Tune arithmetic progression threshold.

src/stratigraphy/util/depthcolumn.py

@@ -265,11 +265,11 @@ def significant_arithmetic_progression(self) -> bool:
         if len(self.entries) < 6:
             return self.is_arithmetic_progression()
         else:
-            # to allow for OCR errors or gaps in the progression, we only require a segment of length 6 that is an
+            # to allow for OCR errors or gaps in the progression, we only require a segment of length 7 that is an
             # arithmetic progression
-            for i in range(len(self.entries) - 6 + 1):
+            for i in range(len(self.entries) - 7 + 1):
                 if BoundaryDepthColumn(
-                    self.noise_count_threshold, self.noise_count_offset, self.entries[i : i + 6]
+                    self.noise_count_threshold, self.noise_count_offset, self.entries[i : i + 7]
                 ).is_arithmetic_progression():
                     return True
             return False
@@ -322,10 +322,14 @@ def is_valid(self, all_words: list[TextLine]) -> bool:
         ):
             return False
         # Check if the entries are strictly increasing.
-        if not all(i.value < j.value for i, j in zip(self.entries, self.entries[1:], strict=False)):
+        if (
+            not all(i.value < j.value for i, j in zip(self.entries, self.entries[1:], strict=False))
+            and len(self.entries) < 15
+        ):
             return False
 
         corr_coef = self.pearson_correlation_coef()
+
         return (
             corr_coef and corr_coef > np.min([1.0382 - len(self.entries) * 0.01, 0.9985]) and corr_coef > 0.95
         )  # Magic numbers obtained using an error analysis on critical borehole profiles. Admittedly, this may
@@ -396,7 +400,7 @@ def correct_typos(self, all_words) -> BoundaryDepthColumn | None:
                     ),
                 )
             )
-        best_columns = max(new_columns, key=lambda column: column[1].pearson_correlation_coef())
+        _removed_entry, best_columns = max(new_columns, key=lambda column: column[1].pearson_correlation_coef())
 
         if best_columns.is_valid(all_words):
             return best_columns

src/stratigraphy/util/find_depth_columns.py

@@ -161,7 +161,6 @@ def find_depth_columns(
     Returns:
         list[BoundaryDepthColumn]: Found BoundaryDepthColumn objects.
     """
-    # print(f"{entries=}")
     numeric_columns: list[BoundaryDepthColumn] = []
     for entry in entries:
         has_match = False
@@ -186,7 +185,7 @@ def find_depth_columns(
             for column in numeric_columns
             if all(not other.strictly_contains(column) for other in numeric_columns)
         ]
-    # print(f"{numeric_columns=}")
+
     numeric_columns = [
         column.reduce_until_valid(all_words)
         for numeric_column in numeric_columns
@@ -195,7 +194,7 @@ def find_depth_columns(
         # that does not match the descriptions
         if not column.significant_arithmetic_progression()
     ]
-    # print(f"{numeric_columns=}")
+
     return sorted(
         [column for column in numeric_columns if column and column.is_valid(all_words)],
         key=lambda column: len(column.entries),