smftools 0.2.1__py3-none-any.whl → 0.2.4__py3-none-any.whl

smftools/preprocessing/filter_reads_on_modification_thresholds.py

@@ -15,7 +15,7 @@ def filter_reads_on_modification_thresholds(
     a_thresholds: Optional[Sequence[float]] = None,
     use_other_c_as_background: bool = False,
     min_valid_fraction_positions_in_read_vs_ref: Optional[float] = None,
-    uns_flag: str = 'reads_filtered_on_modification_thresholds',
+    uns_flag: str = 'filter_reads_on_modification_thresholds_performed',
     bypass: bool = False,
     force_redo: bool = False,
     reference_column: str = 'Reference_strand',
@@ -31,9 +31,9 @@ def filter_reads_on_modification_thresholds(
     - Otherwise, computes the relevant per-read metrics per-reference in batches
       and writes them into adata.obs before filtering.
 
-    Parameters of interest (same semantics as your original function):
+    Parameters of interest :
       - gpc_thresholds, cpg_thresholds, any_c_thresholds, a_thresholds:
-          each should be [min, max] (floats 0..1) or None.
+          each should be [min, max] (floats 0..1) or None. Thresholds are inclusive.
       - use_other_c_as_background: require GpC/CpG > other_C background (if present).
       - min_valid_fraction_positions_in_read_vs_ref: minimum fraction of valid sites
           in the read vs reference (0..1). If None, this check is skipped.
@@ -53,7 +53,7 @@ def filter_reads_on_modification_thresholds(
         col_pref = {
             "GpC": ("Fraction_GpC_site_modified", f"Valid_GpC_site_in_read_vs_reference"),
             "CpG": ("Fraction_CpG_site_modified", f"Valid_CpG_site_in_read_vs_reference"),
-            "C": ("Fraction_any_C_site_modified", f"Valid_any_C_site_in_read_vs_reference"),
+            "C": ("Fraction_C_site_modified", f"Valid_C_site_in_read_vs_reference"),
             "A": ("Fraction_A_site_modified", f"Valid_A_site_in_read_vs_reference"),
         }.get(mod_type, (None, None))
         return (col_pref[0] in adata.obs.columns) and (col_pref[1] in adata.obs.columns)
@@ -99,8 +99,8 @@ def filter_reads_on_modification_thresholds(
             create_cols["Valid_CpG_site_in_read_vs_reference"] = np.full((n_obs,), np.nan)
             create_cols["CpG_to_other_C_mod_ratio"] = np.full((n_obs,), np.nan)
         if "C" in mod_target_bases:
-            create_cols["Fraction_any_C_site_modified"] = np.full((n_obs,), np.nan)
-            create_cols["Valid_any_C_site_in_read_vs_reference"] = np.full((n_obs,), np.nan)
+            create_cols["Fraction_C_site_modified"] = np.full((n_obs,), np.nan)
+            create_cols["Valid_C_site_in_read_vs_reference"] = np.full((n_obs,), np.nan)
         if "A" in mod_target_bases:
             create_cols["Fraction_A_site_modified"] = np.full((n_obs,), np.nan)
             create_cols["Valid_A_site_in_read_vs_reference"] = np.full((n_obs,), np.nan)
@@ -227,7 +227,7 @@ def filter_reads_on_modification_thresholds(
         # any C
         if "C" in mod_target_bases:
             for ref in refs:
-                _compute_for_ref_and_suffix(ref, "any_C_site", create_cols["Fraction_any_C_site_modified"], create_cols["Valid_any_C_site_in_read_vs_reference"])
+                _compute_for_ref_and_suffix(ref, "C_site", create_cols["Fraction_C_site_modified"], create_cols["Valid_C_site_in_read_vs_reference"])
 
