masster 0.5.1__py3-none-any.whl → 0.5.3__py3-none-any.whl

masster/study/h5.py

@@ -974,7 +974,7 @@ def _load_dataframe_from_group(
 
     # Second pass: handle missing columns
     for col in missing_columns:
-        logger.info(f"Column '{col}' not found in {df_name}.")
+        logger.debug(f"Column '{col}' not found in {df_name}.")
         # For missing columns, create appropriately sized array with appropriate defaults
         if col in object_columns:
             data[col] = [None] * expected_length
@@ -1857,6 +1857,26 @@ def _load_study5(self, filename=None):
                     self.logger,
                     object_columns,
                 )
+                
+                # Sanity check: replace any missing rt_original with rt values
+                if self.features_df is not None and not self.features_df.is_empty():
+                    if "rt_original" in self.features_df.columns and "rt" in self.features_df.columns:
+                        null_rt_original_count = self.features_df.filter(pl.col("rt_original").is_null()).height
+                        if null_rt_original_count > 0:
+                            self.logger.info(f"Replacing {null_rt_original_count} missing rt_original values with rt")
+                            self.features_df = self.features_df.with_columns(
+                                pl.when(pl.col("rt_original").is_null())
+                                .then(pl.col("rt"))
+                                .otherwise(pl.col("rt_original"))
+                                .alias("rt_original")
+                            )
+                        else:
+                            self.logger.debug("All rt_original values are present")
+                    else:
+                        if "rt_original" not in self.features_df.columns:
+                            self.logger.debug("rt_original column not found in features_df")
+                        if "rt" not in self.features_df.columns:
+                            self.logger.debug("rt column not found in features_df")
             else:
                 self.features_df = _create_empty_dataframe_from_schema("features_df", schema)
             pbar.update(1)
@@ -2008,12 +2028,12 @@ def _load_study5(self, filename=None):
         )
 
     # Sanitize null feature_id and consensus_id values with new UIDs (same method as merge)
-    self._sanitize_null_ids()
+    _sanitize_nulls(self)
 
     self.logger.debug("Study loaded")
+    
 
-
-def _load_ms1(self, sample_path: str) -> pl.DataFrame:
+def _load_ms1(self, filename: str) -> pl.DataFrame:
     """
     Optimized method to load only MS1 data from a sample5 file for isotope detection.
     
@@ -2030,48 +2050,146 @@ def _load_ms1(self, sample_path: str) -> pl.DataFrame:
     Note:
         Used by find_iso() for efficient isotope pattern detection without full sample loading
     """
-    try:
-        with h5py.File(sample_path, "r") as f:
-            # Check if ms1 group exists
-            if "ms1" not in f:
-                self.logger.debug(f"No MS1 data found in {sample_path}")
-                return pl.DataFrame()
-            
-            ms1_group = f["ms1"]
+    #try:
+    # add .sample5 extension if not provided
+    if not filename.endswith(".sample5"):
+        filename += ".sample5"
+    with h5py.File(filename, "r") as f:
+        # Check if ms1 group exists
+        if "ms1" not in f:
+            self.logger.debug(f"No MS1 data found in {filename}")
+            return pl.DataFrame()
+        
+        ms1_group = f["ms1"]
+        
+        # Load MS1 data efficiently
+        ms1_data = {}
+        for col in ms1_group.keys():
+            ms1_data[col] = ms1_group[col][:]
+        
+        if not ms1_data:
+            self.logger.debug(f"Empty MS1 data in {filename}")
+            return pl.DataFrame()
             
-            # Load MS1 data efficiently
-            ms1_data = {}
-            for col in ms1_group.keys():
-                ms1_data[col] = ms1_group[col][:]
+        # Create DataFrame with proper schema
+        ms1_df = pl.DataFrame(ms1_data)
+        
+        # Apply expected schema for MS1 data
+        expected_schema = {
+            "cycle": pl.Int64,
+            "scan_uid": pl.Int64, 
+            "rt": pl.Float64,
+            "mz": pl.Float64,
+            "inty": pl.Float64
+        }
+        
+        # Cast columns to expected types if they exist
+        cast_expressions = []
+        for col, dtype in expected_schema.items():
+            if col in ms1_df.columns:
+                cast_expressions.append(pl.col(col).cast(dtype))
+        
+        if cast_expressions:
+            ms1_df = ms1_df.with_columns(cast_expressions)
+        
+        self.logger.debug(f"Loaded {len(ms1_df)} MS1 peaks from {filename}")
+        return ms1_df
             
-            if not ms1_data:
-                self.logger.debug(f"Empty MS1 data in {sample_path}")
-                return pl.DataFrame()
-                
-            # Create DataFrame with proper schema
-            ms1_df = pl.DataFrame(ms1_data)
+   #except Exception as e:
+   #     self.logger.warning(f"Failed to load MS1 data from {sample_path}: {e}")
+   #     return pl.DataFrame()
+
+
+def _sanitize_nulls(self):
+    """
+    Sanitize null feature_id and consensus_id values by replacing them with new integer IDs.
+    For feature_id: generates large sequential integers that can be converted by merge/align functions.
+    For consensus_id: uses 16-character UUID strings (as expected by merge function).
+    """
+    import uuid
+    import polars as pl
+    import time
+    
+    # Sanitize features_df feature_id column
+    if hasattr(self, 'features_df') and self.features_df is not None and not self.features_df.is_empty():
+        # Check for null feature_ids
+        null_feature_ids = self.features_df.filter(pl.col("feature_id").is_null()).shape[0]
+        if null_feature_ids > 0:
+            self.logger.debug(f"Sanitizing {null_feature_ids} null feature_id values with new integer IDs")
             
