feat(repeats): adding examples from HGVS repeats recommendations page…

…. Improved repeat analyzer, that now supports 5' and 3' shuffling.

src/hgvs/pretty/console/tx_pos.py

@@ -11,7 +11,10 @@ def display(self, data: VariantData) -> str:
         """show the position of the transcript seq"""
         var_str = ""
 
-        rna_coding = data.var_c_or_n.ac.startswith("NR_")
+        if data.var_c_or_n:
+            rna_coding =  data.var_c_or_n.ac.startswith("NR_") 
+        else:
+            rna_coding = False
 
         count = -1
         for pdata in data.position_details:

src/hgvs/pretty/datacompiler.py

@@ -345,6 +345,8 @@ def data(
         # for p in position_details:
         #     print(f"{p}\t{p.protein_data}\t")
 
+        strand = mapper.strand if mapper else 1
+
         vd = VariantData(
             seq_start,
             seq_end,
@@ -355,7 +357,7 @@ def data(
             tx_seq,
             mapper,
             var_g,
-            mapper.strand,
+            strand,
             var_c_or_n,
             var_p,
             position_details,

src/hgvs/repeats.py

@@ -1,73 +1,236 @@
 # -*- coding: utf-8 -*-
 """ A class to manage conversion of SequenceVariants to repeat representation"""
 
+from dataclasses import dataclass
 import re
-from typing import Optional
 
 from hgvs.pretty.models import VariantCoords
 
 
-def count_pattern_occurence(pattern, string):
-    """Counts how often a pattern can be found in the string."""
-    matches = re.finditer(f'(?={pattern})', string)
-    return sum(1 for _ in matches)
-
-
-def count_repetitive_units(s):
-    length = len(s)
-
-    for i in range(1, length + 1):
-        unit = s[:i]
-        if length % i == 0:
-            if unit * (length // i) == s:
-                return length // i, unit
-
-    return 1, s
+@dataclass(eq=True, repr=True, frozen=True, order=True)
+class RepeatUnit:
+    repeat_count: int
+    repeat_unit:str
+    block_size:int
+    block: str
 
 
 class RepeatAnalyser:
 
-    def __init__(self, fs: VariantCoords) -> None:
+    def __init__(self, fs: VariantCoords, reverse:bool=False) -> None:
 
         self.is_repeat = False
-        self.ref_count = 0
-        self.alt_count = 0
+        self.repeat_units_ref = None
+        self.repeat_units_alt = None
         self.ref_str = fs.ref
         self.alt_str = fs.alt
+        self.reverse = reverse
 
-        self.repeat_unit = self._get_repeat_unit(fs)
-        if self.repeat_unit is None:
+        self.repeat_units_ref = detect_repetitive_block_lengths(self.ref_str, self.reverse)
+        self.repeat_units_alt = detect_repetitive_block_lengths(self.alt_str, self.reverse)
+
+        if len(self.repeat_units_ref) == 0 and len(self.repeat_units_alt) ==0 :
+            return
+
+        # check longest repeat blocks:
+        # we only look at ref to determine if there are repeats 
+        # If ref has no repeat, we don't call this a repeat variant, even if alt would have a repetitive unit
+        longest_r_unit = self._get_longest_repeat_unit(self.repeat_units_ref)
+        if longest_r_unit is None:
+            return
+
+        # filter our too fragmented results
+        expected_size = len(self.ref_str) / 3
+        if longest_r_unit.block_size < expected_size:            
             return
+
+        if longest_r_unit.repeat_unit not in self.alt_str:
+            return
+
 
         self.is_repeat = True
 
-        self.ref_count = count_pattern_occurence(self.repeat_unit, fs.ref)
-        self.alt_count = count_pattern_occurence(self.repeat_unit, fs.alt)
-        self.ref_str = f"{self.repeat_unit}[{self.ref_count}]"
-        self.alt_str = f"{self.repeat_unit}[{self.alt_count}]"
+        ref_repeat = get_repeat_str(self.ref_str, self.alt_str, self.repeat_units_ref, self.repeat_units_alt, self.reverse)
+        alt_repeat = get_repeat_str(self.alt_str, self.ref_str, self.repeat_units_alt, self.repeat_units_ref, self.reverse)
+
+        self.ref_str = ref_repeat
+        self.alt_str = alt_repeat
 
     def __repr__(self):
         return f"{self.ref_str}>{self.alt_str}"
 
