pywombat 0.2.0__py3-none-any.whl → 0.4.0__py3-none-any.whl

pywombat/cli.py

@@ -40,6 +40,11 @@ import yaml
     type=click.Path(exists=True, path_type=Path),
     help="Filter configuration YAML file to apply quality and impact filters.",
 )
+@click.option(
+    "--debug",
+    type=str,
+    help="Debug mode: show rows matching chrom:pos (e.g., chr11:70486013). Displays #CHROM, POS, VEP_SYMBOL, and columns from filter expression.",
+)
 def cli(
     input_file: Path,
     output: Optional[str],
@@ -47,6 +52,7 @@ def cli(
     verbose: bool,
     pedigree: Optional[Path],
     filter_config: Optional[Path],
+    debug: Optional[str],
 ):
     """
     Wombat: A tool for processing bcftools tabulated TSV files.
@@ -80,14 +86,6 @@ def cli(
         if verbose and is_gzipped:
             click.echo("Detected gzipped file", err=True)
 
-        # Read the TSV file (handles both plain and gzipped)
-        df = pl.read_csv(input_file, separator="\t")
-
-        if verbose:
-            click.echo(
-                f"Input shape: {df.shape[0]} rows, {df.shape[1]} columns", err=True
-            )
-
         # Read pedigree file if provided
         pedigree_df = None
         if pedigree:
@@ -95,22 +93,18 @@ def cli(
                 click.echo(f"Reading pedigree file: {pedigree}", err=True)
             pedigree_df = read_pedigree(pedigree)
 
-        # Process the dataframe
-        formatted_df = format_bcftools_tsv(df, pedigree_df)
-
-        if verbose:
-            click.echo(
-                f"Output shape: {formatted_df.shape[0]} rows, {formatted_df.shape[1]} columns",
-                err=True,
-            )
-
-        # Apply filters if provided
+        # Load filter config if provided
         filter_config_data = None
         if filter_config:
             if verbose:
                 click.echo(f"Reading filter config: {filter_config}", err=True)
             filter_config_data = load_filter_config(filter_config)
 
+        # Debug mode: show specific variant
+        if debug:
+            debug_variant(input_file, pedigree_df, filter_config_data, debug, verbose)
+            return
+
         # Determine output prefix
         if output is None:
             # Generate default output prefix from input filename
@@ -128,36 +122,131 @@ def cli(
             else:
                 output = input_stem
 
-        # Apply filters and write output
+        # Use streaming approach with lazy API
+        if verbose:
+            click.echo("Processing with streaming mode...", err=True)
+
+        # Build lazy query
+        lazy_df = pl.scan_csv(input_file, separator="\t")
+
+        # Apply formatting transformations
+        lazy_df = format_bcftools_tsv_lazy(lazy_df, pedigree_df)
+
+        # Apply filters if provided
         if filter_config_data:
-            apply_filters_and_write(
-                formatted_df,
-                filter_config_data,
-                output,
-                output_format,
-                verbose,
-            )
-        else:
-            # No filters - write single output file
-            # Construct output filename with prefix and format
-            output_path = Path(f"{output}.{output_format}")
+            lazy_df = apply_filters_lazy(lazy_df, filter_config_data, verbose)
 
-            if output_format == "tsv":
-                formatted_df.write_csv(output_path, separator="\t")
-            elif output_format == "tsv.gz":
-                csv_content = formatted_df.write_csv(separator="\t")
-                with gzip.open(output_path, "wt") as f:
-                    f.write(csv_content)
-            elif output_format == "parquet":
-                formatted_df.write_parquet(output_path)
+        # Write output
+        output_path = Path(f"{output}.{output_format}")
 
-            click.echo(f"Formatted data written to {output_path}", err=True)
+        if output_format == "tsv":
+            lazy_df.sink_csv(output_path, separator="\t")
+        elif output_format == "tsv.gz":
+            # For gzip, we need to collect and write
+            df = lazy_df.collect()
+            csv_content = df.write_csv(separator="\t")
+            with gzip.open(output_path, "wt") as f:
+                f.write(csv_content)
+        elif output_format == "parquet":
+            lazy_df.sink_parquet(output_path)
+
+        if verbose:
+            click.echo(f"Data written to {output_path}", err=True)
 
     except Exception as e:
         click.echo(f"Error: {e}", err=True)
         raise click.Abort()
 
 
+def debug_variant(
+    input_file: Path,
+    pedigree_df: Optional[pl.DataFrame],
+    filter_config: Optional[dict],
+    debug_pos: str,
+    verbose: bool,
+):
+    """Debug mode: display rows matching a specific chrom:pos."""
+    # Parse debug position
+    if ":" not in debug_pos:
+        click.echo(
+            "Error: Debug position must be in format 'chrom:pos' (e.g., chr11:70486013)",
+            err=True,
+        )
+        raise click.Abort()
+
+    chrom, pos = debug_pos.split(":", 1)
+    try:
+        pos = int(pos)
+    except ValueError:
+        click.echo(f"Error: Position must be an integer, got '{pos}'", err=True)
+        raise click.Abort()
+
+    if verbose:
+        click.echo(f"Debug mode: searching for {chrom}:{pos}", err=True)
+
+    # Read and format the data
+    df = pl.read_csv(input_file, separator="\t")
+    formatted_df = format_bcftools_tsv(df, pedigree_df)
+
+    # Filter to matching rows
+    matching_rows = formatted_df.filter(
+        (pl.col("#CHROM") == chrom) & (pl.col("POS") == pos)
+    )
+
+    if matching_rows.shape[0] == 0:
+        click.echo(f"No rows found matching {chrom}:{pos}", err=True)
+        return
+
+    # Determine which columns to display
+    columns_to_show = ["#CHROM", "POS"]
+
+    # Add VEP_SYMBOL if it exists
+    if "VEP_SYMBOL" in matching_rows.columns:
+        columns_to_show.append("VEP_SYMBOL")
+
+    # Extract column names from expression if filter config provided
+    if filter_config and "expression" in filter_config:
+        expression = filter_config["expression"]
+        # Extract column names from expression using regex
+        # Match patterns like "column_name" before operators
+        column_pattern = r"\b([A-Za-z_][A-Za-z0-9_]*)\b\s*[=!<>]"
+        found_columns = re.findall(column_pattern, expression)
+
+        for col in found_columns:
+            if col in matching_rows.columns and col not in columns_to_show:
+                columns_to_show.append(col)
+
+    # Select only the columns we want to display
+    display_df = matching_rows.select(
+        [col for col in columns_to_show if col in matching_rows.columns]
+    )
+
+    # Replace null and NaN values with <null> and <NaN> for display
+    for col in display_df.columns:
+        if display_df[col].dtype in [pl.Float32, pl.Float64]:
+            # For numeric columns, handle both NaN and null
+            display_df = display_df.with_columns(
+                pl.when(pl.col(col).is_null())
+                .then(pl.lit("<null>"))
+                .when(pl.col(col).is_nan())
+                .then(pl.lit("<NaN>"))
+                .otherwise(pl.col(col).cast(pl.Utf8))
+                .alias(col)
+            )
+        else:
+            # For non-numeric columns, only handle null
+            display_df = display_df.with_columns(
+                pl.when(pl.col(col).is_null())
+                .then(pl.lit("<null>"))
+                .otherwise(pl.col(col).cast(pl.Utf8))
+                .alias(col)
+            )
+
+    # Display the results
+    click.echo(f"\nFound {matching_rows.shape[0]} row(s) matching {chrom}:{pos}:\n")
+    click.echo(display_df.write_csv(separator="\t"))
+
+
 def load_filter_config(config_path: Path) -> dict:
     """Load and parse filter configuration from YAML file."""
     with open(config_path, "r") as f:
@@ -377,6 +466,30 @@ def parse_impact_filter_expression(expression: str, df: pl.DataFrame) -> pl.Expr
                     if col_name not in df.columns:
                         raise ValueError(f"Column '{col_name}' not found in dataframe")
 
+                    # Check for null value
+                    if value.upper() == "NULL":
+                        col_expr = pl.col(col_name)
+                        if op == "=":
+                            return col_expr.is_null()
+                        elif op == "!=":
+                            return ~col_expr.is_null()
+                        else:
+                            raise ValueError(
+                                f"Operator '{op}' not supported for null comparison, use = or !="
+                            )
+
+                    # Check for NaN value
+                    if value.upper() == "NAN":
+                        col_expr = pl.col(col_name).cast(pl.Float64, strict=False)
+                        if op == "=":
+                            return col_expr.is_nan()
+                        elif op == "!=":
+                            return ~col_expr.is_nan()
+                        else:
+                            raise ValueError(
+                                f"Operator '{op}' not supported for NaN comparison, use = or !="
+                            )
+
                     # Try to convert value to number, otherwise treat as string
                     try:
                         value_num = float(value)
@@ -957,5 +1070,142 @@ def format_bcftools_tsv(
     return melted_df
 
 
+def format_bcftools_tsv_lazy(
+    lazy_df: pl.LazyFrame, pedigree_df: Optional[pl.DataFrame] = None
+) -> pl.LazyFrame:
+    """
+    Format a bcftools tabulated TSV using lazy operations for streaming.
+
+    This is a simplified version that collects minimally for complex operations.
+    """
+    # For complex transformations like melting, we need to collect temporarily
+    # but we do this in a streaming fashion
+    df = lazy_df.collect(streaming=True)
+    formatted_df = format_bcftools_tsv(df, pedigree_df)
+    return formatted_df.lazy()
+
+
+def apply_filters_lazy(
+    lazy_df: pl.LazyFrame, filter_config: dict, verbose: bool = False
+) -> pl.LazyFrame:
+    """Apply quality and expression filters using lazy operations."""
+    quality_config = filter_config.get("quality", {})
+    expression = filter_config.get("expression")
+
+    # Apply quality filters
+    if quality_config:
+        # Filter: sample_gt must contain at least one '1' (default: true)
+        filter_no_alt = quality_config.get("filter_no_alt_allele", True)
+        if filter_no_alt:
+            lazy_df = lazy_df.filter(
+                pl.col("sample_gt").str.contains("1")
+                | pl.col("sample_gt").str.contains("2")
+            )
+
+        # Apply minimum depth filter
+        if "sample_dp_min" in quality_config:
+            min_dp = quality_config["sample_dp_min"]
+            lazy_df = lazy_df.filter(
+                pl.col("sample_dp").cast(pl.Float64, strict=False) >= min_dp
+            )
+
+        # Apply minimum GQ filter
+        if "sample_gq_min" in quality_config:
+            min_gq = quality_config["sample_gq_min"]
+            lazy_df = lazy_df.filter(
+                pl.col("sample_gq").cast(pl.Float64, strict=False) >= min_gq
+            )
+
+        # VAF filters for heterozygous (0/1 or 1/0)
+        if (
+            "sample_vaf_het_min" in quality_config
+            or "sample_vaf_het_max" in quality_config
+        ):
+            # Check if genotype is het (contains one '1' and one '0', no '2')
+            is_het = (
+                (pl.col("sample_gt").str.count_matches("1") == 1)
+                & (pl.col("sample_gt").str.count_matches("0") == 1)
+                & (~pl.col("sample_gt").str.contains("2"))
+            )
+
+            het_conditions = []
+            if "sample_vaf_het_min" in quality_config:
+                het_conditions.append(
+                    pl.col("sample_vaf") >= quality_config["sample_vaf_het_min"]
+                )
+            if "sample_vaf_het_max" in quality_config:
+                het_conditions.append(
+                    pl.col("sample_vaf") <= quality_config["sample_vaf_het_max"]
+                )
+
+            if het_conditions:
+                het_filter = het_conditions[0]
+                for cond in het_conditions[1:]:
+                    het_filter = het_filter & cond
+
+                lazy_df = lazy_df.filter(~is_het | het_filter)
+
+        # VAF filter for homozygous alternate (1/1)
+        if "sample_vaf_homalt_min" in quality_config:
+            is_homalt = pl.col("sample_gt") == "1/1"
+            lazy_df = lazy_df.filter(
+                ~is_homalt
+                | (pl.col("sample_vaf") >= quality_config["sample_vaf_homalt_min"])
+            )
+
+        # VAF filter for homozygous reference (0/0)
+        if "sample_vaf_hom_ref_max" in quality_config:
+            is_hom_ref = pl.col("sample_gt") == "0/0"
+            lazy_df = lazy_df.filter(
+                ~is_hom_ref
+                | (pl.col("sample_vaf") <= quality_config["sample_vaf_hom_ref_max"])
+            )
+
+        # Apply same filters to parents if requested
+        apply_to_parents = quality_config.get("apply_to_parents", False)
+        if apply_to_parents:
+            # Father filters
+            if "sample_dp_min" in quality_config:
+                min_dp = quality_config["sample_dp_min"]
+                lazy_df = lazy_df.filter(
+                    (pl.col("father_dp").is_null())
+                    | (pl.col("father_dp").cast(pl.Float64, strict=False) >= min_dp)
+                )
+
+            if "sample_gq_min" in quality_config:
+                min_gq = quality_config["sample_gq_min"]
+                lazy_df = lazy_df.filter(
+                    (pl.col("father_gq").is_null())
+                    | (pl.col("father_gq").cast(pl.Float64, strict=False) >= min_gq)
+                )
+
+            # Mother filters
+            if "sample_dp_min" in quality_config:
+                min_dp = quality_config["sample_dp_min"]
+                lazy_df = lazy_df.filter(
+                    (pl.col("mother_dp").is_null())
+                    | (pl.col("mother_dp").cast(pl.Float64, strict=False) >= min_dp)
+                )
+
+            if "sample_gq_min" in quality_config:
+                min_gq = quality_config["sample_gq_min"]
+                lazy_df = lazy_df.filter(
+                    (pl.col("mother_gq").is_null())
+                    | (pl.col("mother_gq").cast(pl.Float64, strict=False) >= min_gq)
+                )
+
+    # Apply expression filter if provided
+    if expression:
+        if verbose:
+            click.echo(f"Applying expression filter: {expression}", err=True)
+
+        # We need to collect temporarily to use parse_impact_filter_expression
+        df = lazy_df.collect(streaming=True)
+        filter_expr = parse_impact_filter_expression(expression, df)
+        lazy_df = df.lazy().filter(filter_expr)
+
+    return lazy_df
+
+
 if __name__ == "__main__":
     cli()

{pywombat-0.2.0.dist-info → pywombat-0.4.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pywombat
-Version: 0.2.0
+Version: 0.4.0
 Summary: A CLI tool for processing and filtering bcftools tabulated TSV files with pedigree support
 Project-URL: Homepage, https://github.com/bourgeron-lab/pywombat
 Project-URL: Repository, https://github.com/bourgeron-lab/pywombat

pywombat-0.4.0.dist-info/RECORD

@@ -0,0 +1,6 @@
+pywombat/__init__.py,sha256=iIPN9vJtsIUhl_DiKNnknxCamLinfayodLLFK8y-aJg,54
+pywombat/cli.py,sha256=dg38E39VpdJhKQt3aGSHwSiLWn1W8JnUkcsy3ZUHD5w,43518
+pywombat-0.4.0.dist-info/METADATA,sha256=ZKPTIp9ud2AIVbcujg4ciq900DX-UkGs5oafa41jxTQ,4982
+pywombat-0.4.0.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
+pywombat-0.4.0.dist-info/entry_points.txt,sha256=Vt7U2ypbiEgCBlEV71ZPk287H5_HKmPBT4iBu6duEcE,44
+pywombat-0.4.0.dist-info/RECORD,,

pywombat-0.2.0.dist-info/RECORD

@@ -1,6 +0,0 @@
-pywombat/__init__.py,sha256=iIPN9vJtsIUhl_DiKNnknxCamLinfayodLLFK8y-aJg,54
-pywombat/cli.py,sha256=PZKV6FoqZyGgG7_mMIO2FzyeONdBaCqnhDATYsQJqMo,33899
-pywombat-0.2.0.dist-info/METADATA,sha256=7Qg2XnaTM92pmIewu5fw_vrcQW5JCVkkj2q6mNC9v88,4982
-pywombat-0.2.0.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
-pywombat-0.2.0.dist-info/entry_points.txt,sha256=Vt7U2ypbiEgCBlEV71ZPk287H5_HKmPBT4iBu6duEcE,44
-pywombat-0.2.0.dist-info/RECORD,,