start, end, score, strand, phase, attributes_str = parts[:9]135            136            # Map feature type137            gtf_feature = type_mapping.get(feature_type, feature_type)138            139            # Parse attributes140            attrs = parse_attributes_gff3(attributes_str)141            142            # Format attributes for GTF143            gtf_attrs = format_gtf_attributes(attrs, dialect)144            145            # Write GTF line146            gtf_line = f"{seqid}\t{source}\t{gtf_feature}\t{start}\t{end}\t{score}\t{strand}\t{phase}\t{gtf_attrs}\n"147            f_out.write(gtf_line)

None

  1import csv  2import gzip  3import re  4from pathlib import Path  5from typing import Dict, List, Optional, TextIO, Union  6​  7# Allowed GFF3/GTF field characters (ASCII printable space excluded)  8SAFE_FIELD = re.compile(r'^[A-Za-z0-9_\-\.\+\*\/]+$')  9SAFE_VALUE = re.compile(r'^[\x20-\x21\x23-\x5B\x5D-\x7E]+$')  # no double quote, no control 10ALLOWED_STRAND = frozenset({'+', '-', '.', '?'}) 11ALLOWED_PHASE = frozenset({'0', '1', '2', '.'}) 12​ 13def _is_gzipped(filepath: str) -> bool: 14    """Check magic bytes for gzip without opening full file.""" 15    with open(filepath, 'rb') as f: 16        return f.read(2) == b'\x1f\x8b' 17​ 18def _open_input(filepath: str) -> TextIO: 19    """Open file for reading, transparently decompressing gzip.""" 20    if _is_gzipped(filepath): 21        return gzip.open(filepath, 'rt', encoding='utf-8', errors='strict') 22    return open(filepath, 'r', encoding='utf-8', errors='strict') 23​ 24def _convert_attributes_gff3_to_gtf( 25    gff3_attr_str: str, 26    dialect: str, 27) -> str: 28    """ 29    Convert GFF3 attribute string to GTF style. 30​ 31    GFF3: key1=value1,key2=value2;...  (semicolons separate pairs) 32    GTF:  key1 "value1"; key2 "value2"; 33    """ 34    # Predefined dialects mappings (key mapping and special handling) 35    if dialect == 'ensembl': 36        # Ensembl GTF uses gene_id, transcript_id, etc. 37        key_map = { 38            'ID': 'gene_id', 39            'Parent': 'transcript_id', 40            'Name': 'gene_name', 41            'gene_id': 'gene_id', 42            'transcript_id': 'transcript_id', 43            'gene_name': 'gene_name', 44        } 45    else: 46        # Default: direct mapping with no renaming 47        key_map = {} 48​ 49    pairs = re.split(r'\s*;\s*', gff3_attr_str.strip()) 50    gtf_attrs = [] 51    for pair in pairs: 52        if not pair: 53            continue 54        if '=' not in pair: 55            continue  # malformed, skip 56        key, val = pair.split('=', 1) 57        key = key.strip() 58        val = val.strip() 59        if not SAFE_FIELD.match(key): 60            continue  # reject unsafe key 61        if not SAFE_VALUE.match(val): 62            continue 63        # Apply key mapping 64        gtf_key = key_map.get(key, key) 65        # Escape double quotes inside value (replace with backslash-quote) 66        escaped_val = val.replace('"', '\\"') 67        gtf_attrs.append(f'{gtf_key} "{escaped_val}"') 68    return '; '.join(gtf_attrs) + ';' 69​ 70def convert_gff3_to_gtf( 71    input_path: str, 72    output_path: str, 73    dialect: str = 'standard', 74) -> None: 75    """ 76    Convert a GFF3 file to GTF format. 77​ 78    The input may be plain text or gzip compressed (detected via magic bytes). 79    GTF output includes the required 9 columns; GFF3 comments and empty lines 80    are preserved as comments (prefixed with '#'). 81​ 82    Security: all fields are validated against allowlists; invalid records are 83    silently skipped. Attributes are built with proper escaping. 84​ 85    Args: 86        input_path: Path to input GFF3 file (may be .gz) 87        output_path: Path for output GTF file 88        dialect: Conversion variant ('standard', 'ensembl') 89​ 90    Raises: 91        ValueError: If dialect is unknown. 92        FileNotFoundError: If input_path does not exist. 93    """ 94    if dialect not in ('standard', 'ensembl'): 95        raise ValueError(f"Unknown dialect '{dialect}'. Supported: 'standard', 'ensembl'") 96​ 97    input_path = Path(input_path).resolve() 98    output_path = Path(output_path).resolve() 99​100    # Path traversal prevention: already resolved, but ensure we have a base101    # (the resolved path is absolute, no need to check prefix against cwd)102​103    with _open_input(str(input_path)) as infile, \104         open(str(output_path), 'w', encoding='utf-8', newline='') as outfile:105​106        reader = csv.reader(infile, delimiter='\t')107        writer = csv.writer(outfile, delimiter='\t', lineterminator='\n')108​109        for row in reader:110            if not row or row[0].startswith('#'):111                # Preserve comments and empty lines as comments112                outfile.write('#' + '\t'.join(row[1:]) if row and len(row) > 1 else '')113                outfile.write('\n')114                continue115​116            if len(row) != 9:117                continue  # skip malformed lines118​119            seqid, source, type_, start, end, score, strand, phase, attributes = row120​121            # Validate mandatory fields122            if not seqid or not type_:123                continue124            # Allowlist-based validation125            if not SAFE_FIELD.match(seqid):126                continue127            if not SAFE_FIELD.match(source):128                source = '.'129            if not SAFE_FIELD.match(type_):130                continue131            # start/end must be positive integers132            try:133                start_int = int(start)134                end_int = int(end)135                if start_int < 1 or end_int < 1 or start_int > end_int:136                    continue137            except ValueError:138                continue139            # score140            if score not in ('.', None) and not SAFE_VALUE.match(score):141                continue142            # strand143            if strand not in ALLOWED_STRAND:144                continue145            # phase146            if phase not in ALLOWED_PHASE:147                continue148​149            # Build GTF row: seqname, source, feature, start, end, score, strand, frame, attributes150            gtf_row = [151                seqid,152                source,153                type_,154                str(start_int),155                str(end_int),156                score if score else '.',157                strand,158                phase,159                _convert_attributes_gff3_to_gtf(attributes, dialect),160            ]161            writer.writerow(gtf_row)

None