-    def _get_repeat_unit(self, fs: VariantCoords) -> Optional[str]:
-        """Takes fully justified coordiantes and tries to detect a repeat in them."""
-        # analyze for repeat:
-        if len(fs.ref) == len(fs.alt) > 0:
-            # seems we cant shuffle. is an SVN or delins
-            return None
-
-        if len(fs.alt) > 0:
-            if fs.alt in fs.ref:
-                if fs.ref.startswith(fs.alt):
-                    d = fs.ref[len(fs.alt) :]
-                    if count_pattern_occurence(d, fs.ref) > 1:
-                        c, u = count_repetitive_units(d)
-                        return u
-            elif fs.ref in fs.alt:
-                if fs.alt.startswith(fs.ref):
-                    d = fs.alt[len(fs.ref) :]
-                    if count_pattern_occurence(d, fs.ref) > 1:
-                        c, u = count_repetitive_units(d)
-                        return u
+    def _get_longest_repeat_unit(self, repeat_units:list[RepeatUnit])->RepeatUnit:
+        lru = None
+        for ru in repeat_units:
+            if not lru:
+                lru = ru
+                continue
+
+            if ru.block_size > len(lru.block):
+                lru = ru
+
+        return lru
+
+
+def detect_repetitive_block_lengths(sequence: str, reverse: bool = False) -> list[RepeatUnit]:
+    """Detects the length of repetitive blocks in a string, with an option to search from left to right or reverse.
+
+    In reverse mode, it creates the largest possible blocks of the smallest possible units.
+    """
+    result: list[RepeatUnit] = []
+    seq_len = len(sequence)
+
+    if reverse:
+        i = seq_len  # Start from the end of the sequence
+        while i > 0:
+            matched = False
+            # Iterate over block sizes from smallest to largest
+            for block_size in range(1, seq_len // 2 + 1):
+                if i - block_size < 0:
+                    continue  # Not enough characters to form a repeat unit
+
+                # Extract the potential repeat unit ending at position i
+                repeat_unit = sequence[i - block_size:i]
+
+                # Calculate the maximum possible number of repeats for this block size
+                max_possible_repeats = i // block_size
+
+                # Initialize repeat_count to the maximum possible and decrease
+                for repeat_count in range(max_possible_repeats, 1, -1):
+                    start_index = i - repeat_count * block_size
+                    if start_index < 0:
+                        continue  # Not enough characters to form the repetitive block
+
+                    # Extract the substring that could be the repetitive block
+                    substr = sequence[start_index:i]
+
+                    # Build the regex pattern for the current repeat unit and count
+                    pattern = rf'({re.escape(repeat_unit)})' + r'{' + f'{repeat_count}' + r'}'
+
+                    # Check if the substring matches the pattern
+                    if re.fullmatch(pattern, substr):
+                        repetitive_block = substr
+                        ru = RepeatUnit(
+                            block=repetitive_block,
+                            block_size=len(repetitive_block),
+                            repeat_unit=repeat_unit,
+                            repeat_count=repeat_count
+                        )
+                        result.append(ru)
+                        # Move the index `i` backward by the length of the repetitive block
+                        i -= len(repetitive_block)
+                        matched = True
+                        break  # Found the largest block for this unit size
+
+                if matched:
+                    break  # Proceed to the next position `i`
+
+            if not matched:
+                # No repeat found, remove one character
+                ru = RepeatUnit(
+                    block=sequence[i - 1],
+                    block_size=1,
+                    repeat_unit=sequence[i - 1],
+                    repeat_count=1
+                )
+                result.append(ru)
+                i -= 1  # Move back by one character if no match is found
+
+    else:
+        i = 0  # Start from the beginning of the sequence
+        while i < seq_len:
+            matched = False
+            for block_size in range(1, seq_len // 2 + 1):
+                # Build the regex pattern for the current block size
+                pattern = rf'(.{{{block_size}}})\1+'
+                match = re.match(pattern, sequence[i:])
+
+                if match:
+                    repetitive_block = match.group()  # The full repeating pattern
+                    repeated_unit = match.group(1)    # The repeating unit
+                    repetition_count = len(repetitive_block) // len(repeated_unit)
+
+                    # Add the repetitive block and its details to the result
+                    ru = RepeatUnit(
+                        block=repetitive_block,
+                        block_size=len(repetitive_block),
+                        repeat_unit=repeated_unit,
+                        repeat_count=repetition_count
+                    )
+                    result.append(ru)
+
+                    # Move the index `i` forward by the length of the repetitive block
+                    i += len(repetitive_block)
+                    matched = True
+                    break  # Found a match, break the loop for current block size
+
+            if not matched:
+                i += 1  # Move forward by one character if no match is found
+
+    return result
+
+
+def get_repeat_str(
+    sequence: str,
+    other_seq: str,
+    primary_repeat_unit: list[RepeatUnit],
+    other_repeat_unit: list[RepeatUnit],
+    reverse: bool = False
+) -> str:
+    if len(primary_repeat_unit) == 0 and len(other_repeat_unit) == 0:
+        return None
 