         # A
         if "A" in mod_target_bases:
@@ -283,15 +283,15 @@ def filter_reads_on_modification_thresholds(
             filtered = filtered[filtered.obs["GpC_to_other_C_mod_ratio"].astype(float) > 1]
         if lo is not None:
             s0 = filtered.n_obs
-            filtered = filtered[filtered.obs["Fraction_GpC_site_modified"].astype(float) > lo]
+            filtered = filtered[filtered.obs["Fraction_GpC_site_modified"].astype(float) >= lo]
             print(f"Removed {s0 - filtered.n_obs} reads below min GpC fraction {lo}")
         if hi is not None:
             s0 = filtered.n_obs
-            filtered = filtered[filtered.obs["Fraction_GpC_site_modified"].astype(float) < hi]
+            filtered = filtered[filtered.obs["Fraction_GpC_site_modified"].astype(float) <= hi]
             print(f"Removed {s0 - filtered.n_obs} reads above max GpC fraction {hi}")
         if (min_valid_fraction_positions_in_read_vs_ref is not None) and ("Valid_GpC_site_in_read_vs_reference" in filtered.obs.columns):
             s0 = filtered.n_obs
-            filtered = filtered[filtered.obs["Valid_GpC_site_in_read_vs_reference"].astype(float) > float(min_valid_fraction_positions_in_read_vs_ref)]
+            filtered = filtered[filtered.obs["Valid_GpC_site_in_read_vs_reference"].astype(float) >= float(min_valid_fraction_positions_in_read_vs_ref)]
             print(f"Removed {s0 - filtered.n_obs} reads with insufficient valid GpC site fraction vs ref")
 
     # CpG thresholds
@@ -301,15 +301,15 @@ def filter_reads_on_modification_thresholds(
             filtered = filtered[filtered.obs["CpG_to_other_C_mod_ratio"].astype(float) > 1]
         if lo is not None:
             s0 = filtered.n_obs
-            filtered = filtered[filtered.obs["Fraction_CpG_site_modified"].astype(float) > lo]
+            filtered = filtered[filtered.obs["Fraction_CpG_site_modified"].astype(float) >= lo]
             print(f"Removed {s0 - filtered.n_obs} reads below min CpG fraction {lo}")
         if hi is not None:
             s0 = filtered.n_obs
-            filtered = filtered[filtered.obs["Fraction_CpG_site_modified"].astype(float) < hi]
+            filtered = filtered[filtered.obs["Fraction_CpG_site_modified"].astype(float) <= hi]
             print(f"Removed {s0 - filtered.n_obs} reads above max CpG fraction {hi}")
         if (min_valid_fraction_positions_in_read_vs_ref is not None) and ("Valid_CpG_site_in_read_vs_reference" in filtered.obs.columns):
             s0 = filtered.n_obs
-            filtered = filtered[filtered.obs["Valid_CpG_site_in_read_vs_reference"].astype(float) > float(min_valid_fraction_positions_in_read_vs_ref)]
+            filtered = filtered[filtered.obs["Valid_CpG_site_in_read_vs_reference"].astype(float) >= float(min_valid_fraction_positions_in_read_vs_ref)]
             print(f"Removed {s0 - filtered.n_obs} reads with insufficient valid CpG site fraction vs ref")
 
     # any C thresholds
@@ -317,15 +317,15 @@ def filter_reads_on_modification_thresholds(
         lo, hi = _unpack_minmax(any_c_thresholds)
         if lo is not None:
             s0 = filtered.n_obs
-            filtered = filtered[filtered.obs["Fraction_any_C_site_modified"].astype(float) > lo]
+            filtered = filtered[filtered.obs["Fraction_C_site_modified"].astype(float) >= lo]
             print(f"Removed {s0 - filtered.n_obs} reads below min any-C fraction {lo}")
         if hi is not None:
             s0 = filtered.n_obs
-            filtered = filtered[filtered.obs["Fraction_any_C_site_modified"].astype(float) < hi]
+            filtered = filtered[filtered.obs["Fraction_C_site_modified"].astype(float) <= hi]
             print(f"Removed {s0 - filtered.n_obs} reads above max any-C fraction {hi}")
-        if (min_valid_fraction_positions_in_read_vs_ref is not None) and ("Valid_any_C_site_in_read_vs_reference" in filtered.obs.columns):
+        if (min_valid_fraction_positions_in_read_vs_ref is not None) and ("Valid_C_site_in_read_vs_reference" in filtered.obs.columns):
             s0 = filtered.n_obs
-            filtered = filtered[filtered.obs["Valid_any_C_site_in_read_vs_reference"].astype(float) > float(min_valid_fraction_positions_in_read_vs_ref)]
+            filtered = filtered[filtered.obs["Valid_C_site_in_read_vs_reference"].astype(float) >= float(min_valid_fraction_positions_in_read_vs_ref)]
             print(f"Removed {s0 - filtered.n_obs} reads with insufficient valid any-C site fraction vs ref")
 
