From 404dc2c9ed1187083ee10a8800c414a7b84a0e02 Mon Sep 17 00:00:00 2001
From: Emily Kellison-Linn <4672118+emilykl@users.noreply.github.com>
Date: Wed, 6 Sep 2023 15:57:34 -0400
Subject: [PATCH 1/2] improve performance by calling file.tell() only when
 needed

---
 lasio/reader.py | 39 ++++++++++++++++++++++++++++-----------
 1 file changed, 28 insertions(+), 11 deletions(-)

diff --git a/lasio/reader.py b/lasio/reader.py
index 9255c90..1a66b12 100644
--- a/lasio/reader.py
+++ b/lasio/reader.py
@@ -300,18 +300,18 @@ def find_sections_in_file(file_obj):
     file_pos = int(file_obj.tell())
     starts = []
     ends = []
-    line_no = 0
-    line = file_obj.readline()
-    # for i, line in enumerate(file_obj):
+    line_no = -1
+    line = True
+
     while line:
+        line = file_obj.readline()
+        line_no = line_no + 1
         sline = line.strip().strip("\n")
         if sline.startswith("~"):
+            file_pos = int(file_obj.tell()) - full_line_length(line, file_obj)
             starts.append((file_pos, line_no, sline))
             if len(starts) > 1:
                 ends.append(line_no - 1)
-        file_pos = int(file_obj.tell())
-        line = file_obj.readline()
-        line_no = line_no + 1
 
     ends.append(line_no)
     section_positions = []
@@ -320,6 +320,18 @@ def find_sections_in_file(file_obj):
     return section_positions
 
 
+def full_line_length(line, file_obj):
+    """Return the full length of the line in characters, adjusted
+    based on the type of newline separators used ('\\n' vs. '\\r\\n')"""
+    newlines = (
+        file_obj.newlines
+        if not isinstance(file_obj.newlines, tuple)
+        else file_obj.newlines[0]
+    )
+    newline_adjust = len(newlines) - 1 if newlines is not None else 0
+    return len(line) + newline_adjust
+
+
 def determine_section_type(section_title):
     """Return the type of the LAS section based on its title
 
@@ -454,7 +466,7 @@ def read_data_section_iterative_normal_engine(
     title = file_obj.readline()
 
     def items(f, start_line_no, end_line_no):
-        for line_no, line in enumerate(f, start=start_line_no+1):
+        for line_no, line in enumerate(f, start=start_line_no + 1):
             line = line.strip("\n").strip()
             if line.startswith(ignore_data_comments):
                 continue
@@ -581,10 +593,15 @@ def read_data_section_iterative_numpy_engine(file_obj, line_nos):
     file_obj.seek(0)
 
     # unpack=True transforms the data from an array of rows to an array of columns.
-    # loose=False will throw an error on non-numerical data, which then sends the 
+    # loose=False will throw an error on non-numerical data, which then sends the
     # parsing to the 'normal' parser.
     array = np.genfromtxt(
-        file_obj, skip_header=first_line, max_rows=max_rows, names=None, unpack=True, loose=False
+        file_obj,
+        skip_header=first_line,
+        max_rows=max_rows,
+        names=None,
+        unpack=True,
+        loose=False,
     )
 
     # If there is only one data row, np.genfromtxt treats it as one array of
@@ -592,7 +609,7 @@ def read_data_section_iterative_numpy_engine(file_obj, line_nos):
     # converts the single line data array to an array of arrays(column data).
     if len(array.shape) == 1:
         arr_len = array.shape[0]
-        array = array.reshape(arr_len,1)
+        array = array.reshape(arr_len, 1)
 
     return array
 
@@ -1035,7 +1052,7 @@ def configure_metadata_patterns(line, section_name):
     # 3. double_dots '..' caused by mnemonic abbreviation (with period)
     #    next to the dot delimiter.
     if ":" in line:
-        if not "." in line[:line.find(":")]:
+        if not "." in line[: line.find(":")]:
             # If there is no period, then we assume that the colon exists and
             # everything on the left is the name, and everything on the right
             # is the value - therefore no unit or description field.

From 41dc97316ae88a0dcff5c1f90c5bfcb389c7c920 Mon Sep 17 00:00:00 2001
From: Emily Kellison-Linn <4672118+emilykl@users.noreply.github.com>
Date: Thu, 7 Sep 2023 14:16:54 -0400
Subject: [PATCH 2/2] handle files with no newline at EOF(?)

---
 lasio/reader.py | 16 +++++++++++-----
 1 file changed, 11 insertions(+), 5 deletions(-)

diff --git a/lasio/reader.py b/lasio/reader.py
index 1a66b12..8ded9b7 100644
--- a/lasio/reader.py
+++ b/lasio/reader.py
@@ -297,7 +297,7 @@ def find_sections_in_file(file_obj):
         or line breaks.
 
     """
-    file_pos = int(file_obj.tell())
+    file_pos = 0
     starts = []
     ends = []
     line_no = -1
@@ -314,9 +314,10 @@ def find_sections_in_file(file_obj):
                 ends.append(line_no - 1)
 
     ends.append(line_no)
-    section_positions = []
-    for j, (file_pos, first_line_no, sline) in enumerate(starts):
-        section_positions.append((file_pos, first_line_no, ends[j], sline))
+    section_positions = [
+        (file_pos, first_line_no, ends[j], sline)
+        for j, (file_pos, first_line_no, sline) in enumerate(starts)
+    ]
     return section_positions
 
 
@@ -329,7 +330,11 @@ def full_line_length(line, file_obj):
         else file_obj.newlines[0]
     )
     newline_adjust = len(newlines) - 1 if newlines is not None else 0
-    return len(line) + newline_adjust
+    if line.endswith("\n"):
+        length = len(line) + newline_adjust
+    else:
+        length = len(line)
+    return length
 
 
 def determine_section_type(section_title):
@@ -602,6 +607,7 @@ def read_data_section_iterative_numpy_engine(file_obj, line_nos):
         names=None,
         unpack=True,
         loose=False,
+        dtype=float,
     )
 
     # If there is only one data row, np.genfromtxt treats it as one array of