+    if len(primary_repeat_unit) == 0 and len(other_repeat_unit) == 1 and sequence == other_repeat_unit[0].repeat_unit:
+        return f"{sequence}[1]"
+    elif len(primary_repeat_unit) == 0 and len(other_repeat_unit) == 1 and sequence != other_repeat_unit[0].repeat_unit:
         return None
+
+    if len(primary_repeat_unit) > 0 and len(other_repeat_unit) > 0:
+        return_str = assemble_repeat_string(sequence, primary_repeat_unit, reverse=reverse)
+        return return_str
+
+    if len(other_repeat_unit) == 0 and len(other_seq) > 0:
+        return_str = assemble_repeat_string(sequence, primary_repeat_unit, reverse=reverse)
+        if len(return_str) > 0:
+            return return_str
+
+    return None
+
+def assemble_repeat_string(sequence: str, repeat_units: list[RepeatUnit], reverse: bool = False) -> str:
+    return_str = ""
+    primary_repeat_unit = repeat_units.copy()
+    seq = sequence
+
+    if reverse:
+        while len(seq) > 0:
+            found_unit = None
+            for ru in primary_repeat_unit:
+                if seq.endswith(ru.block):
+                    return_str = f"{ru.repeat_unit}[{ru.repeat_count}]" + return_str
+                    seq = seq[:-len(ru.block)]
+                    found_unit = ru
+                    break
+            if not found_unit:
+                # remove one character if no matching repeat unit is found
+                seq_char = seq[0]
+                seq = seq[:-1]
+                return_str = f"{seq_char}[1]" + return_str
+
+    else: # forward direction
+        while len(seq) > 0:
+            found_unit = None
+            for ru in primary_repeat_unit:
+                if seq.startswith(ru.block):
+                    return_str += f"{ru.repeat_unit}[{ru.repeat_count}]"
+                    seq = seq[len(ru.block):]
+                    found_unit = ru
+                    break
+            if not found_unit:
+                # remove one character if no matching repeat unit is found
+                seq_char = seq[0]
+                seq = seq[1:]
+                return_str += f"{seq_char}[1]"
+
+
+    return return_str
+