     # A thresholds
@@ -333,15 +333,15 @@ def filter_reads_on_modification_thresholds(
         lo, hi = _unpack_minmax(a_thresholds)
         if lo is not None:
             s0 = filtered.n_obs
-            filtered = filtered[filtered.obs["Fraction_A_site_modified"].astype(float) > lo]
+            filtered = filtered[filtered.obs["Fraction_A_site_modified"].astype(float) >= lo]
             print(f"Removed {s0 - filtered.n_obs} reads below min A fraction {lo}")
         if hi is not None:
             s0 = filtered.n_obs
-            filtered = filtered[filtered.obs["Fraction_A_site_modified"].astype(float) < hi]
+            filtered = filtered[filtered.obs["Fraction_A_site_modified"].astype(float) <= hi]
             print(f"Removed {s0 - filtered.n_obs} reads above max A fraction {hi}")
         if (min_valid_fraction_positions_in_read_vs_ref is not None) and ("Valid_A_site_in_read_vs_reference" in filtered.obs.columns):
             s0 = filtered.n_obs
-            filtered = filtered[filtered.obs["Valid_A_site_in_read_vs_reference"].astype(float) > float(min_valid_fraction_positions_in_read_vs_ref)]
+            filtered = filtered[filtered.obs["Valid_A_site_in_read_vs_reference"].astype(float) >= float(min_valid_fraction_positions_in_read_vs_ref)]
             print(f"Removed {s0 - filtered.n_obs} reads with insufficient valid A site fraction vs ref")
 
     filtered = filtered.copy()

smftools/preprocessing/flag_duplicate_reads.py

@@ -13,7 +13,7 @@ import pandas as pd
 import matplotlib.pyplot as plt
 from tqdm import tqdm
 
-from ..informatics.helpers import make_dirs
+from ..readwrite import make_dirs
 
 # optional imports for clustering / PCA / KDE
 try:
@@ -77,7 +77,7 @@ def flag_duplicate_reads(
     sample_col: str = "Barcode",
     output_directory: Optional[str] = None,
     metric_keys: Union[str, List[str]] = ("Fraction_any_C_site_modified",),
-    uns_flag: str = "read_duplicate_detection_performed",
+    uns_flag: str = "flag_duplicate_reads_performed",
     uns_filtered_flag: str = "read_duplicates_removed",
     bypass: bool = False,
     force_redo: bool = False,

smftools/preprocessing/invert_adata.py

@@ -1,6 +1,6 @@
 ## invert_adata
 
-def invert_adata(adata, uns_flag='adata_positions_inverted', force_redo=False):
+def invert_adata(adata, uns_flag='invert_adata_performed', force_redo=False):
     """
     Inverts the AnnData object along the column (variable) axis.
 

smftools/preprocessing/load_sample_sheet.py

@@ -2,7 +2,7 @@ def load_sample_sheet(adata,
                       sample_sheet_path, 
                       mapping_key_column='obs_names',
                       as_category=True,
-                      uns_flag='sample_sheet_loaded',
+                      uns_flag='load_sample_sheet_performed',
                       force_reload=True
                         ):
     """

smftools/preprocessing/reindex_references_adata.py

@@ -0,0 +1,37 @@
+## reindex_references_adata
+
+def reindex_references_adata(adata, 
+                             reference_col="Reference_strand", 
+                             offsets=None, 
+                             new_col="reindexed",
+                             uns_flag='reindex_references_adata_performed', 
+                             force_redo=False):
+    
+    # Only run if not already performed
+    already = bool(adata.uns.get(uns_flag, False))
+    if (already and not force_redo):
+        return None
+    
+    if offsets is None:
+        pass
+    else:
+        # Ensure var_names are numeric
+        var_coords = adata.var_names.astype(int)
+
+        for ref in adata.obs[reference_col].unique():
+            if ref not in offsets:
+                pass
+            else:
+                offset_value = offsets[ref]
+
+                # Create a new var column for this reference
+                colname = f"{ref}_{new_col}"
+
+                # Add offset to all var positions
+                adata.var[colname] = var_coords + offset_value
+
+    # mark as done
+    adata.uns[uns_flag] = True
+
+    print("Reindexing complete!")
+    return None