-            # Apply expected schema for MS1 data
-            expected_schema = {
-                "cycle": pl.Int64,
-                "scan_uid": pl.Int64, 
-                "rt": pl.Float64,
-                "mz": pl.Float64,
-                "inty": pl.Float64
-            }
+            # Find the maximum existing feature_id (convert strings to int if possible)
+            max_existing_id = 0
+            existing_ids = self.features_df.filter(pl.col("feature_id").is_not_null())["feature_id"].to_list()
+            for fid in existing_ids:
+                try:
+                    int_id = int(fid)
+                    max_existing_id = max(max_existing_id, int_id)
+                except (ValueError, TypeError):
+                    # Skip non-integer IDs
+                    pass
             
-            # Cast columns to expected types if they exist
-            cast_expressions = []
-            for col, dtype in expected_schema.items():
-                if col in ms1_df.columns:
-                    cast_expressions.append(pl.col(col).cast(dtype))
+            # Generate new sequential integer IDs starting from max + timestamp offset
+            # Use timestamp to ensure uniqueness across different sanitization runs
+            base_id = max(max_existing_id + 1, int(time.time() * 1000000))  # Microsecond timestamp
+            new_int_ids = [str(base_id + i) for i in range(null_feature_ids)]
+            uid_index = 0
             
-            if cast_expressions:
-                ms1_df = ms1_df.with_columns(cast_expressions)
+            # Create a list to store all feature_ids
+            feature_ids = []
+            for feature_id in self.features_df["feature_id"].to_list():
+                if feature_id is None:
+                    feature_ids.append(new_int_ids[uid_index])
+                    uid_index += 1
+                else:
+                    feature_ids.append(feature_id)
             
-            self.logger.debug(f"Loaded {len(ms1_df)} MS1 peaks from {sample_path}")
-            return ms1_df
+            # Update the DataFrame with sanitized feature_ids
+            self.features_df = self.features_df.with_columns(
+                pl.Series("feature_id", feature_ids, dtype=pl.Utf8)
+            )
             
-    except Exception as e:
-        self.logger.warning(f"Failed to load MS1 data from {sample_path}: {e}")
-        return pl.DataFrame()
+            self.logger.debug(f"Successfully sanitized {null_feature_ids} feature_id values")
+    
+    # Sanitize consensus_df consensus_id column
+    if hasattr(self, 'consensus_df') and self.consensus_df is not None and not self.consensus_df.is_empty():
+        if "consensus_id" in self.consensus_df.columns:
+            null_consensus_ids = self.consensus_df.filter(pl.col("consensus_id").is_null()).shape[0]
+            if null_consensus_ids > 0:
+                self.logger.debug(f"Sanitizing {null_consensus_ids} null consensus_id values with new UIDs")
+                
+                # Generate new UIDs for null values using the same method as merge()
+                new_uids = [str(uuid.uuid4()).replace('-', '')[:16] for _ in range(null_consensus_ids)]
+                uid_index = 0
+                
+                # Create a list to store all consensus_ids
+                consensus_ids = []
+                for consensus_id in self.consensus_df["consensus_id"].to_list():
+                    if consensus_id is None:
+                        consensus_ids.append(new_uids[uid_index])
+                        uid_index += 1
+                    else:
+                        consensus_ids.append(consensus_id)
+                
+                # Update the DataFrame with sanitized consensus_ids
+                self.consensus_df = self.consensus_df.with_columns(
+                    pl.Series("consensus_id", consensus_ids, dtype=pl.Utf8)
+                )
+                
+                self.logger.debug(f"Successfully sanitized {null_consensus_ids} consensus_id values")
+
+    # Sanitize rt_original in features_df by replacing null or NaN values with rt values
+    if hasattr(self, 'features_df') and self.features_df is not None and not self.features_df.is_empty():
+        if "rt_original" in self.features_df.columns and "rt" in self.features_df.columns:
+            # Check for null or NaN values in rt_original
+            null_or_nan_rt_original = self.features_df.filter(
+                pl.col("rt_original").is_null() | pl.col("rt_original").is_nan()
+            ).shape[0]
+            if null_or_nan_rt_original > 0:
+                self.logger.debug(f"Sanitizing {null_or_nan_rt_original} null or NaN rt_original values with rt values")
+                self.features_df = self.features_df.with_columns(
+                    pl.when(pl.col("rt_original").is_null() | pl.col("rt_original").is_nan())
+                    .then(pl.col("rt"))
+                    .otherwise(pl.col("rt_original"))
+                    .alias("rt_original")
+                )
+                self.logger.debug(f"Successfully sanitized {null_or_nan_rt_original} rt